_ .o . .. I... . §.A¢.u. 1. ‘ .. . . ... . . . . . 7 c 0.. . Q l I. .. . . . .............. .......... ........ _ .‘. ..7... ..... ... .. o... ..._. ...H....,... l... . ._ .. ... . _. . .._ . .. ... .. .. ....... .. .......... . . . w...o.o..‘,.:‘6.¢ d. .. . .... . ... . .. . .. . .. . . . i... .. I. ... V. . L. A . .... .. . . . . ... .. I . _. ogo‘9‘1...§ o . .. . . . o . .. .._ .7 . o . . . ... o . . ; ,. . .. . O. n O. O . . o o o o o o c--- - o-in -C*'-qq’m.’.’ ,fi -t-..“‘.~“.‘-‘" a r . . . . V . . 7 _ 7 . _ .7. . . . . . . . . . c O . _ . . . . . o . . . . 7 0 . _ . . . . ._ . _ . . . . . . . . 7 . . . ‘ n u - x . _ . . o 7 . v . . n .x - _ u u . _ — . . . 1—- 5 \ o . . . . . n . . . . . . _ . . . n n v o. . - .. . . . 7 . . ‘ . . . . . . . o n o . . . . o O 7 - O . . 7. . . . . . . . . . . . _ . . . .. . . . o 13,. ~ ilCUL ANALY- - MIC 1 {SAN STATE UN L 1 Thesis for the Degree of-_M..S .1973 mm m AN msmmc no «a MIGRATION FROM AG -- _' ,. g . . 9 v . . n 7. - . J y - I an . . . 7 a J . . . . . 1 _ .- :. ~ .3 A . . I ......u:l .~.o.. .. , 7 u o 1 . . _. . .. .n... C . ,7 . . u . Cu}. .. . .. ..ri . . o . A .. . . 0...: . . Ant. . .. A. . . a . :7. .v {fl-1.. — . _. arr... ...>o.. o‘ck. 7...“... .. u .r s 7. .3». . 000 . 0 AV n . o 4. ..a..IrY x . .. .4.o . (.33- .31. a. .a."§ .0. ... to”; :anmod ... .1345”... JHUHMwonmmw..wu-€ swap; Kora“ an» 7. .... . . . ...4\..... «0.. la. v...\.a “I. .uo“-QVH.. L. .-Ob ‘1‘". ”I".:I.. .MI.V .f‘DV ILV «.15~o‘nmuflfi 2%.; Ilsbl”"11(@fljfllllllflllfllflljllfljlllflglullljlflll him 2658 50% ABSTRACT MIGRATION FROM AGRICULTURE: AN HISTORICAL ANALYSIS By Donald William Gailey Agriculture has been a crucial factor in the economic growth of the nonfarm sector of the United States. Aside from providing food and fiber, this contribution has taken form in the transfer of human capital to the nonfarm sector. Undoubtedly, this massive transfer of productive human resources has permitted a more rapid rate of capital accumulation and product output in the nonfarm sector. However, migration from the agricultural sector to the non- farm sector in the United States has not been without private and social costs. Migration from agriculture has been a selective process. Most farm people migrating to the nonfarm sector move only short distances to obtain employment. However, certain segments of those migrating from agriculture move long distances. These long distance migrants, composed of the young, males, nonwhites, and low income persons tend to move to central city metropolitan areas. The selectivity of the migration process creates attendant Donald W. Gailey problems relative to the viability of receiving areas--the urban problems. The selectivity of the process also has implications for the social and economic health of rural areas. The basic objective of this study was to investi- gate the relationship of various demographic and economic characteristics of actual off-farm movers to the distance they migrated in transferring to exclusively nonfarm employment. Results of this analysis provide a clearer understanding of the impact of migration on rural and urban communities. Meaningful public policies can be formu- lated with this additional information source. Multiple linear regression equations for the nation and for each of the five regions were employed for this analysis. Data employed were derived from the one per cent continuous Work History Sample maintained by the Social Security Administration. Distance migrated was the depend- ent variable in the regression analyses. Income change, race, age, farm employment status, farm earings, distance from an SMSA, and nonfarm industry were the independent variables. With the exception of income change, all inde- pendent variables were categorized and entered the regres- sion equations as dummy variables. From the multiple regression analyses for the nation and the regions, major findings were: (1) Income change was not significantly related to distance migrated, both Donald W. Gailey nationally and by region. Apparently short-run compensation for monetary and opportunity costs was not important in inducing long-distance migration. (2) For the nation as a whole, race was not significant in explaining long- distance migration. However, on a regional basis, Negroes were willing to migrate longer distances than nonNegroes. (3) Long-distance migration for all equations was inversely related to increasing age, particularly for farm individuals 35 years of age and older. (4) Single and multiple job farm wage workers were more responsive to distance than multiple and single job farm operators, both nationally and by region. (5) Farm earings prior to migration for all equations were negatively related to distance. (6) Distance migrated and distance from an SMSA were inversely related for off~farm migrants living within 50 miles of an SMSA but were positively related for greater distances from an SMSA, particularly for the nation as a whole. (7) Off- farm migrants moved greater distances to construction, manufacturing, primary and service industries, and govern- ment employment than to utilities and wholesale and retail trade employment. Since long-distance migration was found to be primarily associated with the young, Negroes, low income persons, and off-farm movers initially securing blue-collar jobs, the implications for urban areas included receiving farm migrants with few marketable skills and uncertain Donald W. Gailey employment possibilities, the probability of more rapid unemployment, and possibly frustration, discontent, and poverty. Alternatively, the impact on rural areas included continual economic deterioration from the loss of potential production and consumption necessary for a viable rural eco- nomy, further inadequacy of basic public services, and possibly rural poverty. The public policy implications of the analysis indicated that mutual cooperation between rural and urban areas must be encouraged. For a lasting solution to the problems attendant to migration, meaningful public policies must be designed to promote a heterogeneous development of rural economies and an improvement of rural living condi- tions to discourage further massive off-farm migration and concentration in metropolitan areas. MIGRATION FROM AGRICULTURE: AN HISTORICAL ANALYSIS By Donald William Gailey A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Agricultural Economics 1973 ACKNOWLEDGMENTS The author wishes to express his appreciation to his two major professors during the course of this study, Drs. Dale E. Hathaway and James B. Johnson. Dr. Hathaway's sug- gestions and criticisms were instrumental in the development and early drafts of the thesis. Dr. Johnson's contributions were invaluable to the completion of the final draft. By meticulously reading and editing the entire thesis, Dr. Johnson provided many helpful suggestions and much needed guidance. No student could ask for a more generous effort from a major professor. A thank you is also extended to Dr. Robert L. Gustafson who provided assistance with quantitative problems and offered worthwhile suggestions for the interpretation of the results of the analysis. Appreciation is also expressed to the Department of Agricultural Economics for the provision of financial assistance and study facilities. The author wishes to express a very special thanks to his parents, Mr. and Mrs. A. T. Gailey. Their continual support and encouragement made possible the completion of this phase of my education. ii Finally, I wish to express my sincere appreciation to my wife, Nikki, and girls, Adella and Brenna, for their unselfish sacrifices throughout the writing of this thesis. Their sacrifices may never be repaid. iii TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . vi LIST OF APPENDIX TABLES . . . . . . . . . . . . . . . vi1 CHAPTER I. INTRODUCTION . . . . . . . . . . . . . . l The Problem Setting . 1 The Problem . . . 8 Objectives of the Study . . . . . . . . . . . 9 The Outline . . . . . . . . . . . . . . . . 10 II. CONCEPTUAL ECONOMIC FRAMEWORK . . . . . . . . . 12 Introduction . . . . . . . . . . . . . . 12 A Theory of Migration . . . . . . . . . . . . 13 Costs of Migration . . . . . . . . . . . . . 16 III. METHODOLOGY . . . . . . . . . . . . . . . . . . 23 Discussion of Hypotheses . . . . . . . . . . 23 The Model . . . . . . . . . . . . . . . . . 29 Statistical Tests . . . . . . . . . . . . . . 37 Outline of the Data . . . . . . . . . . . . . 39 IV. RESULTS OF ANALYSIS . . . . . . . . . . . . . . 44 Introduction . . . . . . . . . . . . . . . . 44 Migration Patterns . . . . . . . . . . . . . 45 The Regression . . . . . . . . . . . . . . . 62 V. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . 83 Summary . . . . . . . . . . . . . . 83 Impact on Urban Areas . . . . . . . . . . . . 88 Impact on Rural Areas . . . . . . . . . . . . 89 Policy Implications . . . . . . . . . . . . . 90 iv BIBLIOGRAPHY . APPENDICES A B C DEFINITION OF TERMS . . REGIONAL DIVISION OF STATES APPENDIX TABLES Page 94 100 101 102 Table 1. LIST OF TABLES Percentage distribution of off-farm migrants by income change, and by region, 1957-60 Percentage distribution of off-farm migrants by race, age, and by region, 1957-60 . . . Percentage distribution of off-farm migrants by farm employment status, and by region, 1957-60 9 o o o o o o o o o o o o o o 0 Percentage distribution of off-farm migrants by farm earnings, and by region, 1957-60 Percentage distribution of off-farm migrants by distance from an SMSA, and by region, 1957-60 . . . . . . . . . . . . .‘. . . Percentage distribution of off-farm migrants by nonfarm industry, and by region, 1957-60 Results of regression equation of demographic and economic characteristics associated with distance migrated, the Nation, 1957-60 Results of regression equation of demographic and economic characteristics associated with distance migrated, by region, 1957-60 vi Page 47 49 52 54 57 59 63 Table 1. LIST OF APPENDIX TABLES Results of regression equation of demographic and economic characteristics associated with distance migrated, the Northeast, 1957-60 . . . . . . . . . . . . . . . . . Results of regression equation of demographic and economic characteristics associated with distance migrated, the North Central, 1957-60 . . . . . . . . . . . . . . . Results of regression equation of demographic and economic characteristics associated with distance migrated, the South, 1957-60 Results of regression equation of demographic and economic characteristics associated with distance migrated, the Plains, 1957-60 Results of regression equation of demographic and economic characteristics associated with distance migrated, the West, 1957-60 vii Page 102 104 106 108 110 CHAPTER I INTRODUCTION The Problem Setting Since the depression decade of the 1930's, agricultural economic development in the United States has been phenomenal. The material record in both production and marketing is remarkable, considering the complexity of the problems with which agriculture has been plagued. Amid price and income uncertainties, agriculture has developed into a technologically advanced and highly productive industry. Through the employment of various technological innovations, an average farmer of today supplies himself and over 40 other persons with farm products. Comparable economic progress has also occurred in the nonfarm sector. This progress has come partially as a result of the accom- plishments in agriculture. Especially in earlier decades, agriculture was a crucial factor in the economic growth of the nonfarm sector., Aside from providing food and fiber, this contribution took the form of a transfer of human Capital to the nonfarm sector. The transfer of these pro- ductive human resourCes was important in setting the stage for further maturity in growth by permitting a more rapid rate of capital accumulation and greater increases in non- farm output. However, so sanguine a view of the deveIOpment in agriculture and its accompanying contribution to the non- farm economy must be interpreted with care. The process of economic development in the United States has been charac- terized by a rapid transformation from a rural, agricultural economy to an urban, industrial one. Such change has involved drastic alterations of social and economic rela- tionships, particularly with respect to migration of human capital from agriculture. In more recent decades, the agri- cultural economy has become increasingly dependent on nonfarm economic growth, full employment, and public agricultural policies. Cities have grown in population with such rapidity that social and economic problems have become almost unman— ageable. As a consequence, the social and economic health of many rural and urban areas may be in question. The historical record of migration from agriculture highlights a massive transfer of people.1 After the depres- sion, the proportion of the total population composed of farm people declined rapidly. From 25.8 per cent in 1933, 1See Appendix A for definitions used in this study. the farm population declined to an estimated 4.8 per cent 2 While some of the of the total U.S. population in 1970. change was due to large increases in the total population, of major importance was migration out of agriculture. For the period 1930 to 1962, the net movement from farms to nonfarm areas and/or occupations was more than 23 million people.3 For the three decades prior to the 1960's, the rate of out-migration was even more pronounced. For the 1930-40 period, the net out-migratiOn rate was 13 per cent of the initial 1930 population, compared with 31 per cent of the initial 1940 population for the 1940-50 period, and 29 per cent of the initial 1950 population for 1950-59.4 In the words of Johnson, "One is almost led to wonder how the crops and livestock were tended!"S zU.S. Department of Agriculture, Farm Population Estimates For 1910-62, by Vera J. Banks, Galvin L. BeaIe, and Gladys K. BoWIes, ERS-130 (Washington, D.C.: Govern- ment Printing Office, 1963), Table l, p. 19, and U.S. Bureau of the Census, Current Population Reports, Series P-27, No. 42, "Farm Population“ (WaShington, D.C.: Govern- ment Printing Office, 1971), Table A, p. 1. 3U.S. Department of Agriculture, Farm Population-- Net Mi ration From the Rura1--Farm Population, 1940-50, by GIadys K. Bowles,’StafisticaI Bulletin No. I76TTWasHington, D.C.: Government Printing Office, 1956), Table 1, p. 16, and U.S. Department of Agriculture, Farm Population Esti- mates For 1910-62, Table 5, p. 23. 4U.S. Department of Agriculture, Farm Population-- Net Mi ration From the Rura1--Farm Population, 1940-50, TaBIe E, p. I3, and Brian B. Perkins, ”Labor'Mobility Between the Farm and the Nonfarm Sector,” (unpublished Ph.D. thesis, Michigan State University, 1964). 5D. Gale Johnson, ”Policies to Improve the Labor Transfer Process," Amer. Econ. Rev., L (May 1960), p. 403. Traditionally, two basic reasons are given for the occurrence of farm-nonfarm migration. These are: (l) the low returns to human effort in agriculture relative to non- farm employment and, (2) the declining demand for labor in agriculture. Notwithstanding the large movements of labor out of agriculture, for many, the returns to human effort in agriculture have not risen relative to labor earnings in the nonfarm sector. The farm population per capita income averaged only 40 per cent of the nonfarm level for the period 1935-39, 54 per cent for the 1940-49 and 1950-59 periods, and 69 per cent for the period 1960-70.6 Based on 1960 data, and after appropriate adjustments for differences in labor capa- city, sex and age composition, labor force participation, relative share of labor earnings, purchasing power of money income, income tax payments, and value of home produced food at retail prices, Hathaway estimated ". . . that the average per capita income on farms would have to be about 88 per cent of nonfarm levels to represent comparable labor earnings for the farm population."7 With the advent of the technological revolution in agriculture in the 1930's, the demand for labor steadily 6U.S. Department of Agriculture, Farm Income Situa- tion, (Washington, D.C.: Government Printing Office, July I97I), Table 7H, p. 50. 7 Dale E. Hathaway, Government and Agriculture: Economic Policy in a Democratic Society (New York: Mac- millian and Company, 1963), Ch. 2, p. 35. declined, especially for certain enterprises and regions. The total man-hours of labor used for farm work declined from 22.3 million in 1930-34 to 6.8 million in 1969. For the same periods, the number of tractors, exclusive of steam and garden, increased from one million to 4.8 million.8 The decline in labor demand was particularly true for milk, feed grains, cotton, and hay and forage, all of which had labor intensive processes replaced by capital intensive processes. Moreover, this decline was the most pronounced in the Southeast, Delta, Southern Plains, and Appalachian regions, in general reflecting the decline in labor demand associated with the enterprises specific to the regions. Along with the magnitude of movements out of agri- culture is the well documented observation that the transfer process itself is very selective.9 Adjustments to nonfarm 8U.S. Department of Agriculture, Changes in Farm Production and Efficiency, Statistical BuIIetin No. 233 (Wishington, D.C.: Government Printing Office, 1970). 