NEEDS AND POTENTIALSFOR '- ESCAPE FROM POVERTY THROUGH] ” ~ RETRAINING‘FOR FAMILIES m -. -. VTHE EASTERN CORN BELT . i Thesis for the. DegreéjofPh. D. " ' MICHIGAN‘STATEUNIVERSITY I ' *MARVINE. KONYH’A} ‘ ' 1970' " We? , LIB}? A 1’ Y . Michigan 8 :ate University l-HESIS This is to certifg that the thesis entitled Needs and Potentials for Escape from Poverty Through Retraining for Families in the Eastern Corn Belt presented bg Marvin E. Konyha has been accepted towards fulfillment of the requirements for PhD degree in Agricultural Economics new x A421 6:, ZZZ. I “”7 Major professor I Date 2/27/70 3 0-169 ABSTRACT NEEDS AND POTENTIAL FOR ESCAPE FROM POVERTY THROUGH RETRAINING FOR FAMILIES IN THE EASTERN CORN BELT BY Marvin E. Konyha The objectives of this study were to determine the extent of poverty level incomes in the East North Central States and to estimate the potential ability of job re- training to provide employment and income sufficient to raise low income families above the poverty line. The study provided an extensive review of the literature pertaining to the nature and extent of rural poverty in the United States. It employed a slight modification of the Orshansky poverty criteria to measure the extent of poverty in the open country area of the East North Central sub— region. Data was provided by the 1967 Rural Life Survey study conducted by the Economic Research Service of the U.S. Department of Agriculture. It was found that sixteen percent of the consumer units and thirteen percent of the individuals over age fifteen received 1966 poverty level incomes, that is, incomes below the 1.00 ratio of income received to income required. For retraining purposes, all individuals below the 1.50 income ratio were included in the low income category. Twenty—six percent of the survey individuals were so classified. Based on their age-health status, sixty-three percent of these Marvin E. Konyha low income persons were considered to be potentially economically viable. The theoretical base for eliminating rural poverty through job retraining was established by application of the investment in human capital extension of marginal productivity theory. Procedures for estimating expected increases in future income streams were evaluated, with the benefit-cost analysis approach being recognized as nwst apprOpriate for estimating the gains from retraining. Tentative empirical estimates indicated that seventy-five percent of the low-income families could theoretically be raised above poverty status through retraining. Minimum skill level reqUirements for earning above poverty incomes were developed based on the specific vocational preparation scale (SVP) applied to occupational titles by the U.S. Department of Labor. Job titles utilized were those for which workers have been retrained under MDTA programs. Respondents' present skill levels, in terms of the SVP scale, were determined from their prior vocational training experience. Compared with required skill levels, ninety-two percent of the males and eighty-six percent of the females with potential viability were found in need of job retraining- Based on years of schooling completed as a measure of general educational develOpment levels, fifty-eight percent 0f the males and two-thirds of the females with retraining needs were considered to have good potential for completing it w 1|. 1| El Marvin E. Konyha job retraining programs. Thirty-seven percent of the males and thirty percent of the females were found to have fair retraining potential, and less than five percent of the respondents were found to have no retraining potential apart from programs providing basic education in addition to vocational training. Theoretically, respondents should evaluate the opportunity for job retraining in terms of expected future income streams. Survey data indicated that low income respondents reacted in a manner consistent with such an economic decision rule in relation to age, educational level, recent unemployment, attitudes toward present job, availability for alternative employment, and potential costs of retraining programs. J Application of all estimates developed in the study to the survey's potentially viable low income respondents indicated that from twenty to twenty—five percent of them could be removed from the poverty category via job re— training. 