. ‘§.“€I :fi. -‘v' 7'. In Andrew J. Petro ABSTRACT ENPLOYNENT SHIFTS AND CITY SIZE: MICHIGAN AND INDIANA. 1950-1963 by Andrew J. Petro A striking feature of the 1950's was the "metropolitan explo- sion". This feature produced many human and economic problems for cities. and has brought urban studies to the forefront. It is in the cities that modern development has historically occurred and where changes in economic activity find their incidence. The economic activity in a city participates as a unit in a system. As such a unit the city is influenced by the national eco- nomy and the city's own characteristics. When the system grows. the city can gain or lose economic activity. or gain but at a slower rate than previously. This change presents problems to the city because the change is reflected in the amount of employment. a matter of pub- lic concern. Since the national growth is distributed unevenly among the cities. a city can obtain a growth rate equal or unequal to that of the nation. A shift technique is used to measure this growth or decline. The employment growth in a city could be expected to grow at approximately the same rate as the national employment. Applying this rate to the city employment would give the expected employment in the city if it had grown at the national rate. The actual emp- loyment in the city can be greater or less than the expected employ- ment. The difference between the actual and expected employment is the employment shift. Q 5"»- -v ... \ Ekflc c.. . AVA +he a‘ ‘n ma vk‘ 1C Andrew J. Petro The problems for a city are to determine the causes of the employment shift, regardless of their favorability, and secondly, to determine whether or not the shift is significantly influenced by the city characteristics of size, industrial composition and population. The shift technique is applied to total employment, to emp- loyment by sectors and to industries in the manufacturing sector. Employment data used are the published data by county or counties for selected cities for the period 1950-1963. Standard regression analysis is used to determine the significance of city size and in- dustrial composition on the employment shifts. The analysis indicates that the employment shifts and population size were uncorrelated and that the composition effect had little impact on the employment shifts. The shift of total employment in Michigan was favorable, that is. the actual employment was greater than the exnected employment; the shift was unfavorable for the cities in Indiana. The use of the shift technique indicates that the more significant source of the total shift was the local-factor shift. This local-factor shift is the effect of the locational advantages of the cities. In both states. the local-factor shift swamped the composition shift; the composition shift was generally one—third or less the size of the local-factor shift. The sector with the largest impact on total employment is the manufacturing sector. The shift was negative for both states; Inachinery (electrical and non—electrical) and automobile and trans- laortation industries provided the largest negative shifts. This de- <:line came from the negative impact of the local-factor effect. The (Bities suffered a relative loss of their locational advantages that Andrew J. Petro was not sufficiently offset by the growth industries within the sector. It is normally thought that the larger size city is more apt to grow. If city size were significant for the shift, then the shift would diverge sharply for the larger size cities. This di- vergence did not appear in the analysis. The use of the shift technique shows that the cause of the employment shift, positive or negative. is the local-factor effect; little impact comes from the composition effect. The relative growth or decline in employment is independent of size and industrial com- position. FFPIOYNEFT SglpTS AND CITY SIZE: IvTICHIGAN ATTD IT‘fDIASEA. 19.50-1963 by Andrew J. Petro A TPESIS Submitted to Ffichiaan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics 1966 .9:— I a“. r~+ TABLE OF CONTENTS Page IJIST OF T.ABIJES O O O O O O O 0 O 0 O O O O O O O O C O O O O O O O O O O O O O O O O O O O O O O O Q 0 O O 0 0 iii- LIST OF FIGLIPES 0 O O O O 0 O O O O O 0 O O O O 0 0 O O O O O O O O O O O O O O O O O O O O O O 0 O O O O O 0 v LIST OF APPENDICES .. ........ . ......... ... ..................... vi CHAPTER I. INTRODUCTION ......................................... 1 II. APPROACHES TO FEOIOIAL ECONOMIC ANALYSIS ............. 6 III. A FEAFEVOEI FOP. THE ANALYSIS OF EFTLOITEET CHAT-ICES . .. 15 IV. Er-TDIOIEEDT SHIFTS AED CITY SIZE 25 V. INDUSTRY ET-TIOpIEIT AND CITY SIZE 34 VI. SHIFTS DI IT‘TDUSTPY EI-TLO‘OEITT .. #6 VII. TEE MAWFACTUEEIC SECTOR SHIFTS .......... 60 VIII. MANUFACTUDIDO SFIFTS AND CITY SIZE .. 71 IX. COI‘FCLUSIOUS ....... 83 AFFETDICES.............. ...... 91 BIBLIOGPAPM 00000000000000000 00000 0000 00000000 0000000000000000 1-17 ii 10. 11. 12. 13. 14. IIST OF TARIES Tot-13.1 mnlo‘ment’ 1950-1963 0.000.000.0000.OOOOOOOOOOOOO Total Eleoyment and Net Shift, Michigan and Indiana.’ 1950-1963 ......OOOOOOO.....OOOOOOOIOOOOOOOOOO Correlation Coefficients of Total Employment and Population for Selected Cities in Michigan and Indiana, 1950-1963 0.00.0000.........OOOOOOOOOOOOOOOOOO Correlation Coefficients of Changes in Tbtal Employment and Population for Selected Cities in Nichigan and Indiana. 1950-1963 .................... Correlation Coefficients of Total Employment Shift and Change in Population for Selected Cities in Iva—011108.11 a.nd h(jiana ’ 1950-1963 0 O O O O C O O O O O O C O O O O O O O O O . Correlation Coefficients of Employment and Pepulation Shifts for Selected Cities in Michigan and Indiana, 1950-1963 0.0.0........OOOOOOOOOOOOOOOOO0.0.0.0.0000... Employment Changes by Industry Sectors: Michigan and Indiana, ]_QSO-]_963 00......O0.0.0.0...00.000.000.000... Pevression and Correlation Coefficients of Sector Employment and City Size for Selected Cities in Vichigan and Indiana .................................. Pegression and Correlation Coefficients of Sector Emoloyment and Citv Size for Selected Cities by S+'at!es 0.......00..........OOOOOOOOOOOOOOO00.00.0000... Regression and Correlation Coefficients of Changes in Industry EMUloyment and City Size for Selected Cities inf-1011175311 and II’Idiflha 00000000000000.000000000 Pevression and Correlation Coefficients of Changes in Industry Employment and City Size for Selected Cities, 1950-1963 ...I0.00.00.00.00......OOCOCOO0.0.... Mean and Kean Deviations of Changes in Employment by sectors, 1950-]_963 I0.0.0..........OOOOOOOOOOOOOOOOO Employment Shifts by Industry Sectors: hichigan and Indiana, 1950—1963 .................................... Employment Shifts of Industry Sectors by States. 1950-1963 ......‘O...’...O......O......COOOOOOOD...OO.. iii Page 16 26 28 29 30 35 no #1 42 47 50 Table 16. 20. 2]. 22. 23. 2h. 25. 26. 27. iv Total Jet Employment Shift for Sectors From Local- Factor and Composition EIIects for Selected Cities in Fichisan and Indiana, lQSO-l963 ........... Total Vet Employment Shift for Qpn+ors from local- W‘CIO” ”“3 Cfimpnsi+ion Effects for Selected Caries. 1-960-10é? ...I0.00.0.0...9000............OOOOOOOOO... Correlation Coefficients of ”et Shifts and City Size, logo-1063 0............O...O..0.0.0.......OOIOOOOOOOO Correlation Coefficients of Shifts and Chances in Citv Size for Selected Cities. 1950-1963 -------~---- Relative Employment Shifts in Phnufacturina. All Cities, 19:30-1963 ......OOOOOOOO......OOOOOOOOOOOOOOO Vet Shifts in Kanufacturins EMployment for Selected Ciflies b‘r S+latleS’ quo-1963 ......OOOOCOC.....O...... ts lhnufacturing. Selected “elative Employment Stif in .60-1963 ......OOOOOOOCOCOOOOO... Cities in Indiana, 19 Relative Employment Shifts in Ppnufacturins, Selected Cities in I'fiphidqh, 1950-1963 0.000000000000000000000 Components of the Manufacturing Vet Shift for Selected Cities. 1950-1963 0.0.00000000000000000000.0.00000000 Recression and Correlation Coefficients: lhnufacturing Shifts and Size for Selected Cities in Michigan and In(3i-qnn’ ]_950-1_963 00.000.00.000.........OOOOOOOOOOOO Regression and Correlation Coefficients: Shifts and Chances in Earnings Per Week for Selected Cities in ifichigan and Indiana, 1950-1963 .................. thufacturing Employment Shifts by Kind of Activitv for Selected Cities in Michigan and Indiana. 1950- 1963 ................................................. Pearession and Correlation Coefficients of abricatinc Shifts and Chances in Earninvs Per Week for Selected Cities, 1950-1963 ........................... Pare Sl 74 76 78 91 Figure LIST OF FIGURES Actual and Expected Employment Over Time ............... Possible Employment by Size and Structure .............. Total Employment and City Size ......................... Total Employment and Citv Size by States ............... a. Distribution of Employment by Industry. All Cities, 1950-1963 0000000000000. 00000 0 00000000 000000 b. Distribution of Employment Change by Industry, All Cities, 1950-1963 0............OOOOOOOOOOOOOOOO. Distribution of Employment by Industry Sectors: Grouped by largest and Smallest Size Cities .......... Distribution of Employment by Industry Sector: Actual and mectw ....0............OOOOOOOOOOOOOOOOO Actual and Expected Employment Grow+h by Size Group .... Actual and Expected Fanufacturing Employment Curves for Selected Cities in Indiana, 1963 ................. Actual and Expected Fanufacturina Employment Curves for Selected Cities in Fichigan, 1963 ................ Page 18 19 31 32 66 Appendix A. B. :3 0 LIST OF APPEEDICES Pare Population, Expected Population and Population Shifts by Cities. Richigan and Indiana ........... 91 Employment by Industry Sector for Selected Cities. Michiaap and Indiana, 1950-1963 .................. 