98ee Dale E. Hathaway, "Occupational Mobility from the Farm Labor Force," in Farm Labor in the United States, ed. by C. E. Bishop (New York: Columbia Un1vers1ty Press, 1967), pp. 71-96; Dale E. Hathaway, ”Migration from Agri- culture: The Historical Record and Its Meaning," Amer. Econ. Rev. (May, 1960), pp. 379-391; D. Gale Johnson, "Labor Mobility and Agricultural Adjustment," in Agricultural Adjustment Problems in a Growing Economy, ed. by Earl 0. Head”, et a1. (Ames, Iowa: Iowa State University Press, 1958), pp. 163-172; Vernon W. Ruttan, "The Human Resource Problem in American Agriculture," in Farming, Farmers, and Markets for Farm Goods, ed. by Karl A. Fox, et al. (New York: Cemmittee for Economic Development, 1962), pp. 73-116; and Larry A. Sjaastad, "Occupational Structure and Migration Patterns," in Labor Mobility and Population in Agriculture (Ames, Iowa: Iowa State University Press, 1961), pp. 8-27. opportunities through migration appear to be significantly affected by biological, social, and economic inheritance. Often, those farm individuals who could benefit most from . off-farm migration do not migrate. Likewise, success in adjusting to a nonfarm environment and remaining permanently employed in the nonfarm sector appear to be related to this. selectivity. In a recent study for the President's National Advisory Commission on Rural Poverty, Hathaway and Perkins found that, although most off-farm movers do not migrate more than 50 miles, long-distance migration was more common among the young, Negroes, males, low income persons, and 10 However, they concluded those from high income counties. that long-distance migrants did not have long-term earnings as high as short-distance movers and that long-term earnings were highest for whites and persons with high incomes in agriculture before moving. Moreover, they found that long- distance migrants were more likely to seek employment in large cities, more likely to migrate again after leaving farm employment, but less likely to return to farm employ-~ nmnt. The findings of Hathaway and Perkins concur with those of other studies. In a study of the social and E 10Dale E. Hathaway and Brian B. Perkins, ”Occupa- tional Mobility and Migration from Agriculture," in Rural ,Egvertyin the United States, Report by the President's National Advisory Commission on Rural Poverty (Washington, D.C.: Government Printing Office, 1968), Ch. 13, pp. 185- 237. economic conditions of Negroes, it was found that, for the period 1950-66, the number of Negroes living in the central cities of metropolitan areas grew sharply while increases in smaller cities, towns, and rural areas were negligible.11 Consistent with these findings are those of a Bureau of the 12 For the civilian, noninstitu- Census migration study. tional population 18 years and over in May, 1958, 17 per cent of nonwhites born on farms had current residences in places of 500,000 population and over, however, only 3 per cent of the whites had current large city residences. More- over, in another Bureau of the Census study, migration rates were found to be higher for males, nonwhites, the young, and persons with low incomes.13 Migration from agriculture in the United States has been a massive but highly selective process. Most of the farm people who change to nonfarm employment migrate short distances. However, long-distance migrants, composed 11U.S. Bureau of the Census, Current Population Re orts, Series P-23, No. 26, BLS Report No. 347, "Social ang Economic Conditions of Negroes in the United States" (Washington, D.C.: Government Printing Office, 1968), p. 4. le.S. Bureau of the Census, Current Population Ren Eorts, Series P-23, No. 25, ”Lifetime Migration Histories of e merican People" (Washington, D.C.: Government Printing Office, 1968), Table 6, pp. 56-69. 13 U.S. Bureau of the Census, Current Population Re- orts, Series P-20, No. 171, "Mobility ofithe Population of the United States: March 1966 to March 1967" (Washington, D.C.: Government Printing Office, 1968). of the young, males, nonwhites, and low income persons, tend to move to central city metropolitan areas. This selective process has implications for the economic and social health of many rural and urban areas. The Problem The foregoing discussion of the magnitude and selectivity of migration from agriculture has emphasized two possibly related but unanswered questions. First, what is the impact of-out-migration on rural areas? There is some evidence to indicate that the rapid depopulation of many rural areas has severely burdened private busi- nesses and organizations, schools, and local units of government in economically providing basic public services. In consequence, for many rural areas, these institutions have suffered functional deterioration if not total physi- cal breakdown. Second, what is the impact of out-migration on urban areas? It would appear that, in view of mounting social and economic problems, large cities are not capable of assimilating the continuing mass migration from rural areas. These questions conspicuously point to the lack of knowledge of the processes of migration from agriculture, both from the point of view of the people involved and their impacts on sending and receiving areas. More informa- tion about the factors that characterize actual off-farm migrants need be obtained, in order to better understand not only farm people's desire to incur the migration experience but their ability to cope with nonfarm life. Furthermore, too little is known.about the relationship between rural and urban problems resulting from migration and if separate or unified efforts are necessary or desirable for their solution. If successful public policy is to be formulated to facilitate solutions to these problems, more accurate knowledge of the processes of migration is needed. Objectives of the Study This study is the second utilizing data derived from records provided by the Social Security Administration. The first, by Hathaway and Perkins, was generally concerned with the experiences of low income farm individuals in find- 14 To add to that analysis, the ing nonfarm employment. general purpose of this study is to investigate the rela- tionship of demographic and economic characteristics of actual off-farm movers to the distance they migrated in transferring from farm to nonfarm employment. To accomplish this, the migration patterns of off-farm migrants and multiple linear regression equations with dummy variables are employed. It is hypothesized that this approach will 14Hathaway and Perkins, "Occupational Mobility and Migration from Agriculture," in Rural Poverty in the United States, Ch. 13, pp. 185-237. 10 yield results from which inferences could be formulated relative to the impact of migration on both sending and receiving areas and thereby a greater understanding of the link between rural and urban problems. The analysis is conducted and interpreted within a framework suggested by economic theory. Specifically, the-following questions are to be answered: 1. What demographic and economic characteristics of actual offnfarm migrants are significant in explaining the distance of migration, for the nation and by region? 2. What effect does out-migration have on rural areas? 3. What is the impact on urban areas? 4. What public policy alternatives might be effective in solving these problems? The Outline The remainder of this thesis is comprised of four chapters. Chapter II presents the conceptual framwork. Particular treatment is given to a theory of migration and investments in migration. Chapter III describes the method of analysis. Several hypotheses are advanced and discussed; in addition, the statistical model, the form of available data, and the statistical tests of hypotheses are'presented. 11 Chapter IV preSents data employed in the study on the percentage distribution of off-farm migrants by demo- graphic and economic characteristics. Analytical results of the statistical analysis are reported. Interpretation of the regression equation estimates is provided. Chapter V summarizes the statistical analysis, provides conclusions relative to the impacts of migration on urban and rural areas, and discusses selected policy implications. CHAPTER II CONCEPTUAL FRAMEWORK Introduction The labor market is an allocative mechanism by which human resources are shifted between various forms of pro- duction and between sectors of the economy. In a perfectly competitive economy, wages provide the mechanism for effi- cient resource allocation, where efficiency can be defined in terms of maximizing net national product. When labor resources are allocated among and within markets such that the value marginal products divided by the apprOpriate wage rates of labor are the same in all employments (i.e., VMPLi/PLi are equal for all i), the allocation is optimal from the point of view of economic efficiency. Hence, malallocation of labor resources implies a level of national income and output below that possible and a slower rate of economic growth over time. An efficient labor market as an allocative mechanism must be able to shift resources in two ways: (I) adjust the general price of labor relative to other resource prices so as to clear the market of labor (the interfactor allocative process); and (2) shift labor from one sector of the labor 12 13 market to another (the intrafactor allocative process).15 Migration of labor from agriculture is an example of a market mechanism to promote both efficient interfactor and intrafactor allocation of human resources. The emphasis of this inquiry is on shifts from specified classifications of farm labor to exclusively nonfarm employment in various industries within the nonfarm sector. A Theory of Migration In general, a consumers' income is payment for work performed by him, whereby satisfaction is derived from the commodities purchased with the income and from leisure. Hence, from an analysis of utility maximization, the optimum amount of work that the consumer will be willing to perform and his demand curve for income can be derived. When viewed in this manner, the theory of migration is essentially a special case of the theory of consumer behavior. Consider a small, perfectly competitive, single labor market. A utility maximizing worker will adjust his offer- ings of labor services such that his marginal rate of sub- stitution of income for leisure is equal to the real wage 16 rate in that market. Since the competitive assumption 15Lowell E. Galaway, "Labor Mobility, Resource Alloca- tion, and Structural Unemployment," Amer. Econ. Rev., LIII (Sept., 1963), pp. 694-716. 16This approach was adapted from Lowell E. Gallaway, "Mobility of Hired Agricultural Labor: 1957-1960," Amer. Jour. Agr. Econ., 49 (Feb., 1967), pp. 32-5]. l4 precludes the existence of more than one wage rate in that market, the individual will simply choose the amount of work that is optimum for him at that wage rate. If this reasoning is extended to a fully employed, two sector economy, should any real wage differential between markets exist, other than that reflected by opportunity costs associated with transfer, the maximizing worker will offer his labor services to the high-wage market. Therefore, maximizing behavior on the part of all workers implies an increase in the quantity of labor employed in the high-wage sector and a decrease in the low-wage sector until an equilibrium real wage differential between the sectors is established. This means that, under the usual assumptions of a perfectly competitive model, migration from agriculture is a sufficient condition for the elimination of all forms of unemployment in agriculture and the maintenance of an equilibrium real wage differential between the agricultural and the nonfarm sectors which reflects opportunity costs associated with movements from one geographic area to another. Although there has not been complete agreement among economists as to whether the market for agricultural labor has been in short-run disequilibrium or dynamic equilibrium with respect to the nonfarm sector, the foundations of migration discussed in Chapter I are consistent with those discussed within this section. The real gap between farm 15 and nonfarm earnings has induced voluntary migration from agriculture (i.e., farm workers willingly elect to migrate to the high-wage nonfarm sector) while declining demand for agricultural labor has involved involuntary migration (i.e., farm individuals forced to change jobs). For the sample employed in this study, it was not possible to accurately distinguish between voluntary and involuntary migration, although the extent of the latter was believed to be of relatively minor importance. Nevertheless, on a theoretical basis, the presence of involuntary migration will not produce a change in an equilibrium wage structure, even if it is a differential one. If involuntary migration did produce other than the equilibrium wage differential, voluntary migration would occur to re-establish the equilib- rium differential. 17 Operation of the labor market according to these theoretical constructs presupposes the existence of several conditions.18 These are: (l) labor units are homogeneous; (2) there are no non-wage elements of money income associated with employment; (3) workers attempt to maximize their utility functions; (4) there are no differences in workers' preference functions; and (5) there are no restrictive 17Gallaway, "Mobility of Hired Agricultural Labor: 1957-1960, " pp. 32- 52. 18Gallaway, ”Labor Mobility, Resource Allocation, and Structural Unemployment," pp. 694-716. 16 noneconomic impediments to migration. But in reality, the labor moarket fails to allocate human resources in con- formance with these theoretical constructs. Essentially, a fully-employed, perfectly competitive market for labor does not exist. Human resources in agriculture are not homogeneous, particularly in skill and educational qual- ities; in addition, only by chance would farm workers have identical preference functions. There are nonwage elements of money income associated with different employments and noneconomic impediments are in existence. However, the absence of some of Gallaway's condi- tions given above does not preclude a reasonably efficient functioning of the labor market. In particular, conditions one, two, and four are unnecessary. As long as workers attempt to maximize their utility functions, as diverse as they may be, and there are no undue restrictive non- economic impediments to migration, the labor market will sufficiently perform its function of allocating labor resources among various employments. Costs of Migration In conjunction with the foregoing theory of migration as an explanation of the functioning of the labor market, movements of labor from agriculture have at least been in the right direction. However, little information is gained to permit the generation of particular testable l7 hypotheses about the functional relationship between distance of migration and demographic and economic characteristics of off-farm migrants. Hence, for predictive purposes for this study, it is necessary to view the migration problem more specifically in a resource allocation framework--a process of investment in human capital. Since it has been assumed many times that migration from agriculture has occurred primarily in response to higher nonfarm returns, it remains to relate the impediments or costs of such an investment. Many economists have argued that a considerable part of the failure of migration to equate returns to comparable farm and nonfarm labor can be accounted for by various eco- nomic and noneconomic impediments. Economic impediments reflect opportunity cost estimates associated with farm and nonfarm employment while noneconomic impediments do not. Although far from inhibiting migration, as evidenced by off- farm migration rates, these impediments discourage transfers of some farm individuals by making the decision to procure employment in a particular nonfarm industry much more com- plex and thereby successful nonfarm job establishment more remote. Perhaps the real impact is the increase in the selectivity of the transfer process. For the general purpose of this study, private economic and noneconomic impediments to migration are of great significance. These impediments can be reclassified into monetary and nonmonetary costs and are shown to be a 18 function of the distance of migration. Then, from this simultaneous relationship, a foundation is established upon which to conduct an economic analysis of the relationship of demographicand economic characteristics and distance of migration. The monetary costs of migration generally include out-of—pocket expenses. The most obvious cash outlays are the costs of transportation (to move individuals and posses- sions) and the increase in living expenses (for food, lodging, and incidentals). Monetary costs can also be associated with searching for a nonfarm job, especially if one is not found immediately, and can include any combina- tion of transfer and living expenses. Although monetary costs probably are of minor importance relative to non- monetary costs, by any standard monetary costs increase as the distance of migration increases. For nonmonetary costs associated with off-farm transfers, two categories are distinguishable: psychic costs and opportunity costs. These two costs are nonmone- tary in the sense that they involve no direct cash outlays. Psychic costs are generated through people's preferences for their existing social and economic environment. Many people are reluctant to experience changes, even if higher returns for their labor can be expected from alternative employments. Sjaastad has argued that psychic costs involve no real resource costs, since they are of the nature of lost 19 consumer or producersurplus on the part of off-farm movers.19 Hence, efforts to compensate for psychic costs are wasteful. However, Sjaastad further argued that psychic costs do influence resource allocation and will explain part of an existing equilibrium earnings differential between labor markets in the economy. Then, to the extent that between farm and nonfarm labor markets an earnings differ- ential widens with distance and part of this differential has a nonmonetary equiValent in tastes and preferences, it is possible for psychic costs to increase with the distance of migration. This would mean that marginal psychic costs per mile of migration are greater than zero. A second form of nonmonetary costs are opportunity costs, which are generated through earnings foregone while searching for and/or traveling to a nonfarm job and thereby represent real resource costs.20 A part of these foregone earnings will be a function of the distance of migratiOn and quite obviously will vary depending upon farm earnings. Since uniform sources of information in the labor market are nonexistent, potential off-farm movers may learn of nonfarm job opportunities through word of mouth from rela- tives and friends or from other informal sources. Time 19Larry A. Sjaastad, "The Costs and Returns of Human Migration," Jour. Pol. Econ., LXX, Supplement (Oct., 1962), pp. 80-93. 