'The study results implied that existing vocational training programs have been most inadequate for preparing this subregion‘s rural workers for employment in today's labor market. The study found sufficient retraining needs, POtentials for completing, and interest in retraining to warrant major efforts in this area. Implications concerning the type of occupational training needed and the type of delivery systems required were briefly explored. Serious Marvin E. Konyha implications were raised for the large percentage of rural poor who could not be removed from poverty through job retraining. NEEDS AND POTENTIALS FOR ESCAPE FROM POVERTY THROUGH RETRAINING FOR FAMILIES IN THE EASTERN CORN BELT BY Marvin EF‘Konyha A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1970 W02 appreci to his sacriii studies ‘ H) T1 providt of Agr Financ joint] the U1 Resea: Wood, com Steve Guide Chai; An-fim 542.307 744° ACKNOWLEDGMENT S Words alone do not adequately express the grateful appreciation the writer extends to his wife, Janet, and to his daughters, Cherie Lea and Wendy Jo, for the many sacrifices they made during the course of his graduate studies and the preparation of this thesis. The writer acknowledges the financial assistance provided for one year of graduate study by the Department of Agricultural Economics, Dr. Lawrence Boger, Chairman. Financial assistance for this thesis study was provided jointly by the Department of Agricultural Economics and the United States Department of Agriculture, Economic Research Service, Economic Development Division. Special appreciation is extended to Dr. Garland P. WOOd, respected friend and chairman of the writer's guidance committee. Other committee members were Dr. Robert D. Stevens, Dr. Hideya Kumata, and Dr. Harvey Choldin. Guidance for this study was provided by Dr. Dale E. Hathaway, chairman, Dr. James T. Bonnen, Dr. Milton Steinmueller, and Dr. J.A. Beegle. Additional assistance was provided by Miss Jeanette Fitzwilliams, Rural Life Survey project leader for the Economic Development Division. The writer extends his appreciation to the faculty and graduate students of the Department of Agricultural Economics for providing him with a stimulating and challenging atmosphere in which to studY- ii III IV TABLE OF CONTENTS Page I. INTRODUCTION. . . . . . . . . . . . . . . . . . . l A. Need for the Study. . . . . . . . . . . . . . l B. Purposes and Objectives . . . . . . . . . . . 4 C. Methodology . . . . . . . . . . . . . . . . . 5 1. Sampling. . . . . . . . . . . . . . . . . 5 2. Survey. . . . . . . . . . . . . . . . . . 8 3. Editing . . . . . . . . . . . . . . . . . 9 4. Analysis of Data. . . . . . . . . . . . . 11 D. Hypotheses to be Tested . . . . . . . . . . . 12 II. RURAL POVERTY AND RETRAINING - LITERATURE REVIEW. 17 A. Rural Poverty . . . . . . . . . . . . . . . . 17 B. Rural Poverty in Regions of Commercial Agriculture . . . . . . . . . . . . . . . . . 20 C. Rural Needs for Job Retraining. . . . . . . . 22 D. Rural Areas' Share in Job Retraining Programs. 0 o o o 0 c o o o o o o o o o o o o 25 III. LABOR THEORY APPLIED TO RURAL POVERTY AND RETRAINING c o o o o o 0 O 0 O 0 0 0 0 0 0 0 0 O 0 2 9 A. Introduction. . . . . . . . . . . . . . . . . 29 B. Marginal Productivity and Bargained Wage Theories. . . . . . . . . . . . . . . . . . . 29 C. Human Capital Investment. . . . . . . . . . . 40 Do CODClUSiOfl. o o o O O o O O 0 O 0 0 0 O O O o 46 IV. ECONOMIC AND SOCIAL CHARACTERISTICS OF THE SURVEY POPULATION . . . . . . . . . . . . . . . . 48 A. Family Composition. . . . . . . . . . . . . . 48 B. Farm and Non-farm Households. . . . . . . . . 50 C. Delineation of Poverty Categories . . . . . . 51 D. Components of the Poverty Status Formula. . . 52 iii 7‘ IV POI RE'I A. B C IIIIIIIIIIIIIIIIIL Page E. The Rural Life Survey Population. . . . . . 56 1. Extent of Poverty - Standard Criteria 56 2. Extent of Poverty for Retraining Purposes. . . . . . . . . . . . . . . 63 3. Selected Characteristics of the Retraining Study Poor . . . . . . . . . 65 F. Summary . . . . . . . . . . . . . . . . . 74 POTENTIAL FOR INCREASED EARNINGS THROUGH RETRAINING. . . . . . . . . . . . . . . . 77 A. Introduction. . . . . . . . . . . . . . . . 77 B. Expected Future Income Streams. . . . . 77 C. Retraining Benefit-Cost Analyses. . . . . . 86 D. Empirical Estimates of Potentials for Increased Earnings Through Retraining . . . 103 1. Full Time Employment Increases. . . . . 103 2. Increases in Consumer Unit Income . . . 105 3. Increased Hourly Earnings . . . . . . . 107 RETRAINING NEEDS AND POTENTIALS OF LOW INCOME PERSONS . . . . . . . . . . . . . . . . . . . . 115 A. Introduction. . . . . . . . . . . . . . . . 115 B. Development of Basic Training Requirements. 116 C. Respondent SVP and GED Levels . . . . . . . 