92 Manufacturins Employment by Two-Dinit Standard Indus +rial Classification for Selected Cities. :‘I'iChidgn and Indiana, 1960-1963 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 100 Source of Net Shift of Employment by Sectors tor seleC+ed Cj_‘tieS, 1960-19k33 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 111 Source of Net Shift of Employment in Bhnufacturirq for. selec+.ed Cities, 1960-1963 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 0 ]_]_2 Local-Eactor Shift in Processinv and Fabricating Industries for Selected Cities, 1950-1963 ........ 113 Pesression and Correlation Coefficients of local- Factor Sbift in Fabricatirv Industries and City Size and Chanses in City Size for Selected Cities. 1950—1963 ................................ llh Correlation Coefficients of Chances in Weekly Earnings and Changes in Total Employment. City Size and Manufacturing EMploymept for Selected Cities. 1950-1963 ................................ 11h Technical Note on the Computation of the Shifts .... 115 vi CHAPTER I INTRODUCTION Nbdern regional analysis has focused on many theoretical and practical problems. In response to such problems. its development has been greatly accelerated so that economists today can use many analya tic forms. types. and degrees of rigor to solve problems.1 ‘Within this development. there has been a shift in emphasis from national- regional analysis to the problems of smaller areas. The change in emphasis. however. is not a rejection of the logic of space and acti- vity at a higher level of aggregation, but an acknowledgment that problems exist in these smaller economic units that are linked to form the region or nation. It is this kind of awareness that has brought the analysis of urban problems to the fore-front.2 This attention to urban problems is coupled with a pragmatic attitude that arose with the changing conditions over the long run. and particularly the imme- diate past. The decade of the 1950's saw a national economy become more concerned with the problems of abundance and a high growth rate than with mass unemployment. The change in concern did not, of course, free 1J. Meyer. "Regional Analysis: A Survey", American Economic Review. Vbl. LIII. march 1963. 2Meyer, ibid.. pp. 26-29. file 2 the nation of any unemployment problems. or problems of other kinds. for example. inflation. However, the elimination of mass unemploy- ment was not the chief concern that it had been in earlier periods.3 The acceleration of technology and organization along with the capa— city and ability to satisfy many wants became an economic fact. The result of the rapid prosress was the acceptance of a high growth rate, with high employment as the goal. The acceptance of the goal. and the problems associated with abundance, have continued into the first part of the 1960's. A striking feature of the period was the so-called "metro- politan explosion". This metropolitan explosion was characterized by a rise in city population at a rate higher than the rate for the na- tion and by the urbanization of areas surrounding the central city.“ High birth rates as well as rural miaration contributed to the abso- lute rise in urban population. The striking feature was not the trend, for this was old and continuous. but the fact that it was ac- celerating. The number of places that had a population of 50,000 or more nearly doubled in that decade.5 These developments in the 1950's produced both human and economic problems for the cities. The prob- lems. while not purely economic, are in many ways the reflections of BE. G. Vatter. The U.S. Economy in the 1950's. w; w; Norton & Company. Inc., New York. 1963. pp. 4-5. “vetter, 92. git., pp. 22-23. R. Vernon, The Changins Eco- npmic Function of the Central City. Committee for Economic Develop- ment. 1959. 5U. S. Bureau of the Census. ngulation Trends in the U1§. 1900 to 1963, Technical Paper No. 10. U.S. Government Printing Office.'Washington. D. C., 1964. Table l, p. 16. t}: so 3 the changing industrial and commercial activities in general. Indeed, the metropolitan growth reflects both the expansion of population and the economic development of the nation that brought about the astound- ing rise in productivity and living standard. The problems associated with "exploding metropolis" have not abated in the early 1960's. and all the necessary measures to deal with them have not been developed.6 These facts have prompted the suggestion that new analytical tools be fashioned.7 Certainly. modern analysis will continue to develop new tools and approaches in various and various degrees of rigor. The distinct changes in the 1950's have made the period fruitful for the development of new tools and insights in the analysis of city growth. This study analyzes metropolitan areas of various sizes for two states over the period of 1950 to 1963. The metropolitan area is used as the unit of aggregation because it is here that modern develop- ment has historically occurred. It is also used because it is here that any changes or adjustments in economic activity. whatever their source. have their incidence. As such. the unit has a form of its own as pattern and structure within a system.8 6V'atter. 22, cit.. p. 22. 7B. Chinitz, "Contrast in Agglomeration: New York and Pitts- burgh". Papers and Proceedings. American Economic Review, Vbl. LI. Phy'l961. p. 279. 8R. Vining. "Delimitation of Economic Areas: Statistical Conceptions in the Study of the Spatial Structure of an Economic System", Journal of American Statistical Association. V01. 48, 1953. p. 52. 2‘3 . t" - ab 0'. .‘J .L'L) (I) 0) I!” L. The metropolitan unit is affected by two broad forces: the national economy and the city's own characteristics. If the condi- tions in the nation change. the city is affected because it is a unit in the system. The extent to which the city is favorably or unfavor- ably affected also depends on the particular make-up of the city: population. size. specialization. industrial composition. etc. As the system grows the city can lose or gain activity and industry. or gain but at a slower rate than it gained previously. This change presents problems to the city because the changes are reflected in the amount of employment. a fact of public concern. As a result many questions arise. Why the change? What is the impact of industrial composition? What about cities of other sizes? What is the effect of population changes? Is the actual employment gain lagging behind the national gain? In any immediate employment problem of a city. however. the questions take the form of specific statements: "we're too small". "Our structure is too one sided". "we need more people to service". etc. Whether or not these observations are valid is debatable. for if a city is to get a particular share of the system's growth. it does so concurrently with other cities that also have a size. structure and population. The share obtained is a proportion relative to the other cities in the system. Only knowing this proportion is it possible to say whether the actual change. even if positive. is greater or less than what it could have been. how it compares with other cities. and how it compares with the nation. Indeed. it is then that the influence of size. industrial composition and population can be examined. 5 A shift technique. which is explained in Chapter II. is used to obtain the relative changes. The shift, by which growth or decline is measured here. is the relative gain or loss in a city's employment as compared with national employment. Through this approach, the cause of the relative changes can be isolated, permitting an analysis of the significance of size. industrial composition and population. The hypotheses in the study are: (1) the shift technique is a useful device to indicate the cause of relative employment changes in cities. and (2) the relative shift in employment is independent of city size and industrial composition. Chapter II explains the shift technique and evaluates alterna- tive approaches in the analysis of employment changes. Chapter III considers the framework of the problem to be analyzed in the selected cities in Michigan and Indiana. and establishes the theoretical basis for the use of the city as an appropriate unit of analysis. The above hypotheses are examined in Chapters IV. V, VI. VII and VIII by deter- mining the relative employment shifts for total employment. for employa ment by sectors and by industries within the manufacturing sector. Relationships between the shifts and size. industrial composition and population are then developed through simpleregression analysis. The results of the analysis are elaborated in the concluding chapter. CHAPTER II APPROACHES TO PEEIONAL ECONONIC ANALYSIS The great concern with growth has led to a relatively rapid rise in the significance and development of analysis at the regional and subreaional levels of activity. In a recent survey article. the surge in importance of regional analysis is described as resulting from a desire for more adequate and analytically useful answers to the economic problems of regions and cities.1 The growth. or lack of growth. of economic activity has led to the search for the explanations of these changes that have differed widely among the various areas. The explanations are based on the functioning of a general equilibrium system. in which the prices are arrived at through a process of mutual determination. As the data change over time. the general system pro; duces new solutions resulting in changes in the volume and composition of economic activity in the different places. Modern regional analysis. in seeking explanations. has shifted emphasis away from the problems of business cycles to that of maxi- mization problems. i.e.. promoting growth not so much in terms of given current problems. but in terms of possibilities of growth not yet 1J. Meyer. "Regional Analysis: A Survey". Amprican Economig_ Review, Vol. LIII. March 1963, p. 20. 