20 Ibid. 20 involved in such a trial and error search can magnify fore- gone earnings. Opportunity costs so generated, even if~ only in terms of travel time, can feasibly increase with the distance of migration and will vary depending upon farm earnings from classes of farm employment. Data on the actual monetary and nonmonetary costs incurred by off-farm migrants are not available, but gen- eral estimates have been made. For capital requirements only, James G. Maddox estimated the following: . . . many farm people can travel as far as five hundred miles from their homes, take ten days to find a nonfarm job and wait a week for their first pay check after they start to work, with a nestegg of no more than $100 a person. Surely these estimates are conservative, even though the estimates are not discounted by what it would cost to live without experiencing a farm-nonfarm transfer. But even taken as fact, for a potential off-farm migrant with a family, a furnished house, a source of home-produced food, and an income of less than $3,000 per year, these costs can weigh heavily in any decision to attempt a transfer to nonfarm employment. L Perhaps a more reliable estimate of the cost of migration is the following: The out- of- -pocket expenses are under $100 for most moves of 100 miles or less. However, out- of- pocket 21James G. Maddox, "Private and Social Costs of the Movement of People Out of Agriculture," Amer. Econ. Rev., (May, 1960), p. 395. 21 costs understate the true costs by omitting foregone earnings, costs of search for new jobs, and, perhaps most importantly, psychic costs. The financial re- sources required by a family to move a hundred miles in North Carolina would appear to be closer to $300 if we include costs of search for both jobs and housing and living expenses incurred before being paid.2 Some indication of the trade-off between income and distance is given from an estimate by Sjaastad: At the mean of the income and distance variables the typical migrant would be indifferent between two destinations one of which was 146 miles more distant than the other, if the average annual labor earnings were $106 (1947-49 dollars) higher in more distant one.23 These latter estimates obviously include compensation for more than monetary costs of migration and represent very high marginal costs per mile of migration. This unex- plained part must be related in part to nonmonetary costs of migration. To summarize, the reality of monetary and nonmone- tary costs complicates the willingness of and decision by farm people to experience off-farm migration. These costs explain in part, in the absence of a perfect market, the existence of income differentials between farm and nonfarm employment, and in this respect nonmonetary costs (psychic 22Paul R. Johnson, "Labor Mobility: Some Costs and Returns," in Rural Poverty in the United States, Report by the Presidenth National Advisory COmmission on Rural Poverty(Washington, D.C.: Government Printing Office, 1968), Ch. 14, pp. 238-247. 23 Sjaastad, "The Costs and Returns of Human Migra- tion," p. 84. 22 and opportunity) are of much more significance. Further, both types of cost, on a private basis, are related func- tionally to the distance of migration. In the absence of actual data, distance of migration serves as a reasonable proxy measure for monetary and nonmonetary costs and con- stitutes a negative incentive to migration. CHAPTER III METHODOLOGY Discussion of Hypotheses The process of migration from agriculture requires resources--investments to enhance human capital. It is assumed in this inquiry that these investments in the form of monetary and nonmonetary costs increase with the dis- tance of migration. It is further assumed that the motiva- tion for incurring these costs was the expectation of higher earnings from farm employment and future well-being in general, although previous research cited revealed that some off-farm migrants did not achieve higher earnings in the nonfarm sector. Under these conditions, it is hypothesized that dis- tance of migration can be explained in part by various economic and demographic characteristics of actual off-farm movers. Therefore, the purpose of the analysis is to test specific-hypothesesabout the relationship of these charac- teristics to the distance of migration. These hypotheses are, in part, suggested by the results of an earlier study con- ducted by Hathaway and Perkins on a similar data base.24 24Hathaway and Perkins, "OccupatiOnal Mobility and Migration from Agriculture," Ch. 13, pp. 185-237. 23 24 The economic and demographic characteristics of actual off-farm movers employed in this study are examined in detail in a subsequent section of this chapter. Briefly, these characteristics include income change, race, age, farm employment status, farm earnings, distance from an SMSA, and nonfarm industry. The change in earnings following farm--nonfarm migration is hypothesized to be positively related to dis- tance migrated" .A positive earnings differential between the farm and the nonfarm sector has long been a reality. And, the incentive for incurring the additional costs from migration could come only from the expectation of higher earnings from nonfarm employment. Unfortunately, the income change variable is computed from short-run total incomes rather than from the more desirable permanent or long-run total incomes. Since the race variable includes a Negro-nonNegro differentiation, it is hypothesized that Negroes would be willing to move greater distances in off-farm transfers than nonNegroes.. In general, farm opportunity costs for Negroes are much less relative to those for nonNegroes, by virtue of discrimination resulting in low farm earnings from low skill farm jobs. And limited economic opportunity in nonfarm employment except in the West and North regions has long been recognized. These economic factors, coupled with problems of social and political integration in most 25 communities and thereby the possibility of low psychic costs among Negores could feasibly account for a direct relation- ship between distance and the Negro racial category. Distance migrated and increasing age are hypothe- sized to be inversely related. In general, the payoff period for recouping investments in migration tends to decrease with age, thereby giving much weight to monetary costs of migration and opportunity costs associated with farm employment. Moreover, locational preferences and family responsibilities tend to increase with age, meaning higher psychic costs. Hence, it would appear reasonable to assume that the initial magnitudes of monetary, psychic, and opportunity costs increase substantially with age and, thereby, preclude distance in off—farm migration. With regard to the employment status of farm indi- viduals, it is hypothesized that wage workers are more willing to incur distance in off-farm transfers than farm Operators. In general, psychic and farm opportunity costs are much less for farm wage workers than for categories of farm operators. These farm wage workers usually have few community ties-and public responsibilities, generally are not home owners, and have little capital invested in farm- ing activities. Hence, there is reason to believe that migration to nonfarm jobs appears to be less difficult for wage workers relative to farm operators. 26 The relationship between farm earnings and the dis- tance of migration is hypothesized to be inverse. In general, the higher the level of farm earnings, even if below nonfarm levels, the greater the risks of successfully establishing nonfarm employment with higher earnings, the greater farm opportunity costs, and therefore the greater the complacency of the farm employed and the less the will- ingness to experience distance. However, given that farm earnings have been low, these factors may indicate that earnings from farm employment are not indicative of the ability of potential off-farm migrants to bear the monetary costs associated with distance. An inverse relationship is hypothesized between distance from an SMSA and distance migrated. First, the general favorable economic conditions around such centers and the greater expected availability of nonfarm employment could indicate a direct response by off-farm migrants the smaller the distance from an SMSA. Such individuals would also be expected to have access to the national labor mar- ket. Second, it would seem reasonable to assume that the closer a potential migrant to an SMSA the smaller the mone- tary costs necessary for migration. Third, close proximity to an employment center would minimize the effect of psychic costs from a new environment and opportunity costs from search for a nonfarm job. 27 These hypotheses are supported by the findings of the study by Hathaway and Perkins, with the exception of the income change variable where no uniform decline in losses or increase in gains with distance migrated was found.25 They found that Negroes were a greater proportion of migrants than nonNegroes and were heavily concentrated in the South. In addition, they found the prOportion Of migrants to decline with advancing age, that farm wage workers constituted a higher proportion of migrants than did farm Operators, and that single and multiple job self- employed farm operators were the least likely to migrate. Finally, their study indicates a decreasing relationship between the proportion of migrants and the earnings Of Off- farm migrants while they were farm employed and that off- farm movers from counties within 50 miles of an SMSA constituted the highest proportion of migrants. Tests Of hypotheses concerning the distribution of off-farm migrants by nonfarm industries over distance are based on the ability of these industries via discounted real wage rates to attract potential Off-farm movers and on the ability of the potential migrants, via skill require- ments, to meet these employment Opportunities. For the years 1957-1959, some indication is given by the following average weekly nominal earnings: construction $110.64, 5Hathaway and Perkins,”"Occupationa1 Mobility and Migration from Agriculture," Ch. 13, pp. 185-237. 28 primary industries $103.35, utilities $92.73, manufacturing $85.12, wholesale and retail trade $76.00, and service industries $66.56.26 However, some care must be taken since these data are not discounted and give no indication of the availability of employment for Off-farm migrants. Moreover, the skill requirements within each individual aggregate are quite varied. Nevertheless, based on these rough nonfarm earn- ings and assuming that most nonfarm migrants are relatively unskilled, it is hypothesized that distance would be directly related to manufacturing, construction, government, primary industries--agriculture (nonfarm), forestry, fisheries, and mining, and service industries--finance, insurance, real estate, and services. For the latter group of industries it is assumed that services such as hotels and lodging places and laundry and cleaning jobs would attract Off-farm migrants in greater numbers than finance, insurance, and real estate jobs. However, an inverse relationship is hypothesized for utilities and wholesale and retail trade. These industries are, in general, more prevalant in or near most local communities and thereby located near sources of farm labor. Finally, no justifiable relationship between 26U.S. Department of Labor, Monthly Labor Review, (Washington, D.C.: Government Printing Office, Dec. 1959 and Dec. 1960), Table C-l. 29 distance and the unknown, unclassified, and military category can be hypothesized. Most of these hypothesized relationships are sup- ported by the findings Of Hathaway and Perkins.z7 They found that the proportion of migrant Off-farm movers was highest for construction, service, and manufacturing industries and for government but lowest for primary industries, utilities, and wholesale and retail trade. Their findings for service and primary industries were contradictory to those hypothesized. The Model The model employed in this study is a multiple linear regression equation supported by comparisons of the proportions Of Off-farm migrants by demographic and eco- nomic characteristics. In order to facilitate the analysis and because the initial relationship between distance and the demographic and economic characteristics, excluding income change, could be nonlinear and/or nonquantifiable, sets of zero-one variables are employed as expressions for the characteristics, with the categories in each set being 28 exhaustive and mutually exclusive. Distance migrated is 27Hathaway and Perkins, "Occupational Mobility and Migration from Agriculture," Ch. 13, pp. 185-237. 28 . . . For a complete d1scu551on of dummy (zero-one) vari- able analysis, see Daniel B. Suits, ”Use of Dummy Variables in Regression Equation," Jour. Amer. Stat. ASSoc., 52 30 the dependent variable in the analysis. Farm employment status, farm earnings, race, age, distance from an SMSA, nonfarm industry, and income change are the independent variables. However, for a multiple linear regression analysis in this form, the parameter estimation can not be accom- plished, because the first column in the matrix of squares and cross products is the sum Of the remaining zero-one columns, thereby making the matrix singular and a unique solution indeterminant. Therefore, in order to permit inversion of the appropriate moment matrix, one zero-one variable from each set is omitted. This essentially trans- forms each estimated parameter (coefficient) into the difference between the actual value of the estimated para- meter Of the category Of a characteristic and the estimated parameter of the omitted category of the characteristic. For interpretation purposes each coefficient becomes an estimate of the expected change in the dependent variable, distance migrated, when one moves from the omitted category to another category, holding other independent variables constant at their geometric means. (December, 1957), pp. 548-551; Robert T. Gustafson, "The Use and Interpretation of "Dummy Variables” in Regressions," Michigan State University, Department of Agricultural Eco- nomics, Mimeo (January 1962); and J. Johnston, Econometric Methods (New York: McGraw-Hill Book CO., Inc., 1963)} 31 The use of the procedure outlined above for the pro- posed model results in the following equation: S C s X 2 b0' + blet + s=I k=2 (ask - asl) ztsk + ut t = 1,2,..., N, the number of Observations in the sample; 5 = 1,2,..., S, the number of category-sets (charac- teristics); W I — 2,3,..., C the number of categories in each s’ category-set 5; b0' is the overall constant equal to the original constant term bo plus the coefficients of each of the omitted categories; b1 is the coefficient of the income change vari- able Xlt; (ask - asl) is the coefficient of the appropriate category of a characteristic and is equal to the difference between the actual estimated coefficient ask of the category and the omitted category a$1 Of the appropriate characteristic; 2 = 1, if the Observation t is in the category k tsk Of the category-set (characteristic) 5, other- wise is 0; and “t is an error term. 32 For this multiple linear regression model employed, the basic classical assumptions made are: l. “t is normally distributed; 2. “t has zero expectation (E(ut) = 0); “t has constant variance (E(ut) = 02); “t is nonautoregressive (E(utuj) = 0); and (fl-BM X1t and Ztsk are nonstochastic and have finite variances and all independent variables are less than perfectly correlated. Given the nature of the sample employed, some of the assumptions are violated, with resulting effects on the desired properties of the least squares coefficient esti- mators and the validity of the tests of hypotheses. Desir- able small (finite) sample properties are (l) unbiasedness, (2) efficiency, and (3) best linear unbiasedness (BLUE) while desirable large (asymptotic) sample properties are (4) asymptotic unbiasedness, (5) consistency, and (6) asymptotic efficiency. Although no explicit attempt is made to adjust for the violations Of the assumptions, it is hoped that the violations Of these basic assumptions were not suf- ficient to invalidate the analysis.29 Consider first the violation of the assumption that the error term “t is normally distributed. The least 29For an excellent discussion of the basic classical assumptions of the linear model and the consequences Of various violations, see Jan Kmenta, Elements Of Econometrics (New York: The Macmillan Company, 1971). 33 squares estimators of the regression coefficients are still best linear unbiased (BLUE) estimators, since this property is independent of the assumption relative to the distribu- tion Of the population. And, from the central limit theorem, it can be shown that, irrespective of the distribution Of the error term, the least square estimators retain all asymptotic properties. Finally, tests of hypotheses are strictly valid only in the case Of large samples. Con- sequently, since the sample employed in this study totaled 12,765 observations for the nation as a whole and no fewer than 1,920 observations in any region, nonnormality of the error term would appear not to be a serious problem, i.e., a large sample was.in use. If the second basic assumption of “t having a zero expectation is dropped, the implication is some specifica— tion error resulting from the incorrectly specified regres- sion equation. While several kinds Of such errors are possible, a most common one results from the omission of a relevant independent variable. For this study, vari- ables such as levels Of education and nonfarm city size greater than one million (to which migrated) should have been used; however, such data were unavailable. Neverthe- less, it can be shown that if the omitted independent variables are not correlated with the included ones, the estimator Of bo', the constant term, will be biased and inconsistent, but the estimators Of the coefficients of 34 the included variables will be unbiased, with all other desirable properties unaffected. .However, the estimators Of the variances of the included variables will have an upward bias, resulting in tests of hypotheses with unusu- ally conservative conclusions. Since a priori the degree of correlation between the possible omitted variables and the included ones is not known, the true impact of the error term “t having a nonzero mean cannot be accurately ascertained. The third basic assumption of the classical regres- sion model is that “t has constant variance, technically known as homoskedasticity. However, this study involves many cross-sectional microeconomic observations which may involve substantial differences in the variation of distance migrated for different groups of individuals. The assumption of homoskedasticity may not be plausible on a priori grounds, with the appropriate model being one with an heteroskedastic error term. If such is the case, the least squares esti- mators can.be shown to be unbiased and consistent but not best linear unbiased (BLUE), efficient, or asymptotically efficient. Thus if the error term is truely heteroskedastic but its true nature is unknown by the researcher, the least squares estimators will have some desirable properties. Nevertheless, the estimated variances Of the estimators are biased and the tests of hypotheses are therefore not valid. Hence for this study, if the assumption of homoskedasticity 35 is not satisfied, inferences about the population coefficients are not correct. Since cross-sectional data are employed and no adjustments are made, analysis results must be interpreted with care.