119 D. Retraining Needs Based on Basic Training Requirements. . . . . . . . . . . . . . . 120 E. Retraining Potentials — Introduction. . . . 124 F. Characteristics of Successful Retrainees. . 126 1. Defining Successful Retraining. . . . . 126 2. Hypotheses. . . . . . . . . . . . . . . 127 3. Characteristics Which Influence 12 8 Retraining Success. . . . . . . . . . . iv PM I. ‘[, c. The and H. Pot VII. RETRAII IN RETI A. In‘ B. In' C. Re VIII . SUMMAI A. II “ow- -5—»»".sm at". w —. q...- S Page G. The Relationship Between Education and Training . . . . . . . . . . . . . . . 131 H. Potentials for Retraining Success. . . . . 135 VII. RETRAINING POTENTIALS BASED ON INTEREST IN RETRAINING. . . . . . . . . . . . . . . . . 141 A. Introduction . . . . . . . . . . . . . . . 141 B. Interest in Retraining . . . . . . . . . . 142 l. The Decision Rule. . . . . . . . . . . 142 2. Hypotheses . . . . . . . . . . . . . . 146 C. Retraining Potentials. . . . . . . . . . . 167 VIII.SUMMARY AND CONCLUSIONS. . . . . . . . . . . . 168 A. Introduction . . . . . . . . . . . . . . . 168 B. Rural Poverty in the Eastern Corn Belt . . 169 C. Job Retraining - The Escape Route From Poverty . . . . . . . . . . . . . . . 172 1. Labor Theory . . . . . . . . . . . . . 172 2. Human Capital Investment . . . . . . . 172 3. Benefits of Job Retraining . . . . . . 173 D. Characteristics and Attitudes Related to Retraining. . . . . . . . . . . . . . . 176 1. Retraining Needs . . . . . . . . . . . 176 2. Retraining Potentials Based on . Individual Characteristics . . . . . . 179 3. Retraining Potentials Based on Attitudes Toward Retraining. . . . . . 180 E. Conclusions of the Study (Recapitulation). 184 1. Conclusions Regarding the Extent 184 Of POverty. o o o o o o o o o 0 0 0 2. Conclusions Regarding Potential 185 Income Gains From Retraining . . . 3. Conclusions Regarding Potentials 185 for Retraining Success.. . . . . . . . V I D. l BIBLIOGRA‘ IPPENDIX A. B. APPENDIX APPENDIX Page , F. Potentials For Escape From Poverty Through Retraining - Empirical Estimates . 186 G. Conclusion . . . . . . . . . . . . . . . . 189 IX. IMPLICATIONS . . . . . . . . . . . . . . . . . 190 A. Introduction . . . . . . . . . . . . . . . 190 B. Implications for Existing Rural Vocational Training Programs . . . . . . . 191 C. Implications for Redirection of Rural Training Activities. . . . . . . . . . . . 193 1. Potentials for Escape From Poverty Through Retraining . . . . . . . . . . 193 2. Retraining for What. . . . . . . . . . 195 3. Implications of Retraining Needs . . . 197 4.- Implications of Retraining Potentials 198 5. Implications of Expressed Interest in Retraining. . . . . . . . . . . . . 198 6. Implications for Trainee Recruitment . 200 D. Conclusion . . . . . . . . . . . . . . . . 200 BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . 202 APPENDIX A . . . . . . . . . . . . . . . . . . . . 210 A. Occupational Titles Used . . . . . . . . . 210 B. SVP Median Levels. . . . . . . . . . . . . 213 APPENDIX B . . . . . . . . . . . . . . . . . . . . 215 APPENDIX C. . . . . . . . . . . . . . . . . . . . . 219 A. General Educational Development Measures . 219 B. Specific Vocational Preparation Measures . 222 Vi 10. 11. 12. 13. LIST OF TABLES Page Number and Percent of Poor Rural Families in the East North Central States - 1959 . . . . . . 1 Number of Counties in Which Median Cash Income of A11 Rural Families, Only Rural Farm Families, and Only Rural Nonfarm Families was less than . 21 $3,000 - 1959 o o o o o o o o o o o o o o o 0 Percentage of Consumer Units in Poverty, by Residence and Family Status . . . . . . . . . . 58 Percentage of Consumer Units in Poverty, by Age- Health and Family Status, and Percent . . . . . . 59 Distribution. . . . . . . . . Number and Percent Distribution of Survey Population by Consumer Unit Status. . . 60 Percentage of Survey Individuals in Poverty by Age-Health and Residence Status, and Per- cent Distributions. . . . . . . . 62 Percentage of Survey Population by "Degree of Poverty" and Age-Health Status, and Percent Distributions . . . . . . . . . . . . . 64 Percentage Distribution of the Survey Poor by Relationship to Educational Medians of the Re- gional Population, by Degree of Poverty and by sex I C O O O 0 O O O O O O O O O I O O O O O 6 6 Relationship Between Degree of Poverty and Post-High School Technical Training for sur— . . . 68 vey Poor; Percent Distributions . . . . . Relationship Between Degree of Poverty and Occupational Skill Index for Survey Poor Who Held Jobs in 1966 and Totals by Sex; Percent Distributions . . . . . . . . . . 