7 realized.2 This pragmatic approach has been inteorated in many ways with the developing tools of general economics: multiplier theory. input-output analysis. and/or mathematical programming and the re- furbished location theory.3 This integration has been necessary to link the mass of data and the formulation of programs for the many areas. In consequence. four broad approaches have been developed to organize data to explain changes in economic activity. Each is di- rected at organizing aggravate data for a given blocked out problem and uses various theoretical tools of general economic analysis. a. The EXport-Base Approach The export-base approach attempts to explain changes in growth in terms of the degree and influence of the export-base industries of a given city or region.“ The growth of the given city or region is initiated by the response of the industries within the unit to the in- crease in demand outside the unit. The result in the unit is an exp pension of economic activity through a multiplier process. The approach stresses the key role of exportable commodities and services. ~The rate of growth will depend on the rate at which the export base expands to meet the increased demand from outside the unit under analysis. Use 2%.. p. 26. 3Ibid.. p. 30. uD. C. Nbrth. "Location Theory and Regional Ebonomic Growth". Journal of Political Economy. vol. 63. June. 1955. R. B. Andrews. "mechanics of the Urban Base: Historical Development of the Base Concept”. Land Economics. Vol. 29. My. 1953; and subsequent articles by R. B. Andrews. land Economics. Vbl. 29. Aug. and Nov.. 1953; V01. 30. Feb.. May. Aug.. and Nov.. 1954; Vol. 31. Feb.. May. Aug. and Nov.. 1955; Vbl. 32. Feb]. 1956. here is of the 1 t'alier forecas- tinted analysi in such waste?! 01" .he and its clearly does no 8 here is made of the multiplier theory by observing the past activity of the unit relative to the rest of the economy. An empirical mul- tiplier is derived and applied to estimates of the economic base to forecast the volume of economic activity. The subsequent development of the approach uses more sophis- ticated multiplier theory. comparative cost techniques. input-output analysis and modern location theory in various forms and combinations.5 In such analysis. acceptance of the economic base is fundamental. The existence of sharp criticism. which attacks the estimate of the size of the economic base. the refinement of the theoretical structure. and its usefulness for planning purposes.6 still has not destroyed the clearly useful classification system the approach provides. While it does not provide a functional relationship of the internal and external aspects. this approach does bring to the study of regional growth the clear fact that a city or region's growth is tied to developments in the total economy. b. The Sectoral Approach The sectoral approach had its beginning in the empirical work of Clark and Fisher. who emphasize the broad sectors of primary. sec- ondary and tertiary activities.7 It focuses on internal development 5Meyer. pp. 913.. pp. 30-35. and literature cited there. A specific example of the combination and techniques is in F. T. Mbore and J. w. Petersen. "Regional Analysis: An Interaindustry Medal of Utah". Review of Economics and Statistics. Vbl. 37. November 1955. 6Meyer. 0 . cit. 7C. Clark. The Condition of Economic Progpesg. London. 9 through the evolving specialization and degree of functional differen- tiation of inputs and less on external shifts in demand. The result- ing explanation of growth is seen in the dynamic progress of the shift- ing sectors. When the approach is coupled with location theory. it is used to explain the development stages of a nation or a region.8 The approach. and its combination with development stages. is useful be. cause of the implications that planning policies and activities can determine the rate of growth and its movement between stages. It is also useful as a framework for aggregation of data and has provided a means of extending the analysis of regions by considering the re- lationship between the growth of the region and the existence of "growth industries" within the unit.9 This approach is limited, how; ever. to those problems where a high level of aggregation is desirable and the lack of external functional ties in the internal evolution is insignificant. c. An Evaluation of Approaches The third approach. mathematical programming. provides the 0 best conceptual device for the aggregation of data:L However. the Ehcmillan. 19h0; A. G. B. Fisher. "Capital and the Growth of Knowa ledge". Economic Journal. V61. #3. September 1933 and "Production. Primary. Secondary and Tertiary". Economic Record, V01. 15. June 1939. 8A. Iosch. "The Nature of Economic Regions". Southern Ebo- nomic Journal. Vol. 5. July 1939; E. M. Hoover. The location of Economic Activity) New Ybrk. NbGraWAHill. 1948. pp. 187-196. 9U.S. Department of Commerce. Regional Trends in the U.S. Eggnogy A Supplement to the Survey of Current Business. Washington. Government Printing Office. 1951. lONeyer.‘gp. cit.. p. 36. 9‘4“. Q... :1. 10 approach is plagued by a lack of the data required and the sheer imp mensity of the inputs of the model to explain economic activity. This approach. as well as the two above, has limited usefulness due to the high degree of aggrezation on the one hand. and its limited scope on the other.11 How each should he applied to a given problem of a re- gion, area or citv, is not a rhetorical question or easily resolved. To planners of growth. an approach limited in scope to the internal is relatively useless since the external is also important. The export- base approach generally makes no attempt to specify where the base ex; port goes. or the places from which the inputs of the basic industries are drawn. Assuming the unit versus the rest of the world minimizes the internal structure and develops no functional relationships. This is similar to the sectoral approach that concentrates on the internal evolving specialization and division of labor. Although both ap- proaches allude to internal-external changes. both preclude by assump- tion any interacting tie between units such that their explanations are more absolute than relative. The programming device is sileY too vast and costly to be generally used. The problems associated with the application of an approach. and the extent of its specific utilization. are indicated by the view- points held about analysis in a given context. These viewpoints de- scribe analysis as either an analytically rigorous quantified model at a high level of aggregation (useful for forecasting). or a means of viewing changes in terms of historical and behavioral characteristics.12 112219;. 396 sources cited there. 123':eyerg 92. 9E0, p. 38. n\~\‘ I: Mn... 0 I ‘ "Urn... n .V. 11 There is much blending in these viewpoints that is no doubt a de- sirable evolution toward an interration conceptually. quantitatively and operationally useful.13 However, differences do exist and arise in relationship to the particular problems examined.1u' d. The Shift Technique The rise in concern for maximizing the possible opportuni- ties makes the shift technique quite useful. It combines the in- ternal and external forces operative on a city or region in a sys- tem of cities or revions. The particular unit is gaining or losing relative to other units as its internal characteristics are conducive to a positive or negative change. This approach is a way to describe the redistribution that would have occurred had the unit grown at the U. 3. rate. It permits an analysis of the relative changes in the unit relative to other units, and not as an isolated unit versus the rest of the world. This technique allows the internal-external forces to operate and measures the relative growth or decline. Several recent studies have made use of the shift technique at a high level of aggre- gation.15 However, at this higher level of aggregation, the relevant economic unit is the city. This has led to conflicting explanations 13W. Isard, Methods of Regional Analysis. New York, Wiley 1960; p. 570. 1”For example: R. M. Lichtenberg, One-Tenth of a Nation, Cambridge. Harvard University Press, 1960 and E. Harris. P. J. Areas Systems. Penn-Jersey Study Paper No. 14. Philadelphia. 15V. R. Fuchs. Changes in the Locatipn of Manufacturing in the U.S. Since 1939, New Haven. Yale University Press 1962; H. S. Perloff, E. S. Dunn. E. E. Lampard, J. F. Muth. Regions, Resources and Economic Growth. Johns Hopkins Press. Baltimore 1960. Lichtenberg. 9p. git. Ix “HA no 12 that stem partly from the definitions used, but is more likely the re- sult of taking the highly aggregated unit as the relevant unit for analysis. The shift technique is the approach used in this study, but at the lower level of aggregation of cities where modern development has historically occurred. The application of the shift technique produces a measure, called the shift, whereby the gain or loss in a city's economic acti- vity is related to the nation's activity.l6 Ehployment data are used to indicate the activity. It is the nation's increase in employment then, that is the benchmark to which the actual city employment can be compared. For the period considered, the percentage increase in national employment is applied to the city's total employment in the initial year. This increase added to the initial'year's employment would give the expected employment in the terminal year. It is in this sense that the word expected is used throughout the analysis. If the city's employment did grow at the national rate, the expected employment is equal to the actual employment. Since the national growth is not in fact distributed uniformly throughout the nation. the actual employ. ment in a city can be greater or less than its expected employment. For example. Kalamazoo's actual employment in 1963 was 65,200. If employment had grown at the national rate. Kalamazoo's expected employ. ment in 1963 would have been 56,100. The difference between the actual and expected, 9,100, is the net shift for total employment. This 16Adapted from Perloff. et al., pp. 70-71. 