‘ By the fourth assumption of the classical linear model, “t is nonautoregressive. This assumption is most frequently violated in the case of the use of time series data. Although this study involved data collected over several years, the nature of the sample of individuals would seem to preclude any carryover effect from time period to time period; hence, autoregression would appear to be of no concern. But if autoregressive error terms did exist, it can be shown that least squares estimators are unbiased and consistent but not efficient, best linear unbiased (BLUE), and asymptotically efficient and further that tests of hypotheses are not valid. The fifth classical assumption is that X1t and Ztsk are nonstorhastic and have finite variances and that all independent variables are less than perfectly correlated. The assumption of finite variances will be assumed to hold for both X and Z 1t tsk' fact storhastic, the important consideration is whether or If the independent variables are in not the variables are independent of the error term “t’ If independence holds, it can be shown that relaxing the assump- tion of nonstochastic independent variables results in the loss of only the best linear unbiasedness (BLUE) property 36 and tests of hypotheses remain valid. However, if an independent variable is correlated with the error term, then all desirable least squares prOperties of estimators are lost and tests of hypotheses are not valid. Correlation problems usually occur in simultaneous equation models, dis- tributed lag models, or in error-in-variables models, consequently, for this study the existence Of stochastic independent variables would appear to_be of no consequence or the true effect unknown. If some independent variables are themselves corre- lated, technically referred to as multicollinearity, the real question is one of degree. For the case of a complete lack of multicollinearity, no problem exists. However, for the case of perfect multicollinearity, it can be shown that the least squares estimators of the regression coefficients are indeterminate. For the intermediate case of some multi- collinearity, it can be shown that the higher the degree, the larger the variances of the least squares estimators and hence the estimates of the regression coefficients are highly imprecise. However, large variances of the esti- mated coefficients may exist even without multicollinearity. For the independent variables employed in this study, undoubtedly some correlation exists, but a priori the pre- cise degree is indeterminate. An equation of the aforementioned functional form is fitted for the nation as a whole and for each of the five 37 major regions. All variables are included in each of these six equations, but no interaction terms are included due to the complexity and difficulty of interpretation involved. The magnitude of the income change coefficient is inter- preted in the usual manner. For all of the zero-one variables, the magnitude of any estimated parameter is interpreted to be the estimated expected change in the dependent variable,-distance migrated Y, when one moves, in category-set s, from category 1 to category k, while holding constant the categories in all other category-sets (characteristics) and the income change variable. However, since 8 > 1, the number of total category-sets, it is not possible to get estimates of the original estimated coef- ficients aSk but this deficiency in the model is believed to be of minor importance, since the direction of the dif- ference coefficient is most important. Statistical Tests The Student's t distribution is used to derive tests for significance of individual regression coefficients. Both the .05 and .01 per cent significance levels are employed. These criteria were selected in the main arbi- trarily, but some consideration was given to the type of analysis being conducted and to the probability Of com- mitting Type I and Type II errors. 38 The significance of the regression coefficients is determined by testing a null hypothesis that the coeffi- cients are equal to zero against the alternative hypotheses that the coefficients are either greater than or less than zero, depending upon the direction hypothesized previously. Therefore, for the income change variable, Ho : bl = 0 against HA : b1 > 0, while for the zero-one variables, . _ = - . - > Ho . ask a$1 0 against the HA . ask aSl < 0, where s = l,..., 6 and k = 2,..., Cs“ The general form of the calculated t statistic is tb = g3, where b is the regression coefficient and Sb is the standard error of b. For appro- priate degrees of freedom and directions of hypothesized relationships, if t 2 t.05 or .01 or t < -t.05 or .01 the null hypothesis is rejected, where t.05 or .01 and -t.05 or .01 are the tabular values of the t distribution. And if so, the alternative hypothesis is accepted, thus meaning that the variable in question was linearly related to the dependent variable, when all other variables are held constant. A thorough discussion of statistical hypotheses testing can be found in Mathematical Statistics by John B. Freund.30 30John E. Freund, Mathematical Statistics, (Engle- wood Cliffs, New Jersey: PrentICe-Hall, Inc., 1962). 39 Outline of the Data The dataemployed in this study are derived from the one per cent continuous Work History Sample maintained by the Social Security Administration. This sample con- tains workers who have been in employment covered by the Old Age Survivors Disability and Health Insurance (OASDHI) system.31 Individuals in the sample remain permanently. From this continuous sample, for the period 1957-1960, individuals who had some form of covered farm employment in one year-and covered exclusive nonfarm employment in the next year are selected. Demographic and economic characteristics of such individuals are tabulated for the conterminous United States by five major regions.32 These data facilitate the tracing of the experience of actual off-farm movers and the assessment of their importance in the transfer process. Distance Migrated.--By using Census coordinates, distance migrated is measured as the direct mileage between the center of population in the county of farm employment 31Detailed treatments of Social Security sample data as sources of farm labor statistics can be found in the following: Uel Blank, "O.A.S.I. Data of the Farm Labor Force,” (unpublished Ph.D. thesis. Michigan State Univer- sity, 1960); Arley D. Waldo, "The Off-farm Employment of Farm Operators in the United States" (unpublished Ph.D. thesis, Michigan State University, 1962); and Perkins, "Labor Mobility Between the Farm and the Nonfarm Sector.” 32See Appendix B for the regional division of states. 40 and the center of population in the county in which the off- farm mover found nonfarm employment. Four discrete mileage categories--l-50; 51-100; 101-150; and 151-3200--are recorded in the sample for appropriate individuals as zero- one variables. In such form, however, distance migrated does not easily facilitate an analysis based on a linear model. Distance migrated can not be used as a continuous response variable; values for distance migrated must be assigned to the discrete categories. Second, because the centers Of population between many adjacent counties are less than 50 miles apart, it is not possible to distinguish between those who migrated 50 miles or less and those who did not migrate. Hence, any mileage value assigned to the 1-50 category would necessarily have to be small in order not to seriously bias the analysis. Consequently, for convenience and in an attempt to mediate these problems, the geometric mean is computed for each separate mileage category and assigned as the appropriate value for each category specified. This procedure results in a low mileage value being assigned to the 1-50 mileage category and a more moderate value to the 151-3200 category. The resulting assigned values are: 7.0713 for the 1-50 mile category; 71.4150 for the 51-100 category; 123.0900 for the 101-150 category; and 695.1300 for the 151-3200 mile category. 41 Income Change.--A change in income in dollars, whether positive or negative, is reported as a continuous variable for each farm individual who transferred to the nonfarm sector. This income change is computed as the total earnings in the second year (nonfarm employment) less the total earnings in the first year (farm employment). For purposes of comparing the proportion of migrants who experienced various levels of income change with the four distance migrated categories, income changes are divided into the following five categories: (1) loss $500 and over; (2) loss $499-l; (3) gain $0-999; and (5) gain $1000 and over. Demographic Characteristics.--Race and age charac- teristics are reported for each farm individual in the year of farm employment, as zero-one variables. The race vari- able includes a Negro-nonNegro differentiation. The age variable, in years, includes the following categories: 24 and under; 25-34; 35-44; and 45 and over. Farm Employment Status.--An individual is farm employed, if in the indicated year prior to off-farm move- ment, some form of covered farm employment was maintained by the Social Security Administration. Five categories of farm employment are distinguished in the sample data, as zero-one variables; (1) farm wage work only; (2) farm self- employment only; (3) farm wage work and nonfarm wage 42 employment; (4) farm self-employment and nonfarm wage employment; and (S) farm self-employment and nonfarm self- employment. Farm Earnings.--Total earnings in dollars from all sources in the year of farm employment are reported in the sample for each Off-farm mover. For the 1957-1960 period, maximum creditable earnings from all types of employment were $4,800. As zero-one variables, five categories are distinguished: $0-1,199; $1,200-l,799; $1,800-2,399; $2,400-2,999; and $3,000 and over. Distance From an SMSA.--For each farm employed individual in the sample, location in miles of the county of farm employment with respect to the nearest Standard Metropolitan Statistical Area (SMSA) is recorded. An SMSA is a Census definition for counties which include cities with population of 50,000 or more. As zero-one variables, four categories are distinguished: 0-50 miles from an SMSA; 51-100 miles; 101-150 miles; and 151 miles and over from an SMSA. Nonfarm Industry.--The industry of the job in which the Off-farm mover had the highest earnings in the year of exclusive nonfarm employment is recorded in the sample. Several industry groups were combined, resulting in the following eight categories, expressed as zero-one variables: (1) primary industries--agriculture (nonfarm), forestry, fisheries, and mining; (2) construction; (3) manufacturing, 43 (4) utilities; (5) wholesale and retail trade; (6) service industries--finance, real estate, insurance, and services; (7) government; and (8) unknown, unclassified, and military. CHAPTER IV RESULTS OF ANALYSIS Introduction The purpose of this study is to investigate the relationship of various economic and demographic charac- teristics of actual off-farm migrants to the distance they migrated in transferring to nonfarm employment. From this analysis it is expected that part of the impact of farm- nonfarm migration on both urban and rural areas can be determined. Moreover, it is believed that this approach can yield a greater understanding of the relationship between rural and urban problems. Data employed in the study, for the period 1957-1960, are derived from records provided by the Social Security Administration. This chapter is concerned with the presentation and interpretation of the results of the analysis. The proportions of Off-farm migrants by demographic and eco- nomic characteristics are compared to support the basic multiple regression analysis. In the regression analysis, distance migrated is the dependent variable whereas income change, race, age, farm employment status, farm earnings, 44 45 distance from an SMSA, and nonfarm industry are the independent variables. Briefly, the hypotheses to be tested are: 1. distance and income change are directly related; 2. Negroes are willing to move greater distances than nonNegroes; 3. distance is an inverse function Of increasing age; 4. classes of farm wage workers are more responsive to distance than classes of farm Operators; 5. distance is an inverse function of farm earnings; 6. distance migrated and distance from an SMSA are inversely related; and 7. distance is a direct function of primary, manu- facturing, construction, and service industries and to government, but is inversely related to utilities and whole- sale and retail trade. Migration Patterns The basic patterns of migration of actual off-farm migrants were determined from the Social Security data by comparing the proportion of off-farm migrants moving longer distances by demographic and economic characteristics. Although the gross tabulations represent a mixture of char- acteristics and can be misleading, it was expected that 46 these proportions would add to an assessment of the relationship between distance migrated and these charac- teristics in the national and regional regression equations. Distance migrated and income change were hypothe- sized to be directly related. However, from Table l, the proportion of Off-farm migrants moving longer distances did not vary uniformly with increasing short-run income changes, both for the nation and in most regions. There- fore, it would appear that some economic factor(s) other than short-run nominal income changes motivated long- distance transfers. For the race variable, Negroes were hypothesized to be willing to migrate longer distances than nonNegroes. From Table 2, for the nation and all regions, the propor- tion of off-farm migrants moving greater distances was larger for Negroes than for nonNegroes. This was particu- larly the case for off-farm movers who migrated 151 miles and over, where roughly one-third Of all Negro migrants traveled the longer distance for the nation as a whole. Surprisingly however, the proportion of Negro migrants corresponding to the 151 miles and over category was smaller for the Plains and South regions relative to the other three, with the South having the lowest percentage. Consequently, these frequency comparisons clearly indicate that long- distance migration was not simply limited to the South and Plains regions. 47 woumhmflz museumfim 9111. nanny. o.o0H u.wH H.< 0.5 o.mo Hence o.ooa m.v~ H.e e.w o.~o po>o use ooo.Hm :fimu o.OOH m.HN v.m m.o m.uo mmmw-oomm came o.ooa m.uH w.v N.m u.wo mmem-ow memo o.ooa m.uH m.e m.n m.on Hm-mmew mmoq o.ooa m.¢H m.m N.o u.mn 96>0 use coma mmog ”mmcmco oEoocH :psom o.ooa o.¢H m.e m.o H.mm HeuOH o.ooa n.eH o.m m.m N.ou go>o mam ooo.Hm :flmu o.ooa N.¢H N.m N.m m.mm mmmw-oomm :Hmw o.ooH v.mH v.v m.m m.vN mmem-ow :flmu o.ooa w.mH H.m N.e o.ee Hm-emem wees o.ooH o.HH v.m m.o m.mu pm>o use comm mmoq "omcmno oEoucH HmHHCOu :unoz o.QOH m.mH u.m m.n m.en Hmpoe o.ooa o.mH m.m m.w N.Nn ho>o tam ooo.Hm came o.ooa o.HH o.m o.n ~.nm mmmw-oomm :Hmo o.o0H e.mH m.e e.e e.ee mmea-oa seem o.ooa o.mH H.m w.n m.mu Hm-mmew mmoa o.oo~ «.ma ¢.m m.o m.on uo>o paw coma mmoq flowcwsu OeoocH ammonuhoz mpo>oe Ho>o can mOHfiE mOHfiE mOHfiE omqmco Osoocfi can :Ofimom HH< mOHfiE HmH omH-HoH ooa-am om-H xn use .omcmno Oeoocfi kn muampmwe Ehmm-mwo mo :ofiusnm .oe-nmmfi .eoamee ppmflw ommucoou6d .H OHan 48 .mxmma< McMOSHOxO .mOpmum vowed: Hancocfipcoo OABN .mpmw kpfipsoom Hmfioom ”oopsomH o.ooa N.HN m.e m.e e.ee fleece o.oo~ m.eN m.m m.m e.He ee>e nee ooo.~a eaeu o.ooa e.o~ m.m o.e w.ee mama-ooma anew o.oo~ m.HN N.m m.e o.ee mmea-oa seem o.ooH w.HN e.e e.e e.ee Hw-amea neon o.oo~ w.eH N.e m.e e.oe ee>o eee coma wees newsmao oEoocH Neeeeez o.ooa m.om 0.0 m.m e.mm Hepoe o.OOH ~.Hm N.m m.a m.em ease wee ooo.aa seem o.ooa w.mm e.e e.e H.mm mama-oomw efieo o.oo~ m.mm m.e m.e w.mm amea-oa seem o.oo~ N.Nm e.e m.w m.~m Ha-aaea ween o.ooa o.eN m.m e.m e.wm ee>e nee coma wees ”omcmgo oEoocH ewe: o.ooH e.m~ m.e w.e e.em fleece o.ooH m.mm m.e w.e e.me ee>o nee ooo.aa seem o.ooH e.mm m.e 0.x o.em mama-ooma eeeu o.ooa m.e~ m.m e.n e.wm emee-oa eeeu o.ooa N.eN o.e o.n e.am Hw-mmea wees e.OOH e.e~ e.m N.w o.