70 Types of Problems Encountered in Finding Employment By the Poor Who Sought Jobs in Past Three Years, by Type of Job Sought; Percent Distributions . . . . . . . . . . . . . . . . 72 Measures of the Availability of the Poor for Alternative Employment in Succeeding Year; Percent Distributions . . . . . . . . . . . . 73 Median Years of Schooling Completed in United States, Selected Age Groups . . . . . . . . . . 81 vii . Expe . Pret MDT] . Posi tra: Tra. . Pre MDT and . Wee Hav Per . Inc and - Po: Ea] Tr: im a flL—l- In 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. Expected Lifetime Earnings Compared - 1959 . . Pretraining and Posttraining Earnings of MDTA Institutional Graduates in 1965 and 1966. Posttraining Earnings Compared with Pre- training Earnings of MDTA Institutional Training Graduates in 1965 and 1966. . . . Pretraining and Posttraining Earnings of MDTA Institutional Training Graduates in 1965 and 1966. . . . . . . . . . . . . . . Weeks Employed Full-time for Respondents Having Some Full-time Work in the Reporting Period, by Respondent Class. . . . . . . . . . Income Status of RLS Consumer Units Before and After Addition of $500 "Retraining" Income Posttraining Status and Mean Posttraining Earnings of 1965-66 MDTA Institutional Training Graduates, by Pretraining Earn- ings Level . . . . . . . . . . . . . . . . . . Relationship Between Consumer Unit Highest Wage Earned and Income Status. . . . . . . . . Relationship Between Pretraining and Post- training Hourly Earnings and Income Status for Rural Life Survey Consumer Units Below 1.00 Income Ratio. . . . . . . . . . . . . . . Relationship Between Pretraining and Post- training Hourly Earnings and Income Status for Rural Life Survey Consumer Units Below 1.50 Income Ratio. . . . . . . . . . . . . . . Relationship Between Potentially Viable Respondents Skill Levels and Median Re- quired Skill Levels, by Income Ratio and by Sex; Percentage Distributions . . . . . Number and Percent of MDTA Enrollees 45 Years Old and Older During Fiscal Year 1968. Relationship Between Formal Occupational Training and Formal Education for Workers With Less Than Three Years of College, 1963. Selected Characteristics of Persons Enrolled in MDTA Institutional Training Projects, Cummulative and for Fiscal Years, 1963—68 . . viii Page 85 90 91 92 104 108 110 110 112 113 122 125 132 134 39. Ret ent Ret Res Int PM '17? rnmbdz'u 28. 29. 30. 31. 32. 33. 34. 35. 360‘ 37. 38. 39. 40. 41. Page Retraining Potentials of Low Income Respond— ents With Retraining Needs, by Sex . . . . . . 137 Retraining Potentials of Higher Income Respondents With Retraining Needs, by Sex. . . 138 Interest in Retraining in Relation to Poverty Status, by Sex; Percent Distribution . 147 Interest in Retraining in Relation to Age, 149 by Sex; Percent Distribution . . . . . . . Interest in Retraining in Relation to Amount of Prior Training, by Sex; Percent Distribution 0 O O O O O O O O O O O O O O O O 15]. Interest in Retraining in Relation to Amount of Prior Training for Respondents With Measurable SVP; Percent Distribution . . . 153 Interest in Retraining in Relation to Employment Status at Time of Interview, by Sex; Percent Distribution. . . . . . . . . 155 Relationship Between Difficulties in Finding Employment and Interest in Retraining, by Sex; Percent Distribution . . . . . . . . . 157 Interest in Retraining in Relation to Stated Availability, by Sex; Percent Distribution 0 O O O O O O O O 0 O O O O O O O 159 Interest in Retraining in Relation to Potential Availability, by Sex; Percent Distribution . . . . . . . . . . . . . . . . 160 Interest in Retraining in Relation to Attitude Toward Current Job, by Sex; Percent Distribution . . . . . . . . . . . . . 163 Percentage Changes in Retraining Interest With Increasing Potential Costs of Retraining, as Percent of Those Interested and of Total Poor Viable Population, by sex. 0 O O -. O O O O O O O O O 0 O O O O I 164 Relationship Between Education and Interest in Retraining, by Sex; Percent Distribution. . 166 Potentials for Viability in Farming Based on Acreage and Its Use, 1966, for Poor Respondents. . . . . . . . . . . . . . . . . . 196 ix 42. Use Unit APPENDIX W-l Num] of ' Pro' W-Z Wei C011 B-l Wei APP B-2 Req Med APP C'1 Gen Cat 0-2 3v: VOC C‘3 svx Tr: Page 42. Use of Public Services by Low Income Consumer Units; Percent With Visits, by Income Ratio 201 APPENDIX TABLES A-l Number of MDTA Projects Approved and Number of Trainees Enrolled in Selected States, by Program Type, Through June, 1965 . . 212 A-2 Weighted Number of MDTA Approved Retraining Courses by SVP Scales, by Sex... . . . . . 