13 measure is not a measure of any kind of physical movement. but a net shift concept determined after these changes have occurred. and can be positive or negative. The net shift of total employment is the result of two causes. First. the city is competing to attract industry regardless of what is happening nationally. The city may attract more or less employment in a given industry. Whether or not it does depends on the city's lo- cational advantages for the given industry. This is called the local- factor effect. Secondly. the industry growing in the nation may be located in the city. in which case the city's employment also grows. If many such growth industries are located in the city. this composi- tion is conducive to the growth of employment. This is called the composition effect. Applying the growthzate of each industry in the nation to each industry in the city. instead of total employment. a shift of employ- ment from the effects can be obtained. For example. applying the rate to each industry in the city of Kalamazoo. expected employment in 1963 would have totaled 57.600. This is less than the actual employment of 65,200. The difference. 7.600. is the shift from the local-factor effect and indicates that each industry in fact grew faster than it did in the whole nation. The difference between the shift of total employment. 9.100. and the local-factor shift. 7.600. is the shift due to the composition effect. The composition shift indicates that emp- Iloyment in the growth industries has exceeded the national average. These shifts are in terms of the national forces. the city's locational advantages. and the industry composition. and are relative to shifts in 1a other cities. The computation of the shifts is determined on the premise that the sum of the shifts from the composition and local-factor ef— fects equals the net total employment shift.17 It is possible for the shifts to be positive or negative. and the net shift of total employ; ment is made up of the combinations of the positive and negative local-factor and composition shifts. Either of the effects can pro- duce a shift that more than offsets the other. It is possible for each industry in a city to have a zero shift. i.e.. where the actual and expected employment are equal. but total employment can still shift because of the composition effect. The shift technique allows the relative changes to be traced to their sources and reveals patterns of employment among cities. The shifts are computed for the cities and grouped by states. 17 See Appendix I .1. i . vll a CHAPTER III A FFAWEHCPK FOR THE ANAIYSIS OF CITY GROJTH A metropolitan area is influenced by the national economy and the area's own particular characteristics. This interaction pro- duces different results over time in different places. The places are the cities where the forces may reinforce each other. or offset advantages and disadvantages. so that the dynamic changes may be faster or slower than the actual pace at the national level. The changes are distributed unevenly throughout the system. The cities are an integrated collection of places. It is on this basis of in- tegration. however. that the particular characteristics of a city interact. and obtain their share of national growth. a. Economic Activity and Employment Whenever there is an impact from the forces. the industry and economic activity in an area alter accordingly. The most direct indication of the change is reflected in the amount of employment.1 The volume of economic activity. and its changes. are here measured and analyzed in terms of employment. Alternative concepts which measure activity are value added and real income. The concern here 1V. R. Fuchs. Changes in the Location of Manufacturing in thg U.S. Since_l929. New Haven.—Iale UniverSiEnyress. 1962. p. 176. 15 16 is not with an analysis of welfare. so a measure of real income is not used. value added provides a significantly different result from emp- loyment only in certain kinds of analysis. and in most analvsis. parti- cularly at the state level. has been found to give the same results.2 The employment data shown in Table 1 indicate employment at various levels of aggregation for the states of Michigan and Indiana. Between 1950 and 1963. the percentage change, compared with the nation. varied with the grouping. Disaggrecatinq the prouped city data would show similar variation. This variation reflects the differential im- pact of the forces. The importance of the city unit to the state's TABLE 1 TOTAL EEPLOYVENT. 1950-1963 “— M J— “ J— 1 fl 1250 196} fi channe United States 59.651.700 67,617,000 13.4 Michigan 2.369.u00 2,652,700 11.9 Indiana. 192719200 1945“, 700 14.4 Michigan. minus large SMA* l.0#2.800 1.373.300 31.? Indiana. minus large SMA* 796.300 926.500 16.h Michigan. selected cities 664.300 8n3,500 27.0 Indiana. selected cities 766.300 812.500 6.1 Source: Appendix B *Standard Metropolitan Area total increase in employment is indicated by progressing to the lower level of agareqation. It is here that any changes or adjustments in M 2Fuchs, ibid.. pp. n3.u8, 176. When the finest detailed data was used. different results were obtained that were due to old vs new plants. and/or inventory changes. Ibid., pp. 76—77. 1? economic activity. whatever the source. have their impact on employment. In both states. the selected cities account for more than fifty percent of the total employment outside the very larse metropolitan areas. b. The Problem Since the city is a part of the whole system. it can gain or lose economic activity, and accordingly employment. as the system grows. The employment may rise because it gains a greater proportion of an activity (industry or sector) that is growing. or declining. nationally. Employment can also increase because the city has the locational ad- vantages favorable for industry. If both are indeed positive. a city's employment growth may be greater than the nation's. Certainly. if both were nesative. the city's employment yrowth would be less than the nation's. The employment growth in the cities could be expected to grow at the same rate as the nation's employment. However. the ex- pected employment wrowth may not occur. Conseouently. if the national rate is applied to the city's actual employment in the beginning period, the actual employment in the terminal period can be greater or less than the expected. This is summarized in Figure 1. Over the period of time, the impact of the broad forces may have a favorable or unfavorable impact on the city's employment. This would place the city's employment in the terminal year at point a. b, or c. in Figure 1. Actual employment that is between the points a and c may, however. present employment problems for the city. It may lose employment. or gain. but at a slower rate than desired. If the data of Table l are indicative, a breakdown of the data to the city 18 level would show the variability of employment changes falling some- where between points a and c. Figure 1 Actual and Expected Employment Over Time Total Employ— ment ,Actual ’Expected a ’ . Actual 1950 1963 Time The fact that the city's employment is between points a and c. means some cities obtained a greater share of the nation's employ- ment growth than their expected employment. This is a redistribution of economic activity and employment among the cities in the system. The extent of the redistribution for a given city is the difference between the actual and expected employment for the period. It is the redistribution that is the shift and expresses the relative partici- pation among the cities of the nation's growth in employment. The source of the shift will be indicated through the use of the shift technique. Is the shift affected by city size and industrial composition? In Figure 2, the relationship between total employment and 19 Figure 2 Possible Emplgyment by Size and Structure A.‘ a ' I Total Employ- ,’ men t: I City Size and Composition size and composition is indicated by curve A. Total employment can increase or decrease for each city size and composition as a result of the shift. That is. curve A can be displaced upward or downward and its slope increased or decreased. If the shift had a positive effect on total employment but was independent of size and composition. curve A' is the appropriate relationship indicating that cities of all sizes and composition increased total employment. However. if size and composition significantly affect the shift. curve A" is the appropriate relationship. indicating the total employment rises according to city size and composition. The present analysis shows that the shift is not significantly affected by the city size and com. position; consequently. curve A' is the appropriate relationship. Standard simple regression analysis is used in the development uni sur ‘1. VOL- “car no? 240,1 ‘ .L 20 of relationships and is applied to the data for all cities in both states and selected cities grouped by states. The data are non- random and the regression is used to indicate directions and impli- cations in the interpretation of patterns. The regression and corre— lation are not used here to establish the probabilities of the nearness of coefficients to parameters in the universe.3 There may be no normal universe of cities. and certainly the diversity of growth among cities suggests this. in which case representation and indeed. randomness. are irrelevant. The period covered is 1949-1963. In setting up the data for this period. a three year average was computed for the initial and terminal years. That is. the initial year. called 1950. consists of an average of 1949, 1950 and 1951; the terminal year. 1963. an average of 1961. 