~e ee>o can come wees Homemao OsooaH meeefle muo>oE uo>o was mOHfiE mOHfiE mOHfiE omcmzo Oeoucfi cam :owmom HH< eefiae HmH omH-HOH OOH-Hm om-H wepwhwfiz oocmumfin .eeseeeeou--.fl manee 49 Table 2. Percentage distribution of Off-farm migrants by race, age, and by region, 1957-60. Distance Migrated 151 Region, race, l-50 51-100 101-150 miles All and age miles miles miles and over movers Northeast Race: Negro 62.0 3.9 2.0 32.2 100.0 nonNegrO 76.3 8.0 4.0 11.7 100.0 Age: 0-24 71.7 8.4 4.1 15.8 100.0 25-34 72.6 8.1 5.1 14.3 100.0 35-44 74.8 7 3 3.2 14.8 100.0 45 and over 81.7 5.9 2.4 10.0 100.0 Total 74.8 7.5 3.7 13.9 100.0 North Central Race: Negro 56.5 1.6 6.5 35.5 100.0 nonNegrO 75.5 6.6 4.3 13.6 100.0 Age: 0-24 68.7 8.2 5.5 17.6 100.0 25-34 70.3 7.2 4.7 17.8 100.0 35-44 74.4 6.8 4.4 14.4 100.0 45 and over 84.8 4.4 2.8 8.0 100.0 Total 75.1 6.5 4.3 14.0 100.0 South Race: Negro 60.8 8.0 5.7 25.5 100.0 nonNegro 73.1 7.4 3.5 16.0 100.0 Age: 0-24 61.2 7.8 6.3 24.6 100.0 25-34 62.9 8.3 4 3 24.5 100.0 35-44 72.2 9 4 3.6 14.8 100.0 45 and over 78.1 5.9 2.7 13.3 100.0 Total 69.6 7.6 4.1 18.7 100.0 50 Table 2.--Continued. Distance Migrated 151 Region, race, 1-50 51-100 101-150 miles All and age miles miles miles and over movers Plains Race:, Negro 53.2 8.5 8 5 29.8 100.0 nonNegro 57.8 7.7 6.1 28.3 100.0 Age: 0-24 50.6 9 2 6.8 33.4 100.0 25-34 50.8 8.9 6.8 33.6 100.0 35-44 60. 9 6.6 6 0 26.5 100.0 45 and over 68. 6 6.1 5.4 19.9 100. 0 Total 57. 6 7.8 6.3 28. 4 100. 0 West. Race: Negro 40.7 8.5 39.0 100.0 nonNegrO 55.7 6.0 30.1 100.0 Age: 0-24 56.6 7.9 7.0 28.4 100.0 25-34 48. 9 8.2 4.5 38.4 100.0 35-44 57.6 7.9 5.2 29.3 100. 0 45 and over 58. 6- 9 2 6.8 25.4 100. 0 Total 55.4 8.3 6.0 30. 3 100. 0 Nation2 Race: Negro 58.9 7.2 5.6 28.3 100.0 nonNegrO 67.2 7.5 4.9 20.4 100.0 Age: 0-24 61.4 8.3 6.0 24.3 100.0 25-34 60.6 8.1 5.1 26.2 100.0 35-44 68.0 7.6 4.6 19.8 100.0 45 and over 75.2 6.1 3.9 14. 8 100.0 Total 66.4 7 5 4.9 21.2 100.0 1Source: Social Security data. 2The continental United States, excluding Alaska. 51 Again from Table 2, the proportion of off-farm migrants moving longer distances declined with increasing age, for the nation as a whole and for all regions except the West. Hence, these frequency distributions lend strong support to the hypothesized age-distance relationship that distance migrated is an inverse function of increasing age. For the farm employment status variable, from Table 3, the proportion of off-farm migrants moving longer distances was greater for classes of farm wage workers than for classes of self-employed farm Operators. This was the case for all regions and the nation, supporting the hypothe- sis that farm wage workers are more likely to migrate longer distances than farm Operators. Moreover, for most distance categories, the frequency of long-distance moves. was greater for individuals who had some form of farm wage and nonfarm wage employment than for those persons who had only farm wage employment. For self-employed individuals, the proportion of Off-farm migrants moving longer distances was greatest for individuals with farm self-employment and/ or farm self-employment and nonfarm wage employment. Farm earnings were hypothesized to be inversely related to distance migrated. Reasonably consistent with this hypothesis, from Table 4, for the nation and all regions, the proportion of off-farm migrants moving longer distances declined as farm earnings increased. Such was especially evident for farm earnings of $2,400 or more, 52 o.ooa n.wH H.v 0.n 0.m0 kuoe o.ooa H.m 0.0 m.H 0.0m u:ofi>o~mfio-waom spawned use ucoaonQEO-mHOm seem o.ooH H.m 0.m m.0 v.om uqosxoflmso 0mm: Ehmmao: 0cm p¢oE>OHmEO-mHOm Shem o.ooa N.¢N v.0 5.x u.H0 udoEAOHQEO 0mm: Enemco: 0cm x903 owmz seem o.ooa m.mH m.H 5.5 m.uu xfiqo “cosmonEO-mHOm sham o.ooa N.m~ 0.4 e.m e.ee sHee xeoz ewe: aeee "magnum ucoezoamEO 59mm £9300 o.ooa o.eH m.v 0.0 H.mu Hence o.ooa 0.0 <.H w.~ v.~m.ucoe>OHmEO-mHom aammco: 0cm pfioE%OHQEO-mHOm 59mm o.ooa 0.m m.~ m.m m.mw uaoeonQEO 0mm: Ehmmco: 0cm paoexoflmEO-mHOm Eumm o.ooa 0.mH 0.0 N.“ 0.50 uaoaonQEO Owe: Epemao: 0:0 #90: 0mm: 59mm o.ooa w.0 o.e m.0 m.~w sfieo eeeeseHQEe-mHen seem o.ooa m.mH m.v m.u w.w0 kano #903 owe: Shem ”mapmum unoeonaEO sham Hmpunoo Apnoz o.OOH m.mH n.m m.n m.eu Hence o.ooa v.m H.H 0.0 m.om pcmsxofimEO-mHOm Enemno: 02m u:oE%OHmEo-mHOm Egan o.ooa m.e. m.m m.H m.om pcoezofimso Owe: Euwmco: 0cm u:oE%o~mEo-maom sham o.ooa ~.mH m.m m.w H.Nn ucoaxofimEO owe: Epmwcoc 0cm M903 Owe: Ehem o.oo~ N.m m.H w.o~ m.we sane enmesoHQEemeem seem e.ooa 0.0a N.e m.0 w.~e sflee gee: ewes seem "mapmwm uqosxofimso Egan ammoauuoz .4 mno>os Ho>o 0cm mOHwe mOHwE mOHflE msumum uaoaxoamso_ahwm 0cm :Ofimom HH< mOHHE HmH omH-HOH ooa-Hm om-H woumhwflz museumflm . H.o0-emmH .eeamee s0 0cm .m3pmam ucoakofimEO Shem >0 mucmhmwe Eamw-mmo mo :ofiusnfihumflw ommunoonom .m Oanmh 53 0000:002 00:0pmwn .0xm0H< 0:003Hoxo .mop0pm 0ouw:= H0u:m:wu:oo 0:50 .0900 5:0:5000 H0fioom ”Quezomfi 0.00H ~.HN 0.0 0.5 0.00 H0pOH 0.00H m.m N.H 0.~ v.00 0:055OH950-MHOm E:0m:o: 0:0 u:mEAOHQEO-mHom 5:00 0.00H 0.5 0.m N.0 0.N0 u:oE5OHQEO 0003 E:0m:o: 0:0 u:mE5OH:Em-mHOm EH00 0.00H 0.0m 0.0 N.w 0.00 u:eE5oH:Eo OM03 E:0m:o: 0:0 Mao: 0003 EH00 0.00: m.o: o.m m.5 ~.ee 5::0 eeeesoHeEe-0Hen seem 0.00H 0.0m 0.0 0.5 m.~0 5H:O M903 0003 8:00 ”map0um 0:0E5OHQEO EH00 N:0302 0.00H m.0m 0.0 m.0 v.00 H0005 0.00H m.m m.m 0.H ~.H0 ::oE5oHQEm-mHOm E:0w:o: 0:0 0:0E5OHQEm-maom 5:00 0.00H 0.0 v.m m.0H 0.05 p:mE5o~:Em 0003 E:0m:o: 0:0 u:oE%o~:Em-mHom 5:00 0.00H m.Nm 0.0 0.0 0.00 u:oE5oHQEm 0003 E:0m:o: 0:0 xhoz 0003 EH00 0.00H H.0 0.0 N.0H 5.m0 5H:o u:oE5OHQEo-mamm EH00 0.00: 0.0m 0.0 5.5 5.0m 5:00 Mao: 000: See: "mzu0um 9:0E5OH050 EH00 p003 0.00H 0.0N 0.0 0.5 0.50 H0005 0.00H H.0 N.H 5.0 H.00 u:o:50~:fim-maom Eh0m:o: 0:0 “:0E5ofimse-mHom EH00 0.00H m.NH 0.0 0.5 0.05 “:0550H050 0003 E:0m:o: 0:0 p:oE5OHQEe-mHom EH00 0.00H N.em 0.5 0.5 0.00 9:0550HQEo 0003 :u0m:o: 0:0 #:03 0003 :90: 0.00: H.eH 5.e 0.0 N.m5 5::0 e:e§5e:0§e-0:em seem 0.00H m.Hm 0.0 0.0 0.00 5H:o Muck 0003 EH00 "mnu0pm u:oE5OHmEo EH00 0:00am muo>os no>o 0:0 moawa 00H“: moafis m3p0um p:OeonQEo Eh0m 0:0 :ofimem HH< moafis HmH 00H-HOH 00H-Hm 00-H .eeseeeeou--.m efleee 54 0000000: 00:00000 0.000 0.00 0.0 0.0 0.00 0eeo0 0.000 0.00 0.0 0.0 0.00 ee>o eee 000.00 0.000 0.00 0.0 0.0 0.00 000. 00- 000. 00 0.000 0.00 0.0 0.0 0.00 000. .00- 000 J0 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000 JT 00 .mm:0:00m EH00 00:00 0.000 0.00 0.0 0.0 0.00 0eeo0 0.000 0.0 0.0 0.0 0. 00 ee>e 0ee 000.00 0.000 0.0 0.0 0.0 0. 00 000 .0T 000 .00 0.000 0.00 0.0 0.0 0. 00 000 .0T 000 .00 0.000 0.00 0.0 0.0 0.00 000 .0T 000. 00 0.000 0.00 0.0 0.0 0.00 000 0T 00 .mm:0:00m EH00 0eeeeeu 00002 0.000 0.00 0.0 0.0 0.00 0eee0 0. 000 0.0 0.0 0.0 0.00 ee>e use 000.00 0. 000 0.00 0.0 0.0 0.00 000. .0T 000. .00 0. 000 0.00 0.0 0.0 0.00 000 .00 000 .00 0.000 0.00 0.0 0.00 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000 .00- 00 .mm:0:00o 5000 000000002 who>oE Ho>o 0:0 00005 00008 00005 mw:0:h0m £00m 0:0 :00000 00< 00000 000 000-000 000-00 00-0 H .00-0000 .000000 50 0:0 .mm:0:00o 500m 50 00:00006 E00m-wmo mo :00030000000 0000:00000 .0 00909 55 .00000< m00050ux0 .000000 0000:: 00000000000 0:9 0000HM02 00000009 0 .0000 >00Hsuom 000000 000H:om0 0.000 0.00 0.0 0.0 0.00 0000H 0.000 0.00 0.0 0.0 0.00 00>0 0:0 000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-00 00m00cH00 EH00 N:00u0z 0.000 0.00 0.0 0.0 0.00 00000 0.000 0.00 0.0 0.0 0.00 00>0 000 000.00 0.000 0.00 0.0 0.00 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-00 00mc0EH00 EH00 0002 0.000 0.00 0.0 0.0 0.00 00000 0.000 0.00 0.0 0.0 0.00 00>0 000 000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-000.00 0.000 0.00 0.0 0.0 0.00 000.00-00 00m:0:H00 EH00 0:000m 0H0>oE H0>o 0:0 0000E 0000E 0000E 0mE0EH00 EH00 0:0 :00m0m 00< 00000 000 000-000 000-00 00-0 .000000000--.0 00000 56 although the results were mixed for lower earnings. Nevertheless, since the proportions for all classes of farm earnings were greater for large distance categories,' the evidence would seem to indicate an inverse relation- ship between distance and farm earnings. Distance migrated and distance from an SMSA were hypothesized to be inversely related. However, from Table 5, the evidence would seem to indicate no clear negative rela- tionship. For all regions and the nation, the proportion of off-farm migrants moving longer distances did not decrease with increasing distances from an SMSA. For the nation, most migration above 50 miles was done by those individuals who were located either 50 miles and under or 151 miles and over from an SMSA. For the nonfarm industry entered by off-farm migrants, from Table 6, patterns of distances migrated for most industry categories corresponded closely with hypothe- sized relationships. For the nation and all regions, utility and wholesale and retail trade jobs entailed pro- portions of migrants moving shorter distances relative to that for the other industries. The construction industry category had the highest proportion of such migrants for all regions and the nation. Finally, the proportions of migrants moving longer distances were reasonably larger for manufacturing, government, and service and primary industries. 57 o.ooa u.wH H.v 0.5 o.mo fiance o.ooa o.ooa o.o o.o o.o go>o paw moHHE HmH o.ooa H.OH v.m 5.5 w.on onwE omH-HOH o.ooa 0.5 m.~ N.m v.vw moawe ooa-Hm o.ooa m.u~ o.m N.m m.wm monE om-o uo use moHHE HmH o.ooa m.w m.m <.v m.mw moawe omH-HOH o.oo~ m.m H.m m.v n.mw moHHE ooa-Hm o.OOH N.©H o.m Q.“ o.Hn moHflE om-o "o cam moHHE Hma o.ooa N.wH o.m o.m m.vn mofifle omH-HOH o.ooa o.nH m.~ o.v p.05 monE ooH-Hm o.ooa v.mH w.m m.w ~.vu moHfiE om-o "os Ho>o was moHHE moHfiE monE o paw monE HmH o.ooa o.mH H.q o.“ m.mn mofifie omH-HoH o.ooa «.mH o.m w.m N.un mofifle ooa-Hm o.oo~ m.mN m.m m.w m.oo mmfifis om-o ”o paw moHfiE HmH o.ooa m.om H.v 5.5 m.no moafle omH-HOH o.ooa m.vm o.v H.u m.mo moawa ooa-Hm o.ooa H.Nm a.o m.w H.Nm mofifie om-o ”o pew monE HmH o.ooa m.mH H.e w.w N.wo moHHE omH-Hoa o.ooa m.wH w.m o.m m.oo moHHE ooH-Hm o.ooH m.Hv m.u w.o v.ev mofiwe om-o uoe Ho>o paw moafle mofifie moHflE ou o.ooa 0.0a 5.m m.v 0.m5 mofihumsvcfi mow>hmm o.ooa m.HH v.0 N.m m.w5 mwmhu Hfimumh cam mammmaozz o.ooa H.~H ~.H m.5 N.m5 mofiuflfiwpo o.ooH m.0H 0.0 0.5 v.m0 nofluoznumcou o.ooa w.m~ 0.H N.m ¢.H5 mownumswcfi knmsfihm o.oo~ H.¢H m.m ~.5 0.05 mcfipsuommscmz ”xpumswcw Ehmmcoz Hmhucou nuhoz o.oo~ m.ma 5.m 0.5 0.05 . Hmuoe o.ooa m.mH H.w 5.N 5.m5 59mpfiHHE 0am .mowmfimmmHucs .czocxcs o.ooa o.mH 5.v m.m o.H5 ucoecho>ou o.ooa H.5H 5.m 0.0 m.O5 mmfihpmswew oufi>gom o.ooa o.v~ 0.N 0.5 0.05 0009» Hangoh 0cm mammofionz o.OOH 0.0H 0.0 H.0H m.m5 moflufififius o.ocH ~.HN m.m 0.0 0.00 coauoahumcou o.ooa 0.mH o.o m.HH 0.V5 mofipumsuefl semeflgm o.oo~ N.OH m.¢ N.0 «.05 mnfihsuuamscmz "xhumswnw Enmmcoz ummonuaoz who>oE Ho>o 0cm moHfiE moafle monE xhumswnw Eammno: 0am cowmom HH< moHHE HmH omH-HOH oouufim omua woumhmfiz oocmpmfln .o0-5mmH 5239H 50 0am .zhpmswcw Epmmco: 50 mpcwhwfls Epmm-mmo we soapsnmhumflw omwunoohom .0 magma 60 00000002 00:00000 0.000 0.00 0.0 0.5 0.50 00000 0.000 0.00 0.5 0.00 0.00 50000005 0:0 .000w0mmw0ocs .czocxcb 0.000 0.00 5.0 0.5 0.00 pd0EGH0>oo 0.000 0.00 0.0 0.5 0.00 m00upmswc0 0U0>000 0.000 0.00 0.0 0.0 0.00 00000 000000 0:0 000000002 0.000 5.00 0.0 0.5 0.00 00000000: 0.000 0.00 0.00 0.00 0.00 :00puspumcou 0.000 0.00 0.0 0.0 0.50 00000050=0 0005000 0.000 0.0m 0.0 0.0 0.00 000030000500: ”xhumswa0 Enamcoz 0:0000 0.000 5.00 0.0 0.5 0.00 00009 0.000 0.00 0.0 0.0 0.55 #0000006 000 .00000mmm0ocs .azocxcs 0.000 0.00 0.0 0.00 0.00 uc0scp0>ou 0.000 0.50 0.0 0.0 0.05 .000upmswa0 0o0>p0m 0.000 0.50 0.0 0.0 0.05 0000» 000000 000 000000003 0.000 0.00 5.0 5.0 0.55 00000000: 0.000 0.00 0.0 0.00 0.00 :00uo:pum:ou 0.000 0.00 0.0 0.0 0.00 m00humswc0 #005000 0.000 0.00 0.0 0.0 ¢.m0 mawhspummsamz "xhumswa0 Epmmcoz £0500 mn0>oE H0>o 0:0 00005 00005 00006 50005000 summcoc 0:0 :00000 00< 00005 000 000-000 000-00 00-0 .0mncfipnou--.o 00000 .00000< 000000000 .000000 00000: 00000000000 000 61 0 .0000 00000000 000000 "0000000 0.000 0.00 0.0 0.5 0.00 00000 0.000 0.00 0.0 5.5 0.00 00000000 000 .000000000000 .030000: 0.000 0.00 0.0 0.0 0.00 000E000>00 0.000 0.00 0.0 0.0 5.00 0000000000 000>00m 0.000 0.00 0.0 0.0 5.05 00000 000000 000 000000003 0.000 0.50 0.0 0.00 0.05 00000000: 0.000 0.00 0.5 0.00 0.00 000000000000 0.000 0.00 0.0 0.0 0.00 0000000000 0000000 0.000 0.00 0.0 0.5 0.00 0000000000002 "00000000 0000002 000002 0 . 0.000 0.00 0.0 0.0 0.00 00000 0.000 0.00 0.0 0.0 0.00 00000000 000 .000000000000 .0300000 0.000 0.00 0.0 0.0 5.00 000E000>0u 0.000 0.00 0.0 0.0 0.00 0000000000 000>000 0.000 0.50 0.0 0.5 0.00 00000 000000 000 000000003 0.000 0.00 0.0 0.00 0.00 00000000: 0.000 0.00 5.0 0.5 5.00 000000000000 0.000 0.00 0.0 0.0 0.00 0000000000 0000000 0.000 0.00 0.0 0.0 0.00 0000000000002 "00000000 0000002 0003 000>00 00>0 000 00000 00000 00000 00000000 0000000 000 000000 00< 00000 000 000-000 000-00 00-0 00000002 00000000 .000000000--.o 00000 62 The Regressions Parameters of the regression equations were estimated for the nation and five major regions. For each equation, dependent-independent relationships and statisti- cal tests employed were the same. The Nation.--The results of the regression analysis for the nation is shown in Table 7. Two things are appar- ent. First, little of the variation in the dependent variable--distance migrated--was explained by the independ- ent variables. Several unspecified variables, such as levels of education, long-term earnings, and nonfarm city size may have assisted substantially in explaining the variation of the dependent variable. Second, even though the R2 value was small, 15 of the 24 independent variables (including the constant term) were significant at the .01 or .05 level. Income change was not significantly related to distance migrated. For those off-farm migrants included in the sample, on the average, holding other independent variables constant, short-run changes in income resulting from nonfarm employment were not important in inducing migration. Hence, it would appear that something other than short-run income changes was necessary to compensate for monetary and opportunity costs associated with migration (whether voluntary or involuntary). 63 mmH.Hm-oe "umppfiEo «gm vwmo.n- Nmno.o- nomH.N mnmn.om- pm>o wcm ooo.mm *«m NmHv.¢- Havo.o- «ONm.m ommH.Hv- mam.Nw-oov.Nw «gm owmm.v- vHvo.o- mme.w nmam.om- mmm.Nw-oow.Ha mz NHmV.H- omao.o- vaN.N Naom.OH- mmN.H*-o0N.Hw ”mw:H:hwo Sham xaco ucoeonQEo -mHom Epmm “wopuwso . - . - . . ucosonQEo-mHmm Summco: «m NmNH N NONo o Nnmw NH mHNe mN- can pcmsonmEo-mHom 59mm . - . - . . ucoskonEm ommz Shawna: mz oomo o maoo o qwmm NH mead H- was ucoeonmEo-mHmm sham . . . . ucmeonmEm 0mm: swam «gm wNOH m Hana o oumo NH meow OOH -coq wan xpoz 0mm: 59mm ««m Hmhm.o QQOH.O mme.NH Hmmo.ow Nflao xhoz mmmz agmm ”msumpm ucoeonmEm Swan #06:: wcw VN "wowpfieo mz wmmo.~- NwHo.o- mwmu.o NNNH.HH- Ho>o can m¢ mz Noam.H weao.o mnmN.u nemH.OH vw-mm «xm Nmon.m nwmo.o momw.o oaon.mm em-mNm ,"m < opmozco: "umppfieo mz «OHm.H vmfio.o ommm.w omhv.NH ohwoz ‘ "oumm mz HHmN.o mNoo.o mHoo.o coco. mmcwcu oEoucH «em ammo.w -- mmow.NH H unawaomowaH -maqmfim «pom mo muchho .wam nowmmouwom N .oo-ummH H.aoflpmz may .woumnmfis mucmpmfiw Sufi: wmumNUOmmm muwumwhmpumpmnu uwaocoum can ofinmmamoEow mo aofiumncm :ONmmohmmh mo muasmmm .u manmh 64 .Hm>oH mo. ogu an open Scum ucmaommfiw Afiunmuwmficwfim no: mm: pcofluflmmmou map mcmoa mz Ho>mH Ho. u ««m Ho>oH mo. u «mm Ho. u oucmofimfizmflm VNN<.NV u m ome.NON u .m.m.m NHNo.o u Nm HON.NH u mcoflgm>powpoN .mxmmfi< mcfiwsauxm .mmumum umuwcb Hmuamcfluaou mcefi mqflpsuummscmz ”wouufieo . . . . Npmufififie ««m mmmv w Hove o unoo ma choc oo mam .wmfimwmmmHocs .azocxcs ««m wwan.m Nmmo.o Nnon.oa mmfiw.mm ucoecpm>ou «cm comm.N uwNo.o NmNm.n wmom.NN mmfihumsvcw oofi>pom mz mmNN.o- «Noo.o- omvm.o mvuw.a- mwmhw Hamuou wcm mammoaonz mz oemm.o- mmoo.o- mmVN.HH nmmm.o- mmfiufiawu: «gm evom.h mnuo.o wNHw.n wmmu.Ho newpUSMumaou mz Hmmu.a wmao.o ONow.¢H oovw.mN mowhpmswcfl xmewhm “Ahumswcfi Ehmmaoz mmHNE om-o "wouufiso «gm mHmo.N- vao.o- NNmo.OH wooo.0N- no>o Una mmHHE HmH «gm vomm.o- mHoo.o- mmHm.w mHHm.mm- mmHHE omH-HOH «gm oemm.HH- vmo~.o- wuon.m mmNm.oo- moHHE ooH-Hm ” pnmwcomouqH -mficwfim «pom mo muchho .wum scammmhmmm N .wosafipcou--.u oHnme 65 The insignificance of short-run income changes in motivating long-distance transfers seems quite reasonable if migrants place greater emphasis on anticipations of higher long-run earnings,.improved working conditions, and better housing, etc. Of particular interest, on a theo- retical as well as practical basis, is anticipation of higher long-run earnings. Following the terminology of Friedman and Stigler, let the average income of'a family over a period of years be its permanent income and the deviation of its current income from this level be its transitory component of income.33 Then a family with a large negative transitory component would be put in a low current income class but with a large positive transitory component the same family would be put in a high current income class. But the family would be foolish to vary expenditures solely on the basis of widely fluctuating annual incomes. Moreover, as an empirical matter family incomes.do undergo fluctuations of substantial magnitude, although such fluctuations are not consistent except perhaps between seasons of a year. For these reasons, budgeting studies of a group of families at any one time normally tend to understate the responsiveness of expenditures to changes 33Milton Friedman, A Theory of the Consumption Func- tion, (New York: National Bureau of Economic ResearCh, 1337), and George Stigler, The Theory of Price, (New York: The Macmillan Company, 1966). 