214 B—l Weighted GED and SVP Values of MDTA Courses Approved Through June, 1965. . . . . . . . . 216—217 B-2 Required GED Levels and the Corresponding Median SVP Levels of MDTA Training Courses Approved Through June, 1965, by Sex. . . . 218 C-1 General Educational Development Categories . . . . . . . . . . . . 221 C-2 SVP Weeks Equivalent of High School Vocational and College Training. . . . 224 C-3 SVP Ratio of Unused Training 226 C-4 SVP Equivalents of Weeks of Vocational . . 227 Training. . . . . . . . . . . . . . A. Wee Man hMeI count 01 such act 1960 Ce] problem in the : Numb Door f State Ohio Indiane Illinoi Michige WISCQM Total 2011ch um - i, “la PHI) I. INTRODUCTION A. Need for the Study Many households, and individuals in those households, in the Eastern Corn Belt are in poverty. Whereas an exact count of these families is not readily available, and any such count is dependent on how "poverty" is defined, the 1960 Census did give an indication of the extent of this problem. As indicated in Table 1, over 580,000 rural families in the five East North Central States had 1959 incomes below Table 1. Number and Percent of Poor Rural Families in the East North Central States - 1959 (poor families are those with net money incomes under $3,000) No. of poor Total no. of Poor families as State rural familiesl rural families2 percent of total Ohio 138,240 640,222 21.6 Indiana 103,594 450,591 23.0 Illinois 134,958 505,280 26.7 Michigan 108,215 509,964 21.2 Wisconsin 98,425 347,277 28.3 Total 583,432 2,453,334 23.8 Source: 1Alan R. Bird and John L. McCoy, White Americans in Egral Poverty; Agricultural Economic Report No. 124 (Washington: U. S. Department of Agriculture, Economic Research Service, November, 1967). ' 2U. S. Bureau of the Census, U. S. Census of ngulation: 1960. General Social and Economic Characteristics, PC(1) - C (five State volumes), table 50. ‘— the crud sented n in under may be c families the comi and You A 2 percent were ru were 14 familie the Sou this re countrk It heuseht Corn 3, Wake t? aCCept oondit POVert the crude family income poverty line of $3,000. This repre- sented nearly 24 percent of all rural families in the region. An understanding of the magnitude of this number of families may be obtained by noting that the number of rural poor families exceeded the 1960 total number of all families in the combined cities of Cleveland, Milwaukee, Indianapolis, and Youngstown. A 1966 Census Bureau survey found that thirty-six— percent of the whole North Central region's poor families were rural residents. Also, the 1966 rural poverty rates were 14.6% among farm families and 16.0% among.nonfarm families in the region - rates which were exceeded only in the South. (By contrast the urban poverty rate was 8.5% in this region, the lowest urban rate of any region in the countryl). It is thought that certain heads of poverty level households, and other household members, in the Eastern Corn Belt possess characteristics and attitudes that would make them prime candidates to undertake retraining and accept different jobs. The successful attainment of these conditions would be expected to raiSe the family above the poverty level and to improve the chances for children of 1Harold L. Sheppard, "A Search for New Directions in the War Against Poverty," Appendix Paper in Toward Economic §§curity for the Poor, Subcommittee on Employment) Manpower, and Poverty of the Committee on Labor and Public Welfare, U. S. Senate (90th Congress, 2nd Session), October, 1968, pp. 85-860 these how For for the work one "trainin he consi in the 1 Wha that mah in this it? Hm What, i; employm, to be 5 within to thes Th above p f°r bet all per being i Provide by its; t° a st It Can In rural j these households to stay out of poverty. For some it would not be-"retraining," but training for the first time in any formal sense. However, if whatever work one may have done previously is considered to be "training," then any attempt to raise one's skill level can be considered retraining. The terms are used interchangeably in-the literature, and will be so treated here. What are the characteristics of rural poor individuals 1 that make them retrainable? How many rural poor individuals in this region actually need retraining and would profit by it? How many would be willing to undergo retraining and what, if anything, would they be willing to pay for new employment Skills? If national anti-poverty programs are to be effective in alleviating the rural poverty which exists within this region of commercial agriculture, then answers to these questions are