1962 and 1963. This was done to overcome in part the initial- terminal method of measuring that produces variation by the specific year chosen. 0. A System of Cities The pertinent economic unit here is the city. However. these cities perform relative to other cities in a complex system that is not completely understood. There is an integrated logic of cities in 3The significance of the correlation in the following chapters is computed at the five percent level using the z transformation for- mula. R. Ferber. Statistical Techniques in Market Research. MbGrawa Hill Book Company. New York. 19U9. p. 381. The significance test is only an indication of the reliability of the correlation. since the data used are not random. I3" 53H 21 space according to functions performed. This can be visualized con- ceptually by placing the cities within a spatial structure. losch has developed such a scheme where he views the towns or cities as "punctiform asglomerations of non-agricultural locations".h They are the physical clusters of activity that arise because of chance. the site of a large industry or the source of a raw material that result in advantages of locating. The towns or clusters are linked together by lines of transport and communications producing a viable interaction in a system of towns. The towns differ in size depending on the different collections of industries and the comparative economies and diseconomies of scale. localization and urbanization. The simple market surrounding the town is the supply area for the functions performed in the town. Each town in turn is related to the functions of the next larger size town. So visualized. the scheme becomes a hierarchy of central places at the focal points of the various levels by size and differentiation of functions. with a concomitant population. The entire collection of these places forms an "ideal type" of economic landscape consisting of simple market regions surrounding each center of production and consumption in a network of markets comprising a regional system.5 Within this system. common centers arise. metropolitan areas. that serve a vast hinterland of smaller size cities and compete with “A. Iosch. The fgprgmic§_9§;;ocatipp, vale University Press, New Haven. 1954, p. 6g: Forfa—mathematihal model. see M. J. Backman, "City Hierarchies and the Distribution of Size". EEBEPHQSLEEEEQ£ITF¥EK and Culturil Change. Vol. VI. No. 3. 1958. SIOSChv 393%., pp. 1214-1370 22 other centers. The size and the functions performed in the outlying environs of a center are a function of distance. production advantages and competition. The scheme is an ideal landscape of the location of points. "punctiform agglomerations". and an extended territory of complementary and reciprocallv related activities in an ordered hier- archv of function and size. Tn establishing this scheme. Losch has assumed an undiffer- entiated plain over which raw materials and soil fertility are evenly distributed.6 Although the system is conceptually useful, the com- plexity of the scheme is displayed when it is used to describe reality. For then the simple conditions must be altered to accept the fact of an irregular topography and an uneven distribution of resources. These natural differences. in addition to transport cost policies and regional differences in skills.7 lead to differences in economies and diseconomies of agglomeration with the subsequent effect of giving irregular shapes to areas in the hierarchy. Indeed. this effect also alters the location-function scheme because of the differences of raw material. productivity and accessibility. The entire scheme is in- fluenced by the government. in varying degrees. for administrative purposes and/or control and intervention.8 The result is the estab— lishment of boundaries that may or may not coincide with the logic of the areas within the economic landscape. 61.05Ch9 9P9 92:1" p. 105. 7Iosch. 92. 933., pp. 139-193. 8Losch. 22. git,, p. 130 and pp. 196—210. 23 The Ioschian scheme not only conceptualizes a system of cities. but also indicates the relative tie between the city and its surround— ing area. The functioning of the unit links the city to the system through its particular size. industrial composition and population. The use of the Loschian scheme. however. must be adjusted according to the form in which data are available. 'we have no Loschian statis- tics. just raw data by states. counties and metropolitan areas. Thus. it is central city and the surrounding area that is defined as the city unit used here. The data for this unit are the data for the county or counties that form the metropolitan area. It is only in this form in which the data are available. However. this unit is not to be presumed a unit determined objectively by a unique set of criteria. A boundary for a city could be delimited by alternative procedures. for example. central city. trade areas. newspaper circulation. Federal Reserve districts. etc. Thus. any unit is actually chosen by an arbi- trary decision related to the problem at hand. The unit used herehas a conceptual basis and is statistically dictated by the data available. An analyst has argued that the state may be a more appro- priate unit for analysis rather than the city or metropolitan area.9 This argument is an attempt to replace an arbitrary approach in the selection of a unit with one based on objective criteria. Since the city and the surrounding area can be split by the arbitrary boundary of the state. it should follow that the portion of the nulti-state 9V. R. Fuehs. "States or SMA's When Studying the Location of Manufacturing". Southern Economic Journal. Vbl. 25. January 1959. pp. 349-355. 24 city in one state would predict the rest of the area growth better than the state. He rejects this conclusion and argues that the state unit is a better predictor for analytical purposes. However. it is debatable whether his conclusion is objective. since the boundary for the multi-state city is also arbitrary. The geographical direction of growth depends on how the accidental split occurs. That is. a corner of a multi-state city area may grow faster than the rest of the city because the corner may expand in the direction of the empty space.10 This growth is not tied to the state unit but is part of the functional relationship of the surrounding area to the city. Indeed. the state is not an integrated entity but an arbitrary boundary accidentally en- compassing a heterogeneous collection of places of activity. A sharp change in the state's growth of activity could be obtained by exclud- ing such a place of activity. However. this need not affect the re- maining units since each performs according to its characteristics within a system.11 loLosch. pp. 912., pp. 204-205. 11R. Vining. "Delimitation of Economic Areas: Statistical Con- ceptions in the Study of the Spatial Structure of an Economic System". Journal of American Statistical Association, vol. “8. 1953. p. 52. CHAPTER IV EEPLOYKENT SHIFTS AND CITY SIZE In the analysis of total employment among cities. the actual changes are obviously important. However, a city's actual level of employment depends in part on the particular city's advantages rela- tive to other cities. Consequently. the actual changes do not pro- vide the additional insight that differential changes among cities provide through their relative change. Therefore, the analysis uses the terms of a city's actual and relative expected changes in total employment. a. Patterns of Employment Changes Table 1 (Chapter III) gave actual employment data at several levels of aggregation. Table 2 shows the actual and expected employ- ment disaggregated by the selected cities for both states. Fbr all cities. the actual increase in total employment is 225,000 over the 1950 average. The total employment shift is 166,300 (an absolute value). less than the actual increase in employment. The shift in- dicates the employment distributed differently than in 1950. On the average. 11.6 percent of the jobs in 1950 were redistributed by 1963. Viewing the data in Table 2 by states shows that the redis- tribution of employment exhibits different patterns. The actual in- crease in employment in Indiana is 6 percent of the employment in 2a 26 m Kapcoed< “condom c.HH oor.¢mHfl ooa.mae.H o.oH ooa.mmm ooo.mmm.H oom.oma.H moHpHo HHa maepoe pcmao m.m 0.0 H ooe.Hsfl ooH.owm H.w coa.m¢ cow.mHm oom.pms mHuHoe .enH m.Ha m.s oom.m+ oom.om 0. am OOH.¢+ ooa.mm ocm.e« paange N.c o.m ooa.H- cow.mm H. s oom.H+ OOH.em oom.m~ .osaon H.m N.N oom.an oom.wm N. o oo:.&+ ooa.dm oou.wm opaom .A o.oH m.HH oom.m- oco.oe m. m ooa.mu cow.oa oom.ma oasam .H 0.0H m.m oom.H+ oom.oH H. em OOH.:+ 00H.Hm ooo.sH :oHsas w.om o.me oom.om- ooe.mHH m. 0? oom.mHu ooo.ww oom.NOH ocom .m o.s e.m ooe.mu ooo.me m. m ooe.m+ ooa.mm ooo.sm .oaooca 0.0H w.m 000.3: 00:.m: m. m ood.H+ ood.Hd 000.0: oHocsz H.o m.o com- OOH.on m. mm oom.mm+ oom.mmm oom.omm .mHoocH m.m s.m oom.Hu cos.oo mH oom.o+ oow.so oom.sm ones: .e e.sH o.mH oom.mHu ooo.sm N. am oom.mu cow.ms oom.ms .mcapm N.aH o.c0H 000.39“ oom.mms 0. am OOH.osH 00m.mow oom.awe mHmpoe .aoaz .w.mw. o.sm OOH.em+ oom.m m. m ooa.mm ooe.os 000.5: .osa cc< s.HH o.“ oom.H+ oom.sm o .om coo.m ooo.om ooo.em .asml.i m.o m.o oom- oom.ae m. NH OOH.s oom.mw ooe.mm .cham s.o m.o oom+ coo.Hm 0. 