66 in permanent income. In other words, current measured income is inadequate as an indicator of long-run income status and as such more permanent or long-run income changes would be expected to be of significant importance in explain- ing migration. Contrary to that hypothesized, the coefficient for the race variable was not significantly different from zero, indicating that, for the nation, Negroes were not more willing to move greater distances in off-farm transfers than nonNegroes. In view of this insignificant result, one of two conclusions is apparent. First, although Negroes his- torically appear to have been more responsive to distance in off-farm occupational mobility than nonNegroes, the distinction is not statistically relevant in predicting who will be more willing to migrate long distances. Perhaps the gross relationship is not due to race at all but due to interactions between other variables. Second, since the parameter estimation in the national equation involved the aggregation of cross-sectional data, the true significance was negated by the dominating influence of migration patterns of Negroes-particularly from other than the South. Nevertheless, this result does not lend itself well to interpretation, expecially since approximately 62 per cent of the Negroes in the sample originated in the South. Moreover, the massive migration of Negroes to center cities of metropolitan areas, as opposed to nonNegroes, is well 67 documented. John F. Kain and Joseph J. Persky found that 58 per cent of Negroes born in the South Atlantic division and now living elsewhere live in the four northeastern SMSA's greater than one million (Buffalo, New York, Phila- delphia, and Pittsburgh).34 Similarly, they found that 40 per cent of the Negro migrants from the East South Central division moved to the five East North Central SMSA's greater than one million (Chicago, Detroit, Cincinnati, Cleveland, and Milwaukee) while 36 per cent of the same group from the West South Central division live in the four Pacific SMSA's greater than one million (Los Angeles, San Diego, San Francisco, and Seattle). In addition to this mass migration, it is well known that nonfarm economic opportunities for Negroes in the South have been limited, opportunity costs for Negroes in farming are low, and rural Negro birth rates have been high. Hence, these pressures would seem to add to the long-standing link between the rural Negro of the South and migration to large northern metropolitan centers. The relationship between distance and increasing age was generally consistent with the inverse relationship hypothesized. It will be recalled that each estimated 34John F. Kain and Joseph J. Persky, "The North's Stake in Southern Rural Poverty," in Rural Poverty in the United States, Report by the President‘s National Advisory Commission on Rural Poverty (Washington, D.C.: Government Printing Office, 1968), Ch. 17, pp. 288-308. 68 parameter (coefficient) of a category of the age variable was computed by subtracting the actual value of the coeffi- cient of the omitted category, which in this case was 24 and under, from the actual value of the estimated parameter. Therefore, from the sign of the regression coefficients, off-farm migrants 25-34 years of age were more willing to move.greater distances than those persons 24 and under. The estimated change in distance migrated on theaverage, holding other variables constant, when one moved from the 24 and under to the 25-34 age category, was approximately 40 miles. The results of this comparison between the 24 and under and the 25-34 years of age category was somewhat con- trary to that expected. Normally, work experience and locational preferences would be greater for the older group as Opposed to that of the 24 and under category. This should hold true irrespective of race and farm employment status. Thereby, psychic and farm Opportunity costs would be expected to be higher for the older migrants. In addi- tion, persons in the older category normally have more accumulated belongings and family responsibilities which would translate into greater money costs--food, lodging, and transportation--associated with migration. But appar- ently, as viewed by the older group, the expected length of the pay-off period from nonfarm employment for regaining these investments in migration was sufficient. Maturity 69 short of 35 may have been an advantage in exposing a wider range of potential nonfarm employment opportunities with better earnings and future advancement possibilities. More- over, for the 24 and under category, the possibility of commuting to nonfarm jobs, the reality of various military obligations, and the existence of dependence on other wage earners may have been true negative factors to long-distance migration. From the initial calculations given in Table 7 above, both the 35-44 and the 4S and over age groups had coefficients that were not significantly different from that of the omitted category, 24 and under. However, this does not mean that these categories would not be significant when compared with some other omitted category. Further t statistics (tb = SE) were calculated (results not shown), with the b obtained by subtracting the value of the regres- sion coefficient of the desired omitted category from the value of the regression coefficient of the retained category, and with the Sb obtained by taking the square root of the sum of the appropriate coefficient variances less twice the appropriate covariance term. From such computations, the 35-44 age category was.found to be significantly different from the 25-34 age group, and the 4S and over age group was significantly different from the 35-44 age category. In both cases the calculated coefficients had negative signs. This result reinforced the hypothesized age-distance 70 relationship by indicating that older persons who leave agriculture do so only with the opportunity of nonfarm employment located near their current employment. For the employment status of farm individuals, as hypothesized, single and multiple job wage workers were found to be more responsive to distance than categories of multiple_and single job farm operators. With the exception of the farm operator-nonfarm wage worker category, all farm employment categories were significant in explaining dis- tance migrated. The estimated parameter of the farm wage work and nonfarm wage employment category was of the great- est magnitude, indicating that, other things the same, persons in this group were willing to migrate over.100 miles further than persons who were farm operators only. It appears that persons with this type of multiple employ- ment status are in the best possible position to success- fully migrate long distances to exclusively nonfarm employment. Such long-distance transfers would be expected to be the result of lower relative costs of migration than self-employed individuals or the simple availability of nonfarm employment through present nonfarm wage jobs. Alternatively, for the farm wage work only category, long- distance moves would be expected to be primarily a function of lower migration costs. The estimated change in distance migrated on the average, holding other variables constant, 71 by moving from the farm self-employed category to the farm wage work only category, was approximately 81 miles. As previously mentioned, farm operators with non- farm self—employment were the least likely to migrate long distances. This finding is of particular interest in view of the method of computation with the omitted category, farm self-employment only, and the resulting negative sign of the multiple job category. Several reasons are immediate. In all probability, farm operators with nonfarm self- employment left agriculture with the intention of devoting full time to their present nonfarm jobs.. Quite likely these nonfarm jobs were within commuting distance of their farm businesses. Consequently, their money and psychic costs were minimized by short-distance migration. Moreover, the nonfarm component would be accompanied by some accumula- tion of experience as well as an added source of money income. And, additional investments in transportation, tools, office facilities, and clothing might have been required. Hence, farm operators with some form of nonfarm self-employment would have high opportunity costs relative to the employment position of single-job farm operators. As hypothesized, all categories of farm earnings prior to migration were found to be negatively related to distance. However, as farm earnings increased, holding .other variables constant, the negative magnitude of the coefficients of the categories increased, indicating a 72 declining response to distance migrated as farm earnings became larger. Only the coefficient for the $1,200-$l,799 category was not significantly different from the omitted category, farm earnings between $0-$l,199. In view of the method by which each of the other parameters was estimated, persons with this lowest level of farm earnings were most responsive to distance in off-farm mobility. A number of reasons are plausible. Hathaway and Perkins found that farm-nonfarm occupational mobility does not necessarily close the money income gap between the poor and the better off.35 Perhaps some potential migrants had similar expectations. In addition, on the average, as farm earnings increase, the greater the opportunity costs associ- ated with farm employment become and thereby the less the willingness of a potential mover to migrate. Moreover, the low level of earnings in farming when weighed against monetary costs of migration could have magnified the nega- tive response, although given that individuals with $0- $l,199 were most responsive to distance, there is reason to believe that earnings from farm employment are not good indicators of the ability of potential off-farm migrants to bear the monetary costs associated with distance. As hypothesized, holding other independent vari- ables constant, distance migrated was inversely related to 35Hathaway and Perkins, ”Occupational Mobility and Migration from Agriculture," Ch. 13, pp. 185-237. 73 distance from an SMSA. The omitted category was 0-50 miles and as indicated by the negative sign of the remaining coefficients, persons coming from this category were the most responsive to distance in migration. Surprisingly however, in view of previous findings, for the three remaining categories, as the mileage from an SMSA increased the negative magnitude of the corresponding coefficients decreased. This means that, aside from the 0-50 category, farm employed persons located further away from SMSAs were the most likely to migrate long distances. Perhaps the best explanation for such findings lies in the functioning of the SMSA itself in the labor market. It is well known that nonfarm employment is much more avail- able in or near SMSAs and access to the national labor market more readily obtained. Hence, persons within 50 miles of the SMSA are best able to participate immediately and directly in the national labor market. Moreover, for people outside the 50 mile area, the SMSA essentially serves as a regional center through which farm migrants are transferred to nonfarm employment. Therefore, assuming the off-farm migrants involved moved to and/or through their nearest SMSAs, the further a migrant from his nearest SMSA, the greater the distance he would have to migrate in order to establish exclusively nonfarm employment. With respect to the nonfarm industry entered, four categories were found to be significantly different from 74 manufacturing, the omitted category, in explaining distance migrated. Off-farm migrants were found to move greater distances to construction and service industries and to government, with construction having the largest positive sign followed by government and service industries, respec- tively. Manufacturing was next in importance. Primary industries, utilities, and wholesale and retail trade were a not significantly different from manufacturing in explain- ing distance migrated. As hypothesized, construction and I r 1 government had the largest positive significant regression i . coefficients, indicating that persons in the former group were willing to migrate over 61 miles further than persons in the manufacturing category while persons in the latter category were willing to migrate an additional 40 miles. This result is quite plausible since for construction industries, discounted real wage rates were relatively high. Moreover, employment opportunities with all local, state, and federal levels of government were generally relatively good. The finding for service industries (finance, insurance, and real estate) was contrary to that expected when compared to the manufacturing category. The coeffi- cient magnitude was the smallest of the significant categories. The positive relationship would seem to indi- cate, however, that many migrants possess the skills necessary for this type employment, especially in view of 75 the less important role of manufacturing industries. The possibility of high earnings might also have been an important factor. From the coefficient magnitudes for the industry variable, there would appear to be four major clusters. These are: (1) utilities and wholesale and retail trade, (2) primary and service industries, (3) government, and (4) construction. Further t statistics were calculated to examine any significantly different responses with distance migrated. From such computations, service indus- tries were found to be significantly different from wholesale and retain trade, government was not signifi- cantly different from service industries, but construction was found to be significantly different from government in explaining distance migrated. These results further rein- forced the hypothesized distance-industry relationship by indicating that long-distance moves were principally to jobs in construction, manufacturing, government, and pri- mary and service industries. The Regions.--A summary of the analyses for the 36 As in the case five major regions is shown in Table 8. of the national equation, little of the variation in the dependent variable distance migrated was explained by the 36Specific regression results are contained in Appendix C. "1 76 mmH.Hm-o$ "emouueo - m2 m2 - - Hm>o use ooo.mm - m2 m2 - mz mam.mm-ooq.me mz - m2 - mz mam.mw-oow.fiw m2 m2 m2 m2 m2 mmh.fiw-oom.fiw "mmcflnhmm Show xaco paoexoamso-maom 59mm "wouuNEo uaoeonmEo-mHom m2 m2 m2 m2 m2 Ehmmao: use ucoeonmEo-mHom Seam ucoeonmEo 0mm: m2 m2 m2 m2 m2 Eammcoa wow pcoezofimso-mHom Show pooE>onso + + + + wz 0mm: Ehquo: can xhoz ommz Ehmm + + m2 + m2 haco x903 6mm: Egon "mswmum ucoeonQEm Ehmm some: can VN "oopufieo m2 m2 m2 wz mz po>o one we m2 m2 - m2 m2 ev-mm + m2 m2 + m2 vm-mN Hom< ohmozaoc ”wopuweo mz . + + + Ohmmz “ovum m2 m2 m2 m2 m2 mmamco «soueH m2 + + + m2 Hawumflou amok mqfimHm npsom Hmuucou ammo moanmfihw> poowaomowaH :pyoz -nppoz muwpmflpouomhmgo UNEW .oo-ummfl .cofimoh kn .oopmhmfle ooampmfiw spa: wopmwoommm :ouo wow ofiammuwosov mo nofipmsco :ONmmohmoh mo muHSmom .w oanme 77 p0: mm: ucowofiwmoou oz“ momma mz mo. onp um pcmuflmw:m«m mm: magmnoflpwfioh 0:“ hoauoaz oumoflwcfi mcmfim one .Ho>oH mo. one we open Eosm unoyommfiw xfipawUNMszHm .mflam:0wpmfioh ogu mo ohspwa on“ one No>oH H mawhspommszmz "coupfiao kgwuwfifie + m2 m2 m2 m2 wow .wofiwfimmmfiucs .azocxcb m2 m2 m2 m2 + pqoecno>ou mz - m2 ,m2 + mofihumswow oofl>uom mz - m2 m2 m2 owmhu Hflmuou was mammoaogz mz - m2 m2 m2 mowuflawub + m2 m2 + + :ONuUSHHmapu m2 m2 m2 + m2 mofipumswdfi xpmaflhm ”xnpmswcH Eummzoz moHHE om-o "copuweo + - + + m2 Ho>o pom mofifie HmH - - - - mz moHNE omH-HOH - - - - mz moHNE coauam " ucownomoqu :unoz -gpnoz .eoseupeoo--.m manae 78 independent variables in each regional equation. Despite this shortcoming, the findings were useful because of the additional information obtained to support the results of the national equation. Short-run income changes were not significantly related to distance migrated in any region, as in the case of the nation as a whole. This suggests that anticipations of greater short—run income changes are.not relevant in explaining long-distance migration irregardless of region of origin or demographic and economic characteristics of persons migrating. The Negro race category was significant in all regions except the West and all coefficients were positive except for the Plains region. This was encouraging in view of theinsignificant relationship found in the national equation. It would appear that migrants who were most responsive to distance were nonNegroes from the Plains or Negroes from.the Northeast, North Central, and the South. However, of all Negroes included in the sample from the latter three regions, over 75 per cent had some type of farm employment in the South prior to migration. Conse- quently, there is good reason to believe that many of the farm individuals who migrated long distances to exclusively nonfarm employment were Negroes from the South. The regional breakdown of the relationship between distance migrated and increasing age was similar to that 79 found in the national equation. For the 25-34 category, as for the nation as a whole, a positive relationship was found in the North Central and West regions. In other words, this more mature group of individuals was_willing to migrate greater distances than the younger 24 and under group. Perhaps the former, through maturity and work experience, were more employable than the latter in better nonfarm jobs at greater distance from their places of origin. However, for the South, the 25+34 age category was not significantly different from the 0-24 omitted category, and an inverse but significant relationship occurred for the 35-44 age category. This indicates that for this region all migrants over 24 years of age were resistant to distance in migration. For the farm employment status of farm individuals, single and multiple job farm wage workers were found to be willing to incur distance in off-farm transfers in the North Central, Plains and West regions, while the farm wage work-nonfarm wage employment category had the only signifi- cant coefficient in the South. Without exception for each region, multiple wage workers were the most responsive to distance. As previously argued, persons in this group were in a good position to successfully establish exclusively nonfarm employment. Perhaps the real input of the original 80 nonfarm wage.jobs was the provision of the necessary basis of experience. The relationships found in all regions between farm earnings prior to migration and distance migrated concurred with the inverse relationship of the national equation. All categories were found to be negatively related to dis- tance, indicating that persons with $0-$l,199 in farm earnings, the omitted category, were willing to migrate longer distances. The South had no significant coeffi- cients while only the North Central region had farm earnings with significant coefficients indicating a negative rela- tionship with distance migrated. These findings of a declining response to distance migratedas farm earnings became greater would seem to confirm the argument that for these persons Opportunity costs in farming were large and thereby the incentive to migrate to exclusively