urgently needed. There are certain value judgments implicit in the above paragraphs. This focus on retraining the rural poor for better paying employment leads to the conclusion that all persons should be job-holders, that all are capable of being job-holders, and that having a job will automatically provide a family with above poverty income. Such a focus, by itself, would imply that all rural poverty was amenable to a strictly "economic" solution. It is recognized in this study that all rural poverty can not be eliminated through job retraining. Some of the rural poor are unable to work for physical reasons, either advanced social u or socie aspirati Solutiox require Such pe: Who eiti trainin eWilpyu retrain poverty the non 3- Pt TI POtentj income {2) t. f°r me] Corn B< they f, mange Io l'ea T 1 advanced age or disability. These.are singled out for special social welfare programs. Others suffer from various emotional or social disabilities, such as a lack of motivations and aspirations or having been engulfed in a "culture of poverty." Solutions to their poverty problems would also be likely to require programs other than, or in addition to, job retraining. Such persons are likely to be among those low income respondents who either express no interest in participating in job re- training programs or, when retrained, would fail to obtain employment. Attention will be drawn to the failure of job retraining to provide an escape-from poverty for all rural poverty victims, but the exploration of programs to deal with the non-retrainable poor is beyond the scope of this study. B. Purposes and Objectives The purposes of this study are: (l) to determine the potentials for job retraining to provide employment and income which would raise families above the poverty line; (2) to determine the needs and potentials for retraining for members of rural, poverty-level households in the Eastern Corn Belt; (3) to examine existing retraining programs as they function in rural areas; and (4) t0 propdse proqram changes to enable them to more effectively aid individuals to realize their potentials. The objectives of this study are: 1. To explore contemporary economic theory as it applies to rural poverty and retraining. 2. To determine specific socio—economic characteristics Math This the pew u. S. De; Opportuh: such as age, education level, and prior work experience of subgroups of the rural population in the Eastern Corn Belt with poverty-level incomes, and thus to determine the needs for escape from poverty. To estimate_the potentials for escape from poverty of those households and individuals in each subgroup through retraining, based on measures of their retrainability. To estimate the potentials for escape from poverty of those households and individuals in each sub- group determined to be retrainable, based on their stated interest in retraining. To estimate approximate numbers of individuals who would participate in retraining programs, based on combinations of estimated needs and potentials for retraining. To prepare recommendations for program changes to improve chances of households and individuals to escape from poverty through retraining which will qualify them for higher paying employment. Methodology 1. Sampling This study is part of a larger study being conducted by the Economic Development Division, Economic Research Service, U. S. Department of Agriculture, for the Office of Economic Opportunity (OEO). The major study is titled the "Rural Life Survey," poverty comerci The Develop] Statist: Survey ' The ide: were co was tak Blafi: Ea Mi Survey," and it has as its primary purpose a study of rural poverty conditions existing in a region which is noted for its commercial agriculture. The Rural Life Survey was conducted for the Economic Development Division by the Agriculture Department's Statistical Reporting Service in conjunction with a Pesticide Survey taken for the Farm Production Economics Division. The identification of households and Pesticide Survey were conducted in March, 1967, and the Rural Life Survey was taken in April - May, 1967. Place: East North Central States: Ohio, Indiana, Illinois, Michigan, Wisconsin (the "Eastern Corn Belt"). ngulation of Rural Life Segments: Farm operator households: Those with sales of less than $10,000. Nonfarm operator households: Open country households, defined as those living in land segments with a density of less than 50 households per segment. ngber of cases in Sample: Planned: 2,000 (about a .6% sample) Farm operator households 2,000 (about a .3% sample) Nonfarm households Total selected: 4,055 No interviews (835) or 860 defective (25) Final number of cases 3,195 Sample Design: 1. Counties were grouped into agriculturally similar contiguous groups containing about 4,000 qualifying farms. 