3H oom.m oom.Hm ooo.ma .xoaa m.mm m.:m ooo.mm+ ooo.om o. we ooa.mm oom.moH oom.os .ocmH e.mH s.m oom.¢+ ooH.em s. Hm cos.mH oom.mm oom.me .aHas m.m o.m coo.H+ ooo.ma s. wH oom.o oom.me ooo.sm .oxoae m.m s.e ooa.¢+ conummH o. oH cos.Hm oos.mmH ooo.:HH .oam .ao s.s o.o cow.m+ oom.mNH H. Ha oom.mm oom.mmH cos. OHH chHa s.ma s.o oom.¢+ OOH.ea 0. am ooa.eH oom.mm coo. mm .pam .com c.a m.H oom.H- oom.om u. s OOH.N oom.ow ooo. on >pHo .m a. .2 Es... Ema. .....e.,....H..:.......wc.a.....is.,......p..::sembzztsts ago a ft... 2:6 1 ..im .....q.+oe i Fm +soscrc Hora...» a. error? rmcr cmo... cm H co: 8 tone _+:o§oHc§ H 3.8. -e!wm.V we emcewmmHu;--zx.agi 144111-;-:1sa:-zzsiaoe:::»--zzzlal aetzl memHuommH «pHHQzH and :HchoHs .eaHzm sea cad samz>oamzm eases N @Hm¢e 27 1950. but 9.3 percent of 1950's jobs were redistributed in 1963. In Michigan. the actual increase in employment is 27 percent. with 1h.2 percent of 1950's jobs redistributed. If the redistribution of emp- loyment is considered in terms of the direction of the shift. the sig- nificance of the pattern becomes clear. In Nichigan. 98 percent of the shift is positive; Indiana's is 90 percent negative. The cities in Michigan. in terms of employment grew faster than the nation and In- diana's slower. The cities included here are the major cities (axe cluding the large standard metropolitan areas) in each state. These cities were a drag on Indiana's growth. In fact. Indiana's actual increase in employment of lh.h percent (Table 1) must have come from the smaller cities not included in Table 2.1 The differences exhibited by the data in Table 2 display dissimilar patterns: the cities in Indiana realized a relative loss from the impact of change; Michigan's a relative gain. b. Population and Employment Shifts Do the changes in population fit the same pattern? All ci- ties in both states had positive increases in population and in sum exceeded the each state's increase.2 Regressing total employment and population for the cities shows a high degree of correlation. as shown in Table 3.3 However. changes in total employment and changes in lIndicative of this is the fact that the smallest size cities in Table 2 had the largest increase in total employment. 2AppendixA. 3Linear form Y = A + BX was used with population as the inde- pendent variable. 28 population are not related. The changes in population do not account TABLE 3 CORRELATION COEFFICIENTS OF TOTAL EMPLOYNENT AID POPULATION FOR SELECTED CITIES IN MICHIGAN AND IEDIAKA. 1950-63 r* r2 Michigan cities .960 .922 Indiana cities .996 .992 All cities (pooled) .989 .978 *Significant at the 5% level for the change in total employment. as shown in Table h. TABLE ’4» CORRELATION COEFFICIENTS OF CHANGES IN TOTAL EMPLOYWENT AND POPULATION FOE SELECTED CITIES IN MICHIGAN AND INDIANA, 1950563 r* r2 Michigan cities .532 .283 Indiana cities .828 .686 All cities (pooled) .700 .490 *Not significant at the 5% level The population of the cities is by definition the size of the city. Each city had a net shift in total employment and this shift is independent of the actual changes in population (city size). as indicated 29 in Table 5. TABLE 5 CORRELATION COEFFICIENTS OF TOTAL EMPLOYMENT SHIFT AND CHANGE IN POPULATION FOR SELECTED CITIES IN MICHIGAN ND INDIANA. 1950-63 r* r2 Michiqan cities .209 .023 Indiana cities .123 .015 All cities (pooled) .201 .040 *Not significant at the 5% level If the cities' population had grown at the same rate as the population in the nation. each city would have had a negative or posi- tive shift in population. similarly derived as the employment shift.)+ Shifts in total employment and population are not highly correlated. The pattern of population shifts does not match the pattern of total employment shifts when the data is disaggregated at the city level. This is contrary to the findines when the total data on employment and population shifts at the state level are used.5 The city is an independent economic unit among many such units and its reaction need uPopulation shift data is in Appendix A. 5H. S. Perloff. E. S. Dunn. E. E. Lampard, J. F. Enth. PeoionS. Resources and Economic Growth. John Hopkins Press. Baltimore 1960. p. 296. The unit of analysis in this study is the state. Accordinclv the employment and population shifts so derived do not consider the variation within the state. 30 not follow the general pattern of the state. In fact. none of the components of the total employment shift. the local—factor and com- position shifts. are hiohlv correlated with the population shift.6 This is shown in Table 6. The fact that it is not correlated to any high degree simply indicates the fact that the cities react inde- pendently of population shifts. not only in terms of total employ— ment shifts. but also in terms of shifts due to the industry com- position and location advantades. TABLE 6 COEEELATION COEFFICIENTS OF ENPIOYNENT AND POPUIATION SHIFTS FOE SELECTED CITIES IN MICHIGAN AND INDIANA. 1950-63 __F* r? Tot. Emp1.: Population Michigan cities .190 .036 Indiana cities .413 .171 All cities (pooled) .395 .156 Loc.-Factor: Population Eichigan cities .226 .051 Indiana cities .112 .013 All cities (pooled) .275 .076 Composition: Population Michigan cities -.l42 .020 Indiana cities .816 .666 All cities (pooled) .465 .216 *Not significant at the 5% level 6 If the population shifts are used to measure changes in .7 It! ,‘ IVI'I\ 31 “emcee «saw spec Amecmnsonp n.3,H m .. a . an as a me , mm; Men («mmm mam mum m m M o No..nmm . u as. up .. . o: . om ONH oea . I . cow. .; w,“ . r _ if. I seem Nwao use escsaeadsm fleece .e. Anomav. 1' h chem . .acoshoaasm w m . as dupes m . 0mm:e;i .:F Aommav seam seas mam . as men he was as m: 1‘ «0. H0. mmwpao HH¢ No. no. mowpflo .ucH mo. mm. mafiaao .nowz new e sewases new, 0 Hawcwwpca ”oedema hp seam Nuao use escshcadsm fleece s edemae mNH Ampcwmponv adv n~ es 00 00H 33 The population (city size) is definitely related to total employment for no other reason than that the possible employment is bounded by the population size. According to the definition of the city unit used here (county or counties that form the area). commut- ing workers do not sianificantly alter the size of the available work force. Commuters would pose a difficulty if the city and its population were defined by a oeoaraphical "ring". of say ten miles. about a central core. In that case. however. the city unit is de- limited by alternative techniques. Fiyure 3 shows that a one percent rise in population (city size) is associated with a 60 percent rise in total employment. When the data are grouped by states. the re- gression is twisted up or down from the pooled regression. indicat- ing some variability between the states (Figure 4). The reeression is a good fit with very little curvature and a small error (2%)of the slope. However. the data are of total employment and population and are static. What is important over time is the change and redistri- bution. Tables 4. 5. and 6 indicated that changes in population do not account for the change in total employment. or. that it is de- pendent on the changes in city size. In terms of the employment shift. change in size had no sianificant impact. To obtain addition- al information on changes and redistributions. the total employment concept must be broken down into its components. This is done in the next chapter by industry sectors. demand. then change in demand is not significant for the shifts. CHAPTER V IETDUSTRY ETZPIOYl-‘E‘IT AND CITY SIZE The shifts that have occurred over the period worked on in- dividual industries within some structure and list of advantages. Prior to the analysis of these shifts. it is useful to see what the industry patterns are in terms of actual changes by sectors. The analysis of patterns will provide an insight into the stability and distribution of employment over the period. and also indicate the comparative usefulness of the shift analysis. The employment pat- terns by sectors for the initial and terminal years show several interesting facets. This chapter is concerned with the analysis of the patterns. relative to city size of employment by industry sector. a. Sector Employment Patterns Table 7 gives employment data by sectors for all cities se- lected in the two states. The actual employment in 1963 is distri- buted over the sectors as was that of 1950. The growth by sectors over this period is stable and according to the existing relative pattern of industry. This is indicated graphically in Figure 5 for all cities. The diversification of industry is quite consistent over the period. However. the absolute change as a percent of 1950 employment 34 35 .m Kflpcodd< "condom .pomoaomwpcs mpmp nonpo Use meacfls mopSHoQH* o.eoa m.ma oos.mmm o.ooa ooo.pme.a o.ooa ooe.ome.a masses s.m.. m.m- ooa.m- N.oa oom.mca N.NH oed.esa ..ecda mo. m.o oed.H e.mm ooo.dme m.se ooa.sme meannessedssa c.we m.mma ooe.doa m.HH ooa.mda 0.0 oom.mm pccscec>eo e.sm c.mc oom.ao e.d ooo.dma w.c oom.sm eeea>sem e.o s.mm ooe.am c.m ooe.oe s.m oem.mm .esH .caa o.d 6.5 oom.om m.sa oom.emm e.ma oes.oem Haeecmuaes: m.m e.d oem.s s.m oee.sw e.m oom.es .seo .essse d.m m.sa eed w e.m ooe.dm e.m oom.om ceasesseeseo session museso mm mesa mm ommH Hmpoeunm v Aavwfimv opsaowpa pacemOngm proe sopoom A V Amy Amp flew Ame Amv AHV m mamoa em one as pcmoaeacsam pen. mo. 0mm. mam. mo. saw. mes. mo. mmo. mom. mpuomme .mecago asam “ - mo. mom. mmm. «0. 0mm. Hmm. foo. mooo. bHo.+ Omwa .oswm sfiwo “ mesm :owpwmoczoo «H.- mow. a::.. me.- mam. em¢.- mo.- sea. mmm.- . mmuomme .omcmno swam “ mo; mom. amp... co... cam. 3K... .5... 03.. H3... 0mg .33 .58 u fidfi sopowmufiquH ec.- mac. oam.- oe.- ems. mme.- Ho.- moo. smo.- mp-omoe-amsem .oacapo " do.l cam. smm.l #0.! «mm. mflw.l Ho.n mOfl. wmr.