nonfarm employment small. The regional results for the association between distance migrated and distance from an SMSA were somewhat mixed. All significant coefficients were negative (indi- cating persons 0-50 miles from an SMSA were responsive to distance) with the exception Of the 151 miles and over category. Contrary to the results for the national equa- tion, a positive coefficient for this latter category was found in the North Central, South, and West regions. This implies that the further a person from an SMSA, even 81 considering the 0-50 category, the greater the willingness to migrate long distances. As with the equation for the nation as a whole, the only plausible explanation seems to be that SMSAs provide easy access to the national labor market. That is, the further a potential migrant from his nearest SMSA, the greater the distance he would have to move in order to establish exclusively nonfarm employment, whether in the SMSA itself or at a greater distance through the SMSA to another national employment center. The relationship between nonfarm industry entered and distance migrated varied considerably by industry and by region. For the South, no category was significantly different from manufacturing, the omitted category, in explaining distance migrated. The government category was important only for migrants from the Northeast. Utilities, wholesale and retail trade, and service industries had negative coefficients for the Plains region, indicating that manufacturing was more important in attracting long- distance migrants. As Opposed to this result, for the Northeast region the service industry category had a posi- tive sign. The result for the construction and primary indus- tries categories was similar to that for the national equation in that a significantly different relationship from the manufacturing category was found in some regions. The coefficient for service industries was significant but 82 positive only for the Northeast region. The construction category coefficient was positive and significant in the Northeast, North Central, and West regions. CHAPTER V SUMMARY AND CONCLUSIONS Summary The basic Objective of this inquiry was to investigate the relationship of various economic and demographic characteristics Of actual Off-farm movers to the distance they migrated in transferring from farm to nonfarm employment. Three general assumptions were made. First, it was assumed that the migration process requires investments in the form Of both monetary and nonmonetary costs which increase with the distance Of migration. Second, the motivation for incurring these costs was the expectation Of better economic well-being from nonfarm employment. Finally, it was believed that from the results Of the analysis the impact Of migration on both sending and receiving communities could be more clearly understood and thereby enable a greater understanding of the link between rural and urban problems. The model employed in this study was a multiple linear regression equation, both for the nation and by region. The data employed were derived from the one per cent continuous Work History Sample maintained by the 83 84 Social Security Administration. Distance migrated was_the dependent variable. Income change, race, age, farm employ- ment status, farm earnings, distance from an SMSA, and nonfarm industry were the independent variables. With the exception Of income change, all independent variables were categorized and entered the regression equations as dummy variables. The hypotheses advanced were: 1. income change and distance are directly related; 2. Negroes are willing to migrate further than nonNegroes; 3. distance migrated and increasing age are inversely related; 4. categories of farm wage workers are more responsive to distance than classes of farm Operators; 5. distance migrated is an inverse function Of farm earnings; 6. distance migrated and distance from an SMSA are inversely related; and 7. distance is a direct function of primary, manufacturing, and construction industries and to govern- ment, but is inversely related to utilities, service industries, and wholesale.and retail trade. From the multiple regression analyses for the nation and the regions, the major findings were as follows: 85 1. Income change was not significantly related to distance migrated, both in the national equation and by region. Apparently short-run compensation for monetary and Opportunity costs was not important in inducing long- distance migration. 2. For the nation as a whole, the race variable (Negro-nonNegrO) was not significant in explaining long- distance migration. However, on a regional basis, the Negro race category was significantly different from the nonNegrO category in all regions except the West and all coefficients were positive except for the Plains region. This contradiction with the national equation was attrib- uted to aggregation along cross-sectional lines. Negroes from the South, representing approximately 62 per cent Of all Negroes in the sample, were found to be most responsive tO distance. Hence, there was good reason to believe that many farm workers who migrate long distances to exclusively nonfarm employment are Negroes from the South. 3., Long-distance migration for all equations was found to be inversely related to increasing age, especially for farm individuals 35 years of age and over. Surprisingly, migrants 25-34 were more responsive to distance than per- sons 24 and under. It would appear that maturity and expectations regarding the sufficient length of the payoff period for recouping migration investments and better ‘EEE”“ 86 nonfarm job opportunities motivated the greater response by the older group. 4. For the national equation as well as by region, single and multiple job farm wage workers were found to be more responsive to distance than categories Of multiple and single job farm Operators. Individuals with farm wage work # and some form Of nonfarm wage employment were willing tO F1 migrate longer distances while farm operators with nonfarm self-employment were the least likely to do so. Moreover, single-job farm Operators were more responsive to distance d than multiple-job self-employed individuals. By virtue Of high Opportunity costs, in all probability farm Operators with nonfarm self-employment left agriculture with the intention of devoting full time to their present nonfarm jobs. 5. All categories of farm earnings prior to migra- tion were found to be negatively related to distance, both for the nation and by region. But the negative magnitude of the coefficients Of the categories increased as farm earnings increased, indicating a declining response to distance migrated as farm earnings increased. Even though the farm earnings reported were relatively low, this result would suggest that the negative incentive came from greater Opportunity costs in farming as earnings increased and thereby the greater the risks from migration. But, again in view Of the low levels Of earnings, this result raises 87 a serious.question as to the adequacy of current farm earnings as an economic indicator of farm well-being. 6.7 Distance migrated was inversely related to distance from an SMSA in the national equation when com- paring the 0-50 mileage category with the other categories. However, beyond the 50 mile belt, farm employed persons located further away from SMSA's were the most likely to migrate long distances. Moreover, for the North Central, South, and West regions, farm employed individuals located 151 miles and over from an SMSA were the most responsive to long-distance migration. Since SMSA's function as regional centers in the national labor market, it would appear that persons located near such centers are able to participate directly in the national labor market. For potential migrants located some distance from the SMSA's, the greater the distance Of migration required for the establishment of exclusively nonfarm employment either in the regional center (SMSA) or in more distant national labor market centers. 7. The relationship on a national and regional basis between distance and nonfarm industries entered was generally consistent with that hypothesized. For the nation, off-farm migrants were found to move greater distances to construction, manufacturing, service industries, and to government, with construction having the largest positive sign. The result for service industries was surprising, 88 especially in view Of the general work skills of the migrants involved. These findings were consistent by regions, except that the primary industries category was positively significantly different from manufacturing in the North Central region and service industries were nega- tively significant in the Plains region. Unfortunately, no industry category was significantly different from manufacturing in the South. Impact on Urban Areas From an historical perspective, given the foregoing summarized results Of the regression analysis, the impact Of rapid Off-farm migration on urban areas can be more clearly explained. The lack of adequate economic preparation of many off-farm migrants for urban employment was suggested by the demographic and economic relationships established by this study. In particular, since long-distance migration was found to be primarily associated with the young, Negroes, low income persons, and Off-farm movers initially securing blue-collar jobs (construction, manufacturing, and service industries), the probability Of continued nonfarm employ- ment would be expected to be small. In general, relative to their urban counterparts, farm youth complete fewer years of formal schooling and receive a somewhat poorer quality education. Moreover, via vocational agriculture, 89 job training has tended to emphasize manual labor skills. Consequently, Off-farm migrants generally are able only to Obtain the lowest paying jobs within each industrial cate- gory and normally would tend tO be laid Off first in the event of an economic decline. All of these contribute to possible unemployment among the relatively poorly educated and poorly skilled and ultimately lead to frustration, f} discontent, and poverty. Impact on Rural Areas The impact Of massive off-farm migration on rural areas has created many problems. Results of this study show that long-distance migration is positively associated with the young and categories Of farm wage and nonfarm wage workers. Young Off-farm movers leave with expectations of a better quality life, taking with them potential produc- tion and consumption necessary for a viable rural economy and the maintenance of a livable community. From the shift in the demand for farm labor, and due to the selectivity of the Off-farm migration process, many individuals remaining in rural areas are unemployed and/or underemployed. This has led to poverty for many. Perhaps the most drastic negative impact Of massive out-migration from rural areas has been the con- tinual economic deterioration. Since many individuals who migrate have nonfarm wage jobs, goods and services produced 90 from such efforts stagnate. Moreover, many businesses are forced to discontinue functioning, thereby creating more unemployment. Or, due to a declining volume Of goods and services demanded, prices eventually must be increased and passed on to the remaining rural population. However, such a situation does not exist for but the very short-run, since individuals are induced to substitute in favor Of items purchased elsewhere. And economic deterioration continues. Since predominately the young leave for urban areas, many Old people remain, to exist on declining busi- ness incomes or pensions. Private businesses and local governments experience diseconomies in the production Of various services. Property tax bases, both farm and non- farm, become totally inadequate for Operation of public services and the remaining Older individuals involved strongly resist any increase in tax rates. Policy Implications Some Off-farm migration will continue to occur in response to natural economic phenomena for certain areas Of the United States. Many rural communities are not expected to continue to exist. However, as out-migration tO urban areas is expected to continue, it would be desir- able to assist out-migrants to make successful transfers. This does not mean that Off-farm movements should be 91 encouraged by large publically provided subsidies, since in many cases the real economic position Of many rural poor is superior to that Of their urban counterparts. However, meaningful public policies should receive further attention, particularly those policies designed to encourage coopera- tion between rural and urban areas. Given that multiple-job farm wage earners have been most willing to migrate long distances, a modest beginning could be the provision Of detailed occupational and educational information regarding skill requirements, job Openings, employer contacts, and social conditions in cities as well as limited employer subsidies for job train- ing. Such a program, designed particularly for rural people, could probably best be administered through the Cooperative Extension Service associated with most land- grant institutions in conjunction with local and state employment agencies. Hopefully, such a program would enable successful nonfarm job establishment and adjustments to urban living. Notwithstanding the merits Of the above stOp-gap proposal, a lasting solution can only come from within rural communities. Policies must be designed to assist development Of rural economies and improvement Of rural living conditions to discourage further massive off-farm Inigration and concentrations in large metropolitan areas. 92 Rural community growth demands a heterogeneous economy if any degree of development is to be achieved. Various industries must be induced to Operate in rural areas to effectively utilize out-migrants from the agri— cultural sector and supplement the remaining agricultural sector. But in order to make rural communities more in- viting tO those moving into the cities, to encourage those who left the agricultural sector to return, and to attract industrial development, rural areas must become better places to live. Public policies must be devised to promote the advantages of rural areas and communities over large urban centers. The country must be advertized as a desirable place to live and rear a family. Rural communities must be revitalized in many ways. Public facilities must be modernized. Schools must be ade- quately staffed and equipped. Residential housing must be made available of sufficient quality and quantity. And recreation facilities must be modernized or constructed anew. Such revitalization would attract new industry and Open new prospects and Opportunities for rural areas to grow and prosper. New national growth policies to develop the resources Of rural areas will require much local initiative and will cost money. Local governments, schools, and other formal and informal organizations must be reorganized to 93 permit efficiency in the delivery Of all needed public services. For long-run success, such a reorganization must involve local leadership. Credit institutions in addition to those in current existence must be developed to provide funds for rural industrial development and for financing improvements in public services. Finally, federal and state expenditures to subsidize rural economies must be ‘1 made, perhaps at the expense Of existing urban programs Of low social value. BIBLIOGRAPHY BIBLIOGRAPHY Bachmura, F. T. "Migration and Factor Adjustment in Lower Mississippi Valley Agriculture: 1940-1950." J. Farm Econ., Vol. 38 (November, 1956), pp. 10247TO42. Baumgartner, H. W.- "Potential Mobility in Agriculture: Some Reasons for the Existence of a Labor-Transfer Problem." J. Farm Econ., Vol. 47 (February, 1965), pp. 74-82. Ben-David, Moshe. "Farm-Nonfarm Income Differentials, U.S., 1960." Unpublished Ph.D. dissertation, Michigan State University, 1967. Bishop, C. E. "Underemployment of Labor in Southeastern Agriculture." J. Farm Econ., Vol. 36 (May, 1954), pp. 258-272. . "Economic Development and Adjustments in South- eastern Low Income Agriculture." J. Farm Econ., Vol. 36 (December, 1954), pp. 1146-1160. "Economic Aspects of Changes in Farm Labor Force." Labor Mobility and Population in Agri- culture. _Ames, Iowa: IOwa StateIUniversity Press, 196I, pp. 36-49. "Dimensions Of the Farm Labor Problem." Farm Labor in the United States. Edited by C. E. Bishop. New YOrk: Columbia University Press, 1967, pp. 1-17. "The Urbanization Of Rural America: Implica- tions for Agricultural Economics." J. Farm Econ., Vol. 49 (December, 1967), pp. 999-1008. Blank, Uel. "O.A.S.I. Data Of the Farm Labor Force." Unpublished Ph.D. dissertation, Michigan State University, 1960. Bogue, Donald J. "A Methodological Study of Migration and Labor Mobility." Scripps Foundation Studies in Population Distribution, NO. 4, June, 1952. 94 95 Boyne, David H. "Changes in the Income Distribution in Agriculture." J. Farm Econ., Vol. 47 (December, 1965), pp. 121371224. Bryant, W. Keith. "Causes of Inter-County Variations in Farmers' Earnings." J. Farm Econ., Vol. 48, Part I (August, 1966), pp. 557-5'7. Diehl, William D. "Farm-Nonfarm Migration in the Southeast: A Costs-Returns Analysis.” J. Farm Econ., Vol. 48 (February, 1966), pp. l-ll. Freund, John E. Mathematical Statistics. Englewood Cliffs, I! New Jersey: Prentice-Hall, Inc., 1962. ‘1 Friedman, Milton. A Theory Of the Consumption Function. . New York: National Bureau of Economic Research, 6 1957. J Fuller, Varden. "Factors Influencing Farm Labor Mobility." Labor Mobility and Population in Agriculture. Ames, Iowa: Iowa State University Press, 1961, pp. 28-35. , and Beale, Calvin L. "Impact Of Socioeconomic Factors on Farm Labor Supply." J. Farm Econ., Vol. 49 (December, 1967), pp. 1237-1243. Gallaway, Lowell E. "Labor Mobility, Resource Allocation, and Structural Unemployment.” Amer. Econ. Review, Vol. LIII (September, 1963), pp. 694-716. Interindustry Labor Mobility in the United States: 1957 to 1960. Social Security Administra- tion Research Report'NO. 18. Washington, D.C.: Government Printing Office, 1967. . "Mobility of Hired Agricultural Labor: 1957- 1960." J. Farm Econ., Vol. 49, Part I (February, 1957), pp. 32-51. . "Geographic Flows Of Hired Agricultural Labor: 1957-1960." Amer. J. Agr. Econ., Vol. 50 (May, 1968), pp. 199-212. Gisser, Micha. "Needed Adjustments in the Supply of Farm Labor." J. Farm Econ., Vol. 49 (November, 1967), pp. 806-815. Guither, Harold D. "Factors Influencing Decisions to Leave Farming." J. Farm Econ., Vol. 45 (August, 1963), pp. 567-5760 96 Gustafson, Robert L. "The Use and Intrepretation of 'Dummy Variables' in Regressions." Michigan State University, Department of Agricultural Economics, Mimeo, January, 1962. Hathaway, Dale E. "Migration From Agriculture: The Historical Record and Its Meaning." Amer. Econ. Review, Vol. L (May, 1960), pp. 379-391. Government and Agriculture: Economic Policy in a Democratic Society. NeinorK: Macmillan and Company, 1965. Urban-Industrial Development and Income Differ— entials Between Occupations." J. Farm Econ., Vol. 46 (February, 1964), pp. 56-66. . "Occupational Mobility From the Farm Labor Force." Farm Labor in the United States. Edited by C. E. BiShop. New York: Columbia University Press, 1967. , and Perkins, Brian B. "Occupational Mobility and Migration from Agriculture." Rural Poverty in the United States, Report by the‘President's Nainnal AdvisorinOmmission on Rural Poverty. Washington, D.C.: Government Printing Office, 1968. . "Farm Labor Mobility, Migration, and Income Distribution." Amer. J. Agr. Econ., Vol. 50 (May, 1968), pp. 542-353} , and Waldo, Arley D. Multiple JObhOlding by Farm Operators. Michigan Agr. Exp.iSta. Bul. 5, 1964. Heady, Earl O. "Facilitating Movements Of Labor Out of Agriculture--Discussion." Amer. Econ. Review, Vol. L (May, 1960), pp. 413-415. "Preface.” Labor Mobility and Population in Agriculture. Ames,'Iowa: IOwa State University Press, 196i, pp. v-vi. . Agricultural Policy Under Economic Development. Ames, Iowa: Iowa State University Press, I962. Johnson, D. Gale. "Functioning Of the Labor Market." J; Farm Econ., Vol. 33 (February, 1951), pp. 75-87. 97 . "Policies and Procedures to Facilitate Desirable Shifts of Manpower." J. Farm Econ., Vol. 33 (November, 1951), pp. 722-729. "Comparability Of Labor Capacities Of Farm and Nonfarm Labor." Amer. Econ. Review, Vol. XLIII (June, 1953), pp.’296-3l3. "Labor Mobility and Agricultural Adjustment." Agricultural Adjustment Problems in a Growing Economy. Edited by’EarI'O._Heady, et a1. Ames, owa: Iowa State College Press, 1958. . ”Policies to Improve the Labor Transfer Process." Amer. Econ. Review, Vol. L (May, 1960), pp. 403-412. Johnson, Paul R. "Labor Mobility: Some Costs and Returns." Rural Poverty in the United States. Report by the Piesident‘s National Advisory COmmission on Rural Poverty. Washington, D.C.: Government Printing Office, 1968. Johnston, J. Econometric Methods. New York: McGraw-Hill Book CO., Inc.,41963. Kain, John F. and Persky, Joseph J. ”The North's Stake in Southern Rural Poverty." Rural Poverty in the United States. Report by the President‘s National Advisory Commission on Rural Poverty. Washington, D.C.: Government Printing Office, 1968. Kmenta, Jan. Elements Of Econometrics. New York: The Macmillan Company, 1971. Maddox, James G. "Private and Social Costs Of the Movement of People Out of Agriculture." Amer. Econ. Review, Vol. L (May, 1960), pp. 393-402. Perkins, Brian B. "Labor Mobility Between the Farm and the Nonfarm Sector." Unpublished Ph.D. dissertation, Michigan State University, 1964. , and Hathaway, Dale E. The Movement of Labor Between Farm and Nonfarm Jobs. Michigan Agr. Exp. Sta. Bul. 13; 1966. Reder, Melvin W. "Facilitating Movements of Labor Out Of Agriculture--Discussion." Amer. Econ. Review, Vol. L (May, 1960), pp. 415-416. 98 Ruttan, Vernon W. "The Human Resource Problem in American Agriculture." Farmingo Farmers, and Markets for Farm Goods. Edifed’by Karl A. Fox, et ai. New York: COmmittee for Economic Development, 1962. Schuh, G. Edward. "An Econometric-Investigation Of the Market For Hired Labor in Agriculture." J. Farm Econ., Vol. 44, Part I (May, 1962), pp. 3U7-32I. Schultz, T. W. Agriculture In An Unstable Economy. New York: McGrawiHill Book Company, Inc., I945. Economic Organization Of Agriculture. New York: McGraw-Hill Book Company, Inc., 1953. . "Investment In Human Capital." Amer. Econ. Review, Vol. LI (March, 1961), pp. 1-17. "Reflections On Investment In Man." J. Pol. Econ., Vol. LXX, Part II, Supplement (chOEer, 19625, pp. 1-8. Sjaastad, Larry A. "Occupational Structure and Migration Patterns." Labor Mobility and Population in Agri- culture. Ames, Iowa: Iowa State University Press, I96I, pp. 8-27. "The Costs and Returns of Human Migration." J. Pol. Econ., Vol. LXX, Part II, Supplement (October, I962), pp. 80-93. Stigler, George. The Theogy of Price. New York: The Macmillan Company, 1966. Suits, D. B. "Use Of Dummy Variables in Regression Equa- tions." J. Am. Stat. Assoc., Vol. 52 (December, 1957), pp. 548-551. Tyrchniewicz, E. W., and Schuh, G. E. "Regional Supply Of Hired Labor in Agriculture." J. Farm Econ., Vol. 48 (August, 1966), pp. 537-556. United States Bureau Of the Census. "Mobility of the Popula- tion Of the United States: March 1966 to March 1967." Current Population Reports, Series P-20, NO. 171. Washington, D.C.: Government Printing Office, 1968. "Lifetime Migration Histories of the American People." Current Population Reports, Series P-23, NO. 25. Washington, D.C.: Government Printing Office, 1968. 99 . "Social and Economic Conditions of Negroes in the United States." 'Current Population Reports, Series P-23, NO. 26. Washington,iD.C.: Government Printing Office, 1968. . "Farm Population." Current Population Reports, Series P-27, NO. 42. Washington, DTC.: Government Printing Office, 1971. United States Department of Agriculture. Farm Population-- Net Mi ration From the Rura1--Farm Population, I94U-55, by GladysKi Bowles, Statistical Bulletin No. I76. Washington, D.C.: Government Printing Office, 1956. . Farm Population Estimates For 1910-62, by Vera J. Banks, Calvin L.iBea1e, and Gladys’K. BOwles, ERS-l30. Washington, D.C.: Government Printing Office, 1963. Changes in Farm Production_and Efficiency, StafiStical Bulletin NO. 233. “Washington, D.C.: Government Printing Office, 1970. Farm Income Situation. Washington, D.C.: Government Printing Offlce, July, 1971. United States Department of Labor. Monthly Labor Review. Washington, D.C.: Government”Printing Office, December 1959 and December 1960. Waldo, Arley D. "The Off-farm Employment Of Farm Operators in the United States." Unpublished Ph.D. disserta- tion, Michigan State University, 1962. APPENDICES APPENDIX A DEFINITION OF TERMS APPENDIX A DEFINITION OF TERMS Migration.--Defined as a change in location, the smallest change Being from one county to another, in moving from some form of farm employment coverage to exclu- sively nonfarm employment. Occupational mobility and Off-farm mobility.-- Both used synonymously with migrafien. Distance migrated.--The distance between the county Of farm employment and the county Of nonfarm employment, measured by using the Census coordinates for the population centers of the counties. Employment status.--Defined as either farm or non- farm. Farm Employment.--Persons who are exclusively farm wage workers, who are exclusively self-employed farm Opera- tors, Or who combine one Of these categories with some form Of nonfarm employment. Nonfarm Employment.--Persons whose employment, either wage, salary, or self-employment, was exclusively in nonfarm industries. Industry of nonfarm employment.--The industry of employment from which the individual Obtained the highest earnings during the year after migration. 100 APPENDIX B REGIONAL DIVISION OF STATES APPENDIX B REGIONAL DIVISION OF STATES Northeast.--Maine, Vermont, New Hampshire, Massachusetts, Connecticut, Rhode Island, New York, New Jersey, Maryland, Delaware, Pennsylvania, District Of Columbia. North Central.--Wisconsin, Michigan, Illinois, Indiana, Ohio, Minnesota, Iowa, Missouri. South.--West Virginia, Virginia, Kentucky, Tennessee, North Carolina, South Carolina, Arkansas, Louisiana, Missouri, Alabama, Georgia, Florida. P1ains.--North Dakota, South Dakota, Nebraska, Kansas, Oklahoma, Texas, Montana, Wyoming, Colorado, New Mexico, Idaho, Utah, Arizona. West.--Washington, Oregon, California, Nevada. 101 APPENDIX C APPENDIX TABLES 102 mmH.HH-o* ”eopoHeo «m wemH.N- NHoo.o- aomo.oH 4mmm.vm- am>o use coonma mz HHoA.o- amHo.o- mmmo.o~ Hme.mH- mam.NH-oov NH mz NeaN.H- HNmo.o- Homo.mH w~o~.v~- mmm.~a-oow.Ha mz 4H44.o- QOHo.o- «moa.oH HONN.N- mmN.HH-oom.HH ”mwchpmo Show ano unoeonQEo -HHom scam ”eopoHeo . - . - . . udoaonmEo-mHom Ehmmcoc mz mama o NHmo o 45mm mm mama mm- use peoasoHQEO-HHom some . - . - . . - u:oE%OHQEo owa: Symmco: mz Hose o mHNO o NOOH mm wmeo HN we“ pemasoHoEo-HHom scam A a m mz ammo.H mwoo.o Howw.om 4mmm.mm -wmmeemw ”New MW¢Mumwaa mz camo.H owmo.o oavw.mm OSHA.Hm NHao Mao: owa: some "mapmum unoEAOHQEO Ehmm nova: one QN ”wouuHEo mz NwON.H- mmmo.o- o~om.mH woma.wH- cm>o use m4 m2 aNHH.o QOHo.o aon.oH Hoam.o aa-mm mz m~4m.o NaHo.o Hmma.aH mwmo.w am-m~ nom< oymozco: ”wouuHEo «4m NNm4.o mwsH.o HmNo.NH amma.mHH osmoz "mowm mz oomH.o mmoo.o mmoo.o mooo.o mmemau «soueH mz NomN.H -- mmHm.mm Namm.oe unmumeoo NoucmoH me muanoz mucOHoHHmooo mucoHuHmmoou mOHanpm> HomecomoucH -chmHm muom mo mpOHHO .wum :OHmmOHwom H moHumHhouownmno OHEocooo one OHamwhmoEOw mo :OHpmsco conmoumoh mo muHsmom .oo-nmmH .umwonphoz on“ .UOHHHNHE oocmpmHo AHHB wouwHoommm .H mHnme 103 .Ho>OH mo. may we OHoN Eopm pcoaomeo zHucmonHcmHm no: mm: unoHOHmmooo may memos wz Ho>OH Ho. u *«m Ho>oH mo. u «mN Ho.o n oucmonchHm vaN.m n m mmmN.OMN u .m.m.m vomo.o u Nm ONm.H u mcoHum>pomnoH man3pommscmz HOOHHHEO . . . . pruHHHE mz mwwN H nmNo o onwm mm HmNo Hm one .wOHmHmmmHocs .czocxcs «m mmmH.N mHmo.o «wuu.eN Nwom.Nm Homecpo>ow ««w ovou.N muoo.o mowH.uH «mov.ov moHHpmsocH OOH>Hom mz HHom.H emmo.o Hnoo.vH omoo.mH coma» HHeHo.H was onmoHonz mz ONNm.o NNHo.o vmvw.oN vao.VH moHpHHHu: «am mmow.m emoo.o moom.NH vam.wo :OHHOOHHmcou mz Nmmm.o «NHo.o Nwmm.Hm noww.0H moHHumsocH xnmEHhm "xhpmswcH Summcoz moHHE om-o ”oOHuHEo m2 NHNN.H- wwNo.o- vewH.uN OHNm.vm- Ho>o one mOHHE HmH mz Nmmm.H 0Hmo.o momw.Hm Homo.vv mOHHE omH-HOH mz omom.H memo.o ova.VH ueNN.NN mOHHE OOH-Hm ” uaooaomowcH -mHome muom mo myopno .oum :onmopmom H .ooschcou--.H oHan 104 r mmH.HH-OH ”eonuHeo *«m owou.v- NOHH.o- moNN.NH mem.wm- pm>o mom ooo.mm gem NBHO.4- MHmo.o- Hsow.eH oquw.AN- mam.NH-ooa.NH gem Homo.v- 0Hwo.o- Hmmm.VH vaw.mm- mmm.Nm-oow.Hm m2 aomn.o- OSHo.o- HHoo.mH NNwN.OH- mmN.HH-oom.Hm umwchgmo Egan zHco ucoeonmEo -mHom Show ”woppHEo E» m - m Eye 0 m2 wou4.o- wHHo.o- HNwO.ON wamq.m- ommm mmommowwmm-cHoqummm . . . . ucoeonQEo 0mm: Epmmco: m2 wmmm o mHHo 0 wave NH Hoom 0 wow pcosonmEo-wHom Show . . . . udoEkonEm mmmz gem Noam e mQNH o ommm 0H mHoc Nw Eemmco: mom Heoz owe: Epmm «am oHVN.m HNmo.o NVHN.mH nmoa.mo NHHco Haoz mum: Show ”msumum pcoeonmEO Egon Hows: one «N "wouuHEo m2 Nmom.H- mHmo.o- moov.NH HOOH.©H- Ho>o com me mz wNHw.H wmmo.o owvo.mH memo.mN ev-mm gem mNoH.m onoo.o HmNn.NH vmvm.mm qm-mN Hom< opwozco: ”wouuHEo 44m oaHN.m owmo.o mmwm.m~ ooao.om ocmmz "comm m2 HHmN.H- mNmo.o- mNoo.o ovoo.o- omcmco oEoo:H ««m NnHm.< -- mmmo.oN mNmu.om pneumcou Nmocon m9 mucmHoz mu:OHonmooo mucoHonwooo moHanhm> ucoocomowcH -chme «pom mo mHOHHO .wum :onmOHmom H muHumeOuomHmao OHEOGOOO one OHcmmHmoEoo mo :oHumzco cOHmmOHNOH mo muHsmom .oo-nmmH .Hmhuaoo cphoz Ox» .woumhmHE oucwumHo :HHB OOHmHOOmmm .N oHnae .Ho>OH mo. on» um OHON Eogw Homaomme AHpcmOHchme Ho: mwz ucoHuHmmoou map momma mz 105 H0>OH Ho. u arm Ho>OH mo. u «mN Ho. u oocmonHcmHm oNom.NH u m oooH.NNN u .m.m.m omwo.o n Nm Nnm.N u mcoHum>pomnoH mcHasuommscmz ”woupHEo . . . . xpmuHHHE mz Nnmw H Hmmo o ommH mN mmHo we one .oonHmmmHoas .czocxc: mz ommm.o meo.o mmoo.wH Nonm.NH Homeopo>ou mz Nmmw.H mumo.o NNNN.¢H mva.oN mOHpumswcH OOH>Hom mz mmoo.H- OHNo.o- mmmw.HH mmuv.HH- memo“ HES?H one onmoHozz mz HwNo.o- mooo.o- movo.mH ovmm.o- moHuHHHuD «m oNON.N mmvo.o mmNH.VH ONON.Hm :oHposhumcou «am Nmoo.N Nnvo.o mmnu.mN mnom.uu mOHpumnon kamEHHm "xhpmswcH Epmmcoz mOHHE om-o "wouuHEo «aw mmmo.m vmwo.o momm.ve mon.NNN Ho>o one moHHE HmH «am memn.m- uwoo.o- mnvo.VH van.Nm- mOHHE omH-HOH «em mwov.m- wmoo.o- mNNm.m oeHv.vm- moHHE OOH-Hm u HomecomoocH -chme «pom mo muchho .vum GOHmmopwom H .eoscheoo--.N oHan 106 mmH.H»-oa ”empoHso mz momo.o- mHoo.o- Homm.0H vam.o- no>o wow ooo.mm mz mmom.o mmoo.o Hmam.mH wNoo.OH mmm.Na-oo4.Na mz mnow.o- mNHo.o- oHVN.oH memo.vH- mam.Nm-oom.Ha mz waN.o- wmoo.o- Nonm.mH wamm.m- mmm.Hm-coN.Hw "mmcHahmo Show zHco pooEHOHmEo -HHom Egon ”OOHHHEO . - . - . . ucoeonQEo-mHOm Epmmcoc mz wmww H mNmo o «mmH mN comm he use pooeonQEo-mHom Spam x a w mz Ammo.H- mmmo.o- m~mm.- 4~om.am 6mm” MWmmmommww-mwwwemmam . . . . pcoeonmEo owa: «m oOHN N unmo o NNHo HN VHov ow Ecmmco: wow H903 own: show mz NQNN.H memo.o wamo.m~ 5455.0m sHeo ego: owa: some "mspmum pcoeonmEo spam Home: use VN "oouuHeo mz movw.H- oowo.o- Nmmm.mH ummu.mN- Ho>o one me «m Novo.N- onqo.o- omNm.¢H enem.om- v¢-mm wz womH.H HoNo.o wvHu.VH omum.NH em-mN Hom< opmozco: "voguHEo «m onm.N wovo.o omom.HH mNmm.NN ogmoz "comm wz wwww.H oomo.o mNoo.o emoo.o omcmno oEounH «em Hmmo.N -- mmom.vN nmnm.uwH Apnmumcou NooqmuH me mucmHoz mucoHuHmmooo mucoHOHmmooo moHanhm> pcowcomooaH -wHGMHm «pom mo mnouno .wum :onmoumom H moHumHuopomHmnu OHEocooo one OHnmmhmoEow mo :oHpmsco :onmoumop mo mszmom .oo-nmmH .gusom on“ .woumhmHE mocmpme :HHz woumHOOmmm .m OHQOH 107 .Ho>oH mo. can an OHON Eonm unopommHO zHucmonchHm go: we: HOOHOHHHOOO on“ momma mz Ho>OH HO. u ««m Ho>OH mO. u «mN HO. u oocmonHcmHm OOOH.OH u m ONm¢.HmN u .m.m.m OBOH.O u Nm nmu.N n mOOHum>homnoH mathpommsnmz "OOHHHEO somOHHHE m2 Homm O- mNOO O- NHHH NN munm OH- wow .OonHmmmHOO: .czocxcn mz OHmm.H NNN0.0 OOO0.0N mHme.uN ucoscho>ow m2 OOON.O- NOO0.0- vam.OH NOON.¢- mOHpumzvcH ooH>Hom mz NNOO.H- mvN0.0- Nuov.mH Omum.eH- oumgu HHOHO.H Ono OHmmoHonz mz mmvm.H- NON0.0- ONmO.eN OOOe.wm- moHuHHHHD mz mumm.O- NBO0.0- OmNO.mH Hmmm.m- coHpusuumcou mz NNm0.0 OHO0.0 OOH0.0N HHOm.H mOHupmsvcH thEHpm "ShamswcH Ehmwnoz moHHE Om-O "OOHHHEO «m ONOO.H Omm0.0 mmNH.NmN eva.OOv no>o Ono mOHHE HmH «am mew.v- “HO0.0- OMO0.0N mOmN.HOH- moHHE OmH-HOH «em «BBO.HH- mNNN.O- ONNm.OH NONH.HNH- mOHHE OOH-Hm " pcownomowcH -mHame «pom mo whopno .Oum :onmOHMOm .emsereoo--.m oHnae 108 mmH.HH-o» "eopuHso mz mOOm.H- mmmo.o- NHmN.NH mmmm.uN- Ho>o Odo OOO.mO mz NNHm.H- ommo.o- eHmH.om OHoo.om- mam.~a-ooa.~m «m OOMH.N- ONv0.0- mva.OH mONm.Hv- OOm.NO-OOO.HO mz HvNo.o- mooo.o- mmov.oH 54mm.o- mmN.HH-OON.HH "mmchpmo Spam cho ucoeonmEo -mHom Show ”OOHHHEO . . . . “doEHOHQEO-wHom Epmmao: mz man H- NeeO O- «New Nm NOHN mm- Ocmxpcmsonmao-mHom Shem . - . - . . - Home OH 60 0 m3 summoo: mz OmOO O ONOO O NOON ON NOOO N was ucoeonMEommHomepmm «am oomN.¢ mmNH.o HmOH.mN HmHH.OOH Seaweed ewm8mtmw ”New “wan «am «OOH.m OOOH.O OHO0.0N mNmm.wO NAHco Hpoz owa: spam ”msumum uqoaonQEo Show Howe: One «N ”OOHHHEO mz HONN.H- mmmo.o- HOmH.mH mOON.ON- po>o won me mz Nme.O- ONO0.0- NNO0.0H Omom.N- ev-mm mz mOmv.H NON0.0 mOOO.mH «NOO.NN vm-mN Hom< Ohmozco: ”OOHHHEO ««m “soo.- waeo.o- Nmm4.mN mHON.oo- oummz “comm mz NmOH.H OHN0.0 mmO0.0 OmO0.0 owzwgo oaoucH «em mmoe.w - NOm0.0m NHvO.NON ucmpmcou NoocmoH me mungoz mu:OHOHmwooo mucoHonmooo mOHanpm> uzovcomowaH -chme muom mo muonuo .vum :onmopwom H .Oo-anH .mchHm ogp .OOpmhmHE oucmpme Hsz wouwHoommm moHumHHouomhmao OHEocooo use OHgmmnmoaow mo :OHumsco :onmoamoa mo muHsmom .O OHan 109 .Ho>OH mo. esp um OHON anw poopommHO prcmonHcmHm poo mm: pcoHonwooo ecu mauve mz Ho>OH Ho. u ««m Ho>OH mO. u «m N HO. u oocmonHamHm OmOO.NH u m mumH.mON u .m.m.m OONH.O u Na ONN.N u mcoHpm>HOmnoH wcHHsuommscmz ”OOHHHEO . . . . sgmpHHHe mz NONm O OOHO O Nwom NN Hmmm OH wow .OonHmmeocs .czocxca mz OONN.O- NOH0.0- OOmO.mN OON0.0H- u:oE:po>oo «m HNOm.N- Han.O- mnwm.OH NOOm.wO- moHpumswcH ooH>pom «gm HOHN.m- OOO0.0- ONN0.0H OOO0.0m- OOOHH HHOHO.H use onmoHonz «m Nomm.N- NOO0.0- NHN0.0N NONN.HO- moHuHHHp: mz mNO0.0 OHH0.0 mwmm.wH ONO0.0 :oHuosppmoou mz OOmN.H- me0.0- OOH0.0m NNOn.mO- moHHHmOOcH meEHHm ”xhumswcH Enmwcoz mOHHE om-O "OOHHHEO «m NON0.0H- mOHN.O- mNHN.mH HNNN.OOH- 90>0 One mOHHE HmH «m mOOO.n- OOOH.O- HmON.NH omvN.NmH- mOHHE OmH-HOH «m OmON.O- mHOH.O- NON0.0H uonm.omH- moHHE OOH-Hm u ucowcomowcH -chmHm muom mo mHOHno .Oum :onmopmom H .emzaHocoo--.a oHan 110 «am Ouwm.m- mNmH.O- mmuO.wH NOO0.00H- Ho>o Ocm OOOHmw ««m smac.~- Namo.o- ooam.m~ Hoaa.ao- mom.ma-ooa Na m2 OmO0.0- OHN0.0- OOON.HN NOOO.HN- OOm.NO-OOO.HO mz NOHH.H- mON0.0- OOmu.OH OOmO.NN- OON.HO-OON.HO ”mmchuwo Ewen cho pcoeonmEo -wHom Ehmm "OOHHHEO . . . . pcoEAOHmEo-mHom Eammco: mz NONO O ONOO O Omen OO HOOO v Oem HooeonmEo-mHom seem % a m 0 m2 H4NO.O meo.o smN4.oo «moa.am “mm” mWommowamw-%%MW:smmm Eko QEO owe: tam mmao.m masw.o OHOO.OH owmm.NAH Eamoco: Onwaxgow mam: scam aam OOHO.m OOOH.O Omom.mw NOM0.00H cho Hhoz owa: Eamm “msumum “amaonQEo Show Hows: One «N ”OOHHHEO mz HHN0.0 NOH0.0 NOmm.NH OomO.HH Ho>o Ono mv mz OOOm.H OON0.0 OOO¢.OH umNN.ON OO-mm «gm NNNO.m OONH.O OmvH.NH NmOO.Nw Om-mN Hom< oumozco: ”OOHHHEO mz mNnm.O mNO0.0 mOO0.00 HOO0.0H oumoz ”comm mz oqov.o- omoo.o- Haoo.o mHoo.o- mmemgo osoueH mz BNom.H - nmmv.Ov OOOM.NO ucmumooo Noocon me mugmHoz mpcOHOHmmooo mucoHonmooo mOHanhm> HOOOOOQOOQH -chmHm «pom mo gouge .Opm :onmopwom H mOHpmeouomumso OHEocooo Ono oHnmmHmoEOO mo coHumsco :OHmmOHmou mo mHHzmom .Oo-NmOH .umoz ma“ .OoumHmHE oocmpmHO HHHS OOHOHOOmmm .m oHnae 111 .Ho>OH mo. may we open Eouw pconomeO xHucmonHOMHm Ho: mm: HOOHOHHHOOO may momma mz Ho>OH HO. u ««m Ho>OH mO. u «mN HO. u moomonHcmHm NHN0.0 u m ONO0.00N n .m.m.m Nnmo.o u Nm «mm.N u mOOHum>homnoH mcHgsuummzcmz "OOHHHEO ram mmmn.~ memo.o wwom.Na OHwo.oHH mam .eoHHHmmmHuamHmwwameO mz OOON.O OmH0.0 OONH.Om mmmO.NN unmeaho>oo mz «NON.H NNN0.0 «NON.ON HOO0.0N mOHHHmOOcH ouH>Hom mz ONON.O- wooo.o- OmO0.0H OOOO.¢- owmhp HHOHO.H One onmoHonz mz mON0.0- OOH0.0- OOON.NN HONO.mN- mOHuHHHH: «gm OON0.0 NOmH.O OOO0.0N NOm0.0NH OOHuoshumaou mz mum0.0 NHO0.0 OOmH.Om «ONO.H mOHHumOOcH meEHHO "xhumswaH symmnoz moHHE om-O ”OOHHHEO «m OOOm.N NHm0.0 mON0.0N HNmN.HO Ho>o Ono moHHE HmH «gm NHvu.m- NON0.0- ONN0.0N OOmH.nn- moHHE OmH-HOH ««m nmNO.m- OHO0.0- ONOO.NH mHOn.Nm- mOHHE OOH-Hm " unowcomowcH -chme muom mo gouge .Oum :OHmmOHmom H .eoscHoeoo-.m oHOmH