| OmOH .oswm bpfio “pmwxm scwthowmscez L ms *s L «a *s ....fl ms *s III 41 mosses flea «caesmwiz - categoma mwOHlommH .¢Z+Jc A) note trkor«1~014 FJFJ ‘ OD CC.) 53.300 ‘ l (lag-L ‘- ' » . '19” I 96 . - 1 Aenendix 3 (Continuee) A. giver.) 1953 (.Mrov.‘ 1054? Moni‘nrl ”NF- Sin—'1‘”? ~-" 41 --, 'z -"_“. "- "'“ "‘ -- V Cit153ector h".- m cm; 1. '1 a .419“. LuV—L..- ..I *- Construction 3,300 2,600 4,000 -1,400 Trans.-Com. 4,900 4,400 4,600 -200 ‘fh01.-Reta11 13,000 14,500 16,000 -1,500 Fin.-1ns. 1,600 2,500 2,300 200 Service 6,500 9,600 9,500 -000 Government 0,600 6,000 6,900 -900 Ehnufacturin; 32,300 24,200 37,300 -13,100 fisco1laneous 9,300 10,700 7,900 2,900 Tota1 76,700 73,r90 Q7,000 13,500 Fort Yavne Construction 3,700 4,100 4,500 —400 Trans.-Com. 6,700 6,700 6,000 300 Thol.-Retail 15.600 19.400 19,100 1,300 Fin.-Ins. 2,600 4,800 3,900 1,000 Service 5,900 9,700 8,500 1,200 Government 4,700 7,100 7,100 0 Eanufacturing 37,800 35,700 43,000 7,300 Miscellaneous 11,000 10,300 9,200 1,100 Total 87,900 97,800 9,700 -1,900 ggfiienapolis Construction 12,100 13,600 14,800 -1,200 Trans.-Com. 23,800 21,500 22,800 -1,300 Hho1.-Retai1 61.100 66,900 70,800 —3,900 Fin.-Ihso 13,200 20,700 19,200 1.500 Service 20,300 32,000 35,600 -3,200 Government 25,700 43,700 38,300 4,900 Manufacturing 103,000 101,100 117,100 -16,000 Eiscellaneous 27,000 29,000 2,500 6,500 Total 290,200 328,900 329.100 -200 97 Appendix 8 (Continued) W City-Sector (Aver.) 1950 (Aver.) 1963 Expected Not Shift 0mncie “I. Construction 900 1,300 1,100 200 Trans.-Com. 1,800 2,200 1,700 500 Hhol.-Rctai1 5,500 7,000 6,400 600 Fin.-1ns. 700 1,200 1,000 200 Service 3,100 3,500 4,500 -1,000 Government 3,000 4,900 4,500 400 Bhnufecturing 19,200 16,100 20,700 -4,600 Iiscellanoous 6,000 5,200 5,700 -500 Total 40,050 41,400 45,400 ~4,000 Anflerson Construction 600 800 700 100 Trans.-Com. 800 900 800 100 fho1.-Retai1 6,300 5,600 7.300 -1,700 Fin.—Ins. 600 000 900 ~100 Service 1,900 2,000 2,800 -800 Government 1,000 1,600 1,500 100 Kanufacturing 23,900 25,000 27,200 -1,400 Liscellaneous 1,900 1,900 1,600 300 Total 37,000 39,400 42,000 -2,600 South Send Construction 3,300 2,300 4,000 -1,?00 Trans.-Com. 5,600 3,300 5,400 -l,600 Sh01.-Retai1 15,100 15,800 17,500 ~1,700 Fin.-Ins. 2,300 4,200 3,400 800 Service 8,400 11,500 11,500 0 Government 4,200 6,400 6,300 100 Eanufacturing 53,400 34,400 60,700 -26,300 hiscellaneous 10,500 7,100 8,800 -1,700 Total 102,800 86,000 116,600 ~30,600 Appendix B (Continued) _—v_-_ 1 fl CitzgSector (Aver.) 1950 Aver.) 1953 Exoected Net Shift Ferion Construction 400 500 500 0 Trans.-Com. 1,400 700 1,300 ~600 Wh01.-Retail 3,100 3,000 3,600 -600 Fin.-Ins. 500 500 700 -200 Service 900 1,200 1,300 -100 Government 1,000 1,600 1,500 100 Phnufacturing 9,000 13,500 10,200 3,400 hiscellaneous 700 0 600 -600 Total 17,000 21,100 19,300 1,000 Terre Feute Construction 1,400 1,400 1,700 -300 Trans.-Com. 4,900 3,600 4,600 -1,000 Thol.-Retai1 9,000 9,200 9.300 -100 Fin.-Ins. 900 1,300 1,300 0 Service 3,900 4,200 5,700 -1,500 Government 3,500 4,400 5,300 -900 Kanufacturing 11,200 10,200 12,700 -2,500 Liscellaneous 9,500 6,500 7,900 -1,400 IaPorte Construction 1,200 900 1,500 -600 Trans.-Com. 1,100 1,000 1,100 -100 Whol.-Retail 4,300 4,200 5,000 -800 Fin.-Ins. 400 600 600 0 Service 1,400 2,300 2,100 +200 Government 1,000 1,000 1,500 -500 Hanufacturin; 12,600 13,500 14,300 -800 Riscellaneous 700 600 600 0 Total 22,700 24,100 25,700 -1,600 99 Ancendix 3 (Continued) C+*-Sector LI 111 Richncnd Construction Trans.-Com. Tho1.--_etai1 Fin.-Ins. Service Government Kanufacturin; liscellencous Total T1 1,}, n ”t u‘h -- ~I __ ». A—~\.'v Construction Trans.-Con. Thol.-Retai1 Fin. -1n3 . Ser‘fi 3e Novel)“. “Pit W1— }firufficcll -_‘ Ii scellane ous Total. Construction Trans.-Con. Wholesale Pet?:1 We ‘-Pcturinc ..‘Lsce1121eonc (A5. (fiver.} 19 1,000 600 14,1300 500 'l , 5,00 --o WK, n.200 (\ ‘ ‘ (DO {1) U1 HNJNC‘M L‘Oxow O to h.)\)‘\)'\) nowogoo K oooooo , 0 ,\0 .92 , 0 .407, 0 920,000 6,090,300 14,73r,000 )1u,113,100 n: J O 20 2,10 1,20 22,70 ‘ . Q ‘0 § . -?~Q:%ko\o f‘ {6 o \o o\ H \o 13v9\p\oc3\0 \) N (DKOWN (Okn O\O CO 16 i 727,3 a; 11,990,..3 67,617,000 OO(DO('D 00000 1,000 600 5.900 700 1,900 1,500 13.900 000 I 25.500 Z\) N I l—‘F—JKHN.‘_ ',__J K») 4rr4>4 O O O O O O D O O (3‘\'\) COKJUI \OWKRKAJ U1 (“AK/O O‘x’J O\\n Ul“\} {TN ...... -300 200 -1,000 200 300 -300 -700 200 -1,400 -300 000 -200 100 -100 -300 5,000 -100 5,600 O [\— ‘— 3 on de urcc: Inoian. s militv Division ,.am1f'rr-‘ct1::‘e1 3 ‘~':. (van ‘9’,— 312131.037‘7181113 100 :‘ DDT v *“~.IV C it- A ...a-..'.J 4'; "A “"711, n:n‘rv-Qw-rn 7.137 C‘.”-","'- rm 7v o-.H~l -. DIGIT SIC F02 SELECTJD CITIES. 1930—1363 1330 1963 ,ggcted Yet Shift L,MOO 7'300 79500 ~200 1,500 1,500 2,000 -300 50 600 000 200 10 100 100 0 1,100 1,500 1,100 000 3,190 2.700 3,600 -900 3,700 6,400 3,900 2,600 600 700 1,100 —400 3,300 700 0,000 .3,300 2,600 700 2,500 -1,000 100 1.300 100 1,200 72.43 117.99 :av City 1,100 1.000 1,300 -300 000 000 000 O 600 200 500 -300 600 100 700 —600 200 200 200 O 500 500 500 0 1,100 900 1,100 -200 200 000 200 200 1,200 700 1,200 -500 930 800 1,700 -900 0,000 0,000 5.300 —1.300 0,200 3,200 0,000 -000 100 300 100 200 65.32 115.56 Appendix C (Continued) 101 C1§y_STC 1950 1933 Expocfcd Vet 3hift Eerie” Harbor 2' 700 1,500 900 700 22-3 300 100 300 100 20 200 000 200 200 25 300 100 300 -200 26 1,500 2.000 1.900 100 27 E00 800 600 200 3 3,000 3,900 3,000 500 30 1.000 1,500 1.200 000 35 5,000 3,000 5,100 -2.100 36 0,600 6,000 8.700 -2,300 37 2,900 2.700 3,000 -700 371 2,000 2,600 2,700 -100 Eisc. 100 1,700 100 1,60 E/‘k 07.1? 9; 99 Flint 20 900 1,300 1.000 300 22-3 000 300 000 -100 25 100 100 100 0 26 100 200 100 100 27 700 900 600 200 29-9 500 600 500 100 33 300 300 300 0 30 8,500 9,200 .800 -600 35 200 600 £00 0 36 100 100 100 0 37 50,000 55,900 60,700 -0,900 371 50,000 55,900 00,300 7,500 Rise 200 400 200 200 E/rk 78.63 100.50 Grand Espids 20 3,200 3,500 3,700 ~200 22-3 1.400 3,000 1.300 1,700 20 1.600 1.000 1.200 -200 25 8,800 7,300 10,000 -?,700 26 1,500 1,000 1.100 300 102 Appendix C (Continued) 0111.310 1950 1963* “$0 cted Net Shift p (D .‘_ 22-3 28-9 33 30 3% 3f.) ’3 J 371 \'0 1.4...JCO (D {\JPO‘AQ 000000 OOCDCJO 600 000 200 200 600 2,200 2.100 900 5.100 5.000 2.200 76.66 800 000 100 9.100 700 2,800 000 2.100 1.100 000 2,50 2,300 1.900 70.78 .100 500 2,100 12,000 6,700 2,300 700 000 7,600 c'\) 109.71 600 000 300 200 800 1.700 2,800 1.600 0,900 0,000 2.100 119.15 1.700 000 100 8,500 1.100 3,000 000 600 2.000 600 3,000 2,100 3.300 110.01 1.500 600 1,000 15,200 9.900 2,500 800 500 3.800 700 000 200 200 600 2,500 2,100 1.500 6.100 0,800 2,500 900 000 100 11.500 500 3.000 000 2,000 1.100 800 3,000 2,200 2.100 600 -100 .70 -3,200 -3.200 -200 -100 -100 3,800 ~100 100 200 -800 ~700 100 -1.200 -000 —000 Appendix C (Continued) 01011310 Exgected Net Shi 10.113121? 20 900 1.300 1.200 100 22-3 100 100 100 O 20 100 200 100 100 27 900 1.100 700 000 23-9 200 000 200 200 33 1.900 2,200 1.900 300 30 700 1,300 900 500 35 1,500 2,000 1.500 900 37 21.200 19.700 25,500 —6,300 371 21.200 19.700 20.300 -1,600 Kisc. 000 600 000 200 E/Tk 73.09 102.30 Ih$k0“on 20 000 600 500 100 25 2.200 1.500 2,000 —900 26 600 1.000 800 200 27 200 200 200 0 23-9 200 1,200 200 1.000 3 5,200 5,300 5.200 100 30 600 800 700 100 5 7.600 6.100 7,900 -1,700 30 000 1.000 800 200 37 5,600 0,300 5,700 —1.000 371 5.000 0,200 5,000 -1,200 Tisc. 1.600 2,300 1,900 500 E/fk 73.25 115.09 3"}‘1113 J 20 1.900 2.100 2.100 0 22-3 100 100 100 0 20 300 100 200 -100 25 700 600 P00 -200 26 200 100 300 -200 27 200 000 200 200 29-9 300 300 300 0 33 9,200 7,600 9,200 -1,300 «(‘4- wiJ 100 Pigy 3*0 1950 19633 Exgocted Vet Shift S271n0w (:00f1“fl0d) 30 600 700 7.0 0 35 3.100 3.900 3,700 700 36 700 Q00 1.300 -500 37 6,400 7, 00 0,000 -700 371 6,600 7.300 6,30' 1,000 1isc. 900 200 900 -700 E/fv 73.19 133.03 P01t '*Ifl: 20 300 000 300 100 22-3 300 300 300 0 25 700 600 900 -300 27 200 200 200 0 29-9 700 700 800 -100 30 200 300 200 100 3' 1.500 90 1,500 -600 37 2,200 700 2,700 -2,000 371 1.500 200 1.000 -1,200 rise. 2.600 600 2.900 -2.300 s/rk 70.79 103.30 Ann Arbor 20 300 500 300 200 20 100 100 100 0 27 600 800 500 300 33 600 700 600 100 30 900 1,200 1.000 200 35 2,100 0,600 2.100 2,500 36 2.100 3.100 0.000 -900 37 8.90 11.200 9.700 1.500 371 3.900 11.2.0 9,500 2.70 Lise. 0,700 2.500 5,200 -2.700 E/fk 73.07 135.32 105 0117-310 1950 1963. :EPected wot Shift Ev20svil1e 20 0,30. 3,000 5,000 -2.000 22-3 1,200 1.500 1.100 000 20 1,100 900 900 00 25 2.100 2.100 2.000 -30 / , ‘ 2C—7 1,0C0 1.000 1,000 -400 O 2- 1.300 1.000 1.100 -100 or ’ _ . C , o " I 27 0.3?“ 3,200 5.200 -g,000 “isc. 900 200 300 -600 2/1: :2.95 110.01 (5,77,; rn'V-Ao ‘ Vd- L . L11 _/ [*3 \ I ‘3 r‘ I Indiaflapolis 20 22-3 20 25 26 27 20 33 30 35 0 J (DE) Eb<3 ‘0 .0meth knkJ‘xkokn\o ’_-J 0:0- 0 O O :0" L) CDC) (DCDfiDCD()(D(J(DCD 04. .% ”D 7; K- 12, 00 0,300 900 1,300 3.600 2.000 8,000 0,000 7.000 11.800 161200 3.500 1.50 2.200 1.500 2.100 3.500 10.900 0.600 1.000 127.72 14 031—" ’(Dm .m Mflvromgd "ooapom com: co: com: cow: com.m oom.H com: Jow.H 00d: 00H.: cow: 00:.H .Hggo .uop .QOHpHmogeoo 502% pmfinm pom 100.0m 00:.H com coo ooo.om 000.0 OOH.N oom.m oom.m ooa.m com- 000.: .Hggo .000 .aopommnHmoOH Eogm pmfigm pom ooa.mm ooq.fl com- com oom.mm 00fi.m oom.fi 000.0 oom.m oom.m oom.H- oom.m .Hgso .pop .pMHsm p02 gonad mopzm szwwwm CQHQXurL mmwmst oommxwmc: cemACcm mvrm-c wCHHh 009:1: pr0 $0000 "moflpwo I CqCHLo.: crc ppom 0:000 copcom hmm oprmm {Infil- lfill g E memH10mmH .mMHHHo omeomqmm Hm mmopozm em enwgeo.n1m no eeHam 01-. mo mommom Q NHQHHQL< .0 wwmrommu "maprom ), . I 1.1.- .p\.\.' ;\ ~ \I \h. \II1 u\ .1) n. x. \.);s\ xx. all a Ii. ,1. .4 (C; r (at ,CL (or co c CCr m cc; H ac- 0H (cc C (:1 r (ow1flcoi Ifioo Eopm wwwfim . J ..\I 1..-: -..ml 1.....I 6. 1 ... I 1- . -.I .-..\.:J.I 6.. ..- 7).,6 CCC ; C CCC (CD a CCD C CC\ 0 (CC 00 C(C m (LL \0 OCC Vr (D :r :01. M sH000H Coam wmflmm .... .\I (y' \3. .l .. ......s\ \ ... Q I \u | ‘1‘. \I \,l..-.. ' \\.. ...4.‘ -\\u ' 1 .I.~.\I OI..I.‘ II cc; u cab cco cc; 0 cc; 0 CC1 m cco \c (cu ; (cc m0 oco . (Cr 00 HLLC o J.‘.o a) 41m. 0 30.4 to: 4.» . L 0 i In.» Crrkr rqu: PI (. t; C “varf‘rur IN.“ :(.f\ “9.: Ct. (..r.\._.lL {\K 4...: Chw Pk \..f\m.p\ O WICW»rr(_— hr. "...r. k .H r .(MIH-rb F... .r.rvo L.’ Irph. “LL rHC "VFW... O'H..HOL1. Hw.w...r._.0mw IF (.C r TSH MLMCHM IV. 96...)“. “mmflw MO I 0C. MMMUHMH 90 4 I J I ..W 1..< .) ... 1W N ..xr .r TCU‘ (H .... 06.01.?00 .)... I)..- .. \.)\u I ~. g- 3...}; 1 .. .. I ) .\t J \a‘ ),..6\ ,1 -l . .c ocu can act (cc H ccm r 0cm r CC: 003 m ch C Cu m Coo cc: CCHpHmOL [:60 000.0% pmflnm _. .. .) . I )1. I .- .1I ,J- . .I 1...;I .31. I . I w... _. .I ..II » .NI 3 \,I ). 4- OCR m ((0 m CCr m OCQ r Och ; CCr V CC: H CCC & (CC 3 00¢ r CON 0 CCC Houcmw -00000 5010 0000m 1.... ‘I \N... '0 .0 . 5" ......‘n‘- . s | \ \..)., A ' “\Il \IQI' .\ a .> I I J). | _\l_.4 ((0 r (CC m COC C C16 0 00c m CC1 CCO H C(& OCU r CC? 00; m cm .Hpao .000 .00000 000 Lon;u :czzb rw.HHrm 01v;. Héwuaqw ocuugnwrw remAc ; mmricz wmwfim gogpwg pru ;ampC "moHCHo I erwxowm ccd 000a CCmpo sopmmm mam mepmm l “1.01. . ..J. 0 be—IC»LLO. < E moofluo -m010u a ..LOH -, .0 02b Erbtilr F .MHHHHU .Hbomgmm mom b0 Hfithm EU“. m NHQZMLD< L0 11. Lcm 0m Cl C) :3- (‘\ C I O (x 4. 04 O C L Q C r—i W C) C \ r—{ O C3 CA I O O O\ I com.:m- 000.3- 003.0H- ooo.mH- oom.mH- mcflpQOHmnwm C C (“N C‘v C ) .) r-{ O C 1 U \ O C) (CL-I" Cl C L\ O O O C .\ l oom.m- com- ooa.w- oom.fi- oom.a- mcflmmooogm . . 1.1 x... I \ .‘ 4.4.\ - \1.\.. . ‘1 . ...1 a \ .1 4.. . .c_\.\l \ ) ..14 4.¢<\. ul. \) )v J . . \I clll; . 3111.14. 0 n.‘,,l|\-.J.I\. ' .n....J.1... < rc +C... +5 CFC." F” bei C 7r (...4 C C. (.C Wtr C.» CCLU C 4... UL. 7£Crr CHLr ..-: mu .r, -rOrH CCer...... C. r .r "L? . t(_ w. -.. .rrC (SC «TCH ufloflm Igovxd OLAeH spzom IwcmwwsH whom umcm>m fi".p\ co. .0. cow.m oca.m eon- ocC a- coo.m- ooa.m- ooa.fl- oou com m :0: oom.«- oom.m- wsfixwoflgpwm coc. can cow . occ H- com com com 00H com- 0;. coo.fi oos.fl- com _:fl mooogm .1. ~ 4‘ Ill . 1. .al ~ ... 11‘} ... '1‘; ..i... '14.! .31. ...l . I x n 4.:.. .. A. 4. \l1.\ .. .. . .1. .‘1I‘.\{ .1 .— InnI1. .l ,. I<\.4 \ 1 V. ' 11:! t .\.u{tu I. ' 1....Jt‘td\/,.Il.fl FCC'CFL necrwprC Ferrpr...» .r.rr\r> w.?.(f\, r -Cxprxr...r_ $x..»n.(rrrkf (CCCr..C.P-fkw {OUKMIU CFJ myTWLQfir +r\.w..~rf_H rfOrx—rrnsrw Arijo VPQCorc oonFufi To Crux! w..\wr\ run» \ u Q \' . ‘ . \I 1 cra +Loa vcwgc rowzmm hwm mappmm .\. --n'w.\.\vfl u\-wl1._1HU ‘31.:3.’ \U .90.“. ”8| ......)sz1 .|_1. C-..C r C Orrft Cr; ELL .7 u( f. (fibrin... .( .... H- C . ..11qu43.) .1 4.1..‘. , 1'3. .4a) .3. VLF-Ht... C(HPM/Ir J‘L Drlc (;Il_lf\rUr>'.\,Cfer FILI BrfithUD JCLILI (LQ P. h N Q.ham< 114 APPENDIX G REGEESSION AND COPRELATIOH COEFFICIENTS OF TOCATpFACTOR SHIFTS IN FARRICATING INDUSTRIES AND CITY SIZE. AND CHANGE. IN CITY SIZE FOR SELECTED CITIES, 1950-1963 a-“ -—.-.-..———-. --.-.....m r---~v._-‘ --—.—-o --- g—__. - h1§gbiqan_ Indiana All Cities r r‘ r f2 r *2. local-ictor Shift: Citv Size —.#88 .238 -.49Q .hSZ -.622 .387 Fabricating (1950) Local-Factor Shift: Chanve, Size .433 .188 -.571 .326 -.436 .190 Fabricating APPENDIX H CORRELATION COEFFICIENTS OF CHANGES IN WEEKLY EARNINGS AND CHANGES IN TOTAL EMPLOYFEFT, CITY SIZE AND MANUFACTURING EMPLOYFJNT FOR SELECTED CITIES, 1950-1963 ””*Q*M.—*snm II.‘ r r2 Change in E/Wk: Change in total emnl. .301 .091 Change in E/Wk: Change in city size .267 .071 Change in E/Wk: Change in mfv. empl. .132 .017 115 A131)? ‘YDT v T ‘ A 4.4.. ~15 .4 T'T‘f‘"“"Tf‘ A "' ’JD’T‘: H n“ ' f"- ' '1 ”(‘7- f‘fiYT fl m'V-IX' T fl (‘7'- hm“ 4 . . :4 ... » _. - HU~H--Uo£J—l A -nad L—o -o-M d\y’& I T LAK—Jv 03:! b.) TH .-.~4. «SJ ,-hind notation of the derivation of the shifts is J. J. develoned on the basis that the total RCL shift is composed of the 1 sum of the local-factor and tne composition shifts. Any residual in a given piece. kositive or negative, between the total employ- a T ment shi t and the shift lue to industry cheneeS' within the total, CI *1 stems from tie sec or growei. Therefore, the residual plus the snilt within industr" is equa to the total shift. In detailed notation, if, Tia = employment in i industry and j city in J‘ initial period N*%i = emplovnent in i industry and j city in u the terminal period 37-1. = national employment in i ir’ldustry =ZJ- I'ji w - = + .I‘ “L w“- , = . F.- -..J total OIL] el.-L§lOJTTTOnI £1 ..Jl )3 H national employment in all industries N. |\ o -- 3i then, the total shift for a city is H O M 1+ LJ H. V v. * ' 1 s ~ \V (1) is 'J - (.‘Ifo/L‘Ioo) .3303 (2.1. ,- The local-factor shift for each city is A '\J V I- 4 ——4 ..J H s a 4 +¢e (.1. I A 3 4 . ,1. o \ I ”I Ho 0 v I-d -—-4 And, 116 = 2:1 (IT§./Ix1i.)_“ifi .. (ML/BL.) 2:13.113} (3)= :1 [Gig/Ni.) - (Ixr’f./:\I..)] Nij Where (3) is the difference between the total shift and the comeosition shift. That is. IT . — “Y . = “I . zz [Ejfo ”.0 - h1*. ‘ 001KT . a. m <-.. m > 1.1.]. lb BIBLIOGRAPHY Andrews, R. B. "Fechanics of the Urban Economic Bas : Historical Develooment of the Base Conceot", land Economics, Vel. 29, 30, 31, and 32, lav ]_953 t.o Feb. 1956. Beckman, N. J. "Citv Hierarchies and the Distribution of Citv Size", Economic Develonnent and Cultural Ch_°nve, Vol. VI, No. 3, 1958. Ports, G. H. "The Enualization of Returns and Becional Growth", American Economic Beview, V01. 50, June 1960. Chinitz, B. "Contrasts in Aedlomeration: New York and Pittsburch", Papers and Proceedinvs, American Economic Review, Vol. 51, Clark, C. The Conditions of Econonic Procress, The Iacmillan Co., Iondon, 1930. Ferber, R. Sfatistical Techninues in Market Research, MCGrawaHill Book Comnany, New York, 1949, p. 381. Fisher, A. G. B. "Canital and the Growth of Knowledge", Ebonggic Journal, Vol. #3, Sent. 1933. . "Production: Primarv, Secondarv and Tertiarv", Economic Record, Vol. 15, June 1939. Fuchs, V. R. Chances in the Location of Tanufacturinc in the United States Since 1929, New Haven, Yale Universitv Press, 1962. . "The Determinants of the Pedistribution of thu- facturinv in the United States Since 1929, Review of Economics and Statistics, V01. 49, May 1962.“ . ”States or SMA'S When Studying the Location of Manufacturing", Southern Economic Journal, V01. 25, Jan. 1959- Gordon, M. S. Ennloynent Exgansion and Ponulation Growth, The California FMperience 1900-1950, University of California Press. Berkeley and Les Anneles, 1959. Harris, B. P-J Area Systems, Penn-Jersev Study Pacer No. 19, Phila- delnhia, 1962. THeover, E. M. The Location of Economic Ae+1VI+v New York, thrawa Hill 19h8. :[sard, W. Nethnds of Pesional Analysis, New Yerk, WileV. 1960. , 117 118 . 19.C°11°9 and Snace EconoMV. Camhridee, Tech— “‘—--—. ----...-- nolovv Press N.I.T. and-Tile1r 1956. Iichtenberfl, R. N. One-Tenth of a Nation, Cambridce, Harvard -—-..-.. .- Universitv Press. 1960? Leech. A. "The 1% ture of Economic Eerions", Southern Economic Journal, Vol. 5, Jnlv 1939. . jfie Economies o? Location, Yale Universitv Press. 3‘».'e1.-r F1 ven . 19 5E. Mclauqh1in, G. E. and Pobock, S. 'hv Industrv loves South, National 4 -W.—‘,” _.___.- __--_“ P19PP1P" Associa+inn, 19QQ Never, 3. nog~inn91 Apo1voig: A SurveV", Anerioan EconoMic EeVie". Vol. LTTT, Fhrch 1963. —” Moore. F. T. and Petersen, J. N. "Peeional An21_vsis: An Inter- 1EdNSTTV 39991 9f VTPN". Davie" of Frononics_ and Stat1q+1cs. Vol. 37, November 1955. North. D. C. "Iocation Theorv and Peeiona1 Growth", Journal of Poli_tic~1_Economv, Vol. 6?, June 1955. --.. PerIOff’ F. 3" Winn, E. SO! Ifimfll‘d, E0 E0 and ”filth, Jo F0 ppfiifihc E0"finr"°° nni Econoni.c Growth, John Perkins Press. pa1+1moro -’ -- .—-..-—-—-——.—¢— - C'— 1960. Podcers. A. "some Aenects of Industrial Diversification in the U.S.", Pannrs and Proceedings, Peeional Sciences Associa- tion. VblT 1, 1955. Vatter. N. G. The U. S. Economv in +he 1950's W.‘W. Nor on and —-—.--.- -—. ‘o..— —.—..-—C - Comm. Inc.. "72115; non; Vernon, R. The_ Chan~finn Economic Function o_f +he Central Cit“. Oo—o. F‘IM .-.—....-- -c-u. Comm1+tee for Economic Develonment.1959. Vininc. R. "Delimitation of Economic Areas: Statistical Concentions in the Study of the Snatial Structure of an Economic System". Journal of American S+a+is+ica1 Association. vol. #8, 1953. United States Department of Commerce, Regional Trends in the U. S. Economz, A Sunolement to the Survev of Current Business. Washington. Government Printing Office. 1951. "Postwar Exnansion of Outnut". Survey of Current. Pns1ress. Annual Review Number. U. S. Department of Commerce. Jan. 1964. fluent“ 119 EEPPBEE? Repogj of the Pregjdggi, 1962. U. S. Bureau of the Census. Ponulatign Trends in the U. S. 19QQ -=—-—-—. to 19§p, Technical Panor No. 10. U. S. Government b-Q- -— printifiq Office, L‘fashinrrton. D. 0. 196a. SOURCES OF DATA Michivan Emplovment Security Commission, Labor Mazket. 19h9-1963. Detroit. Pfichiaan. Dfichivan Emoloyment Securitv Commission. Labor Forcefignd Ehnlovmenfi. --...w _o—-— .- Estimates, “0. 3221, Rev. 1962, Detroit:- Michip’an. tfichivan Denartment of Health, Ponnlation by;County_and Specified Cities. 1090-1960. Lansing. iichiran. Indiana Emolovment Securitv Division, labor Market Letter, Indiananoiis. Indiana. Indiana Emnlovment Security Division. AreafiTrends. Indianapolis. Indiana. Indiana Emoiovment Security Division, County'Ehployment Patterng, Indiananolis. Indiana. Indiana State Board of Health, Poouintionghy'Counties, 3950-1960, Indiananolis, Indiana. United States Census Bureau. Annual Snrvev of Vannfactures. 19h9- 19‘3. United States Census Bureau, County Business Patterns. 120 WWII! 93 03174 4695 l 312 I llHll