2. From each group one county was selected with probability proportional to the 1959 number of farms in economic classes I-IV (i.e. with sales of $5,000 or more). 3. The land area of selected counties was divided into counting areas and segments in accordance with the Master Sample of Agriculture developed in 1943-45.2 The necessary number of sample segments for each selected county was drawn from an accumulated listing of segments using a random start and then successively applying the sampling interval (i.e. total segments in the county divided by the number of segments needed to get the number of farms desired). 4. The segments so selected were then divided into two groups. All households in the first group, consisting of every fourth segment, were asked only the Pesticide questionnaire. Households in the remaining three—fourths of the segments which had sales of $10,000 or more were asked the Pesticide questionnaire. 5. The remaining households of the second group were divided into farm and nonfarm households on the basis 2See R. J. Jessen, "The Master Sample Project and its Use in Agricultural Economics," Journal of Farm Economics, Vol. 29, No. 2 (May, l947);pp. 531—540. Th selecti states . 2 . TI regular Servic. design. Recess. 0f the series Supper housin The p: I defiui and p, Coma: Dorm. detai ibili of the Identification Sheet and then into sub— samples to give 2,000 cases of each using a random start and a given interval. For farm households, every second one was skipped and thenevery fifth one remaining was dropped. This area block sampling procedure allowed proportional selection of counties by the region, but not by individual states. 2. Survey The survey was administered by experienced enumerators, regular part-time employees of the Statistical Reporting Service. The survey instrument was a 28 page Questionnaire, designed as a single integrated document providing the data necessary for the accomplishment of the primary objectives of the study. The questionnaire was also organized into a series of more or less self contained sections which would support one or more specialized studies in such areas as housing, education, food patterns, and health conditions. The present study is one of these. The questionnaire was poverty, farm,and tax record definition oriented. It focused on family work activities and personal characteristics which affected 1966 income. It contained exceedingly detailed income information which permitted the formulation of numerous income concepts and detailed data on family composition which permitted flex- ibility in unit definition. The questionnaire also provided detailed farm information geared to the income-tax return. The: utilized consumer income. I unit we: been re] poverty second I 0f the I referre. He n0te Status Poverty of 1962 in 196: of the Simila; HoweVe: here. to the perSis There are two basic weaknesses in the measures of poverty utilized in this study.. First, although all sources of consumer unit money income were included in the measure of income received, the amount of assets owned by the consumer unit were not taken into account. Thus, a family could have been relatively wealthy and still be classified below the poverty line because of a low level of money income. The second weakness, related to the first, is due to the nature of the one-year cross-sectional study. Lanzilotti has referred to this as the condition of "temporary low income." He noted that about.twenty percent of the 1962 poverty status families in the United States had moved out of poverty in 1963. At the same time, about the same number of 1962 above poverty families moved into the poverty status in 1963.3 It can reasonably be assumed that some percentage of the Rural Life Survey poor families would respond similarly and receive above poverty incomes in 1967. However, no allowances for temporary low income will be made here. Job retraining might very well provide the solution to the problem of temporary low income as well as to persistent poverty problems. 3. Editing The process of editing the Rural Life Survey data was described in an editing manual prepared by the Economic _ 3Robert F. Lanzilotti, "The Poverty Syndrome: A Critical ReV1ew," Michigan State University Business Topics (Spring, 1966), pp. 39-49. .._\.... _. ___,_._._—? Research as much process The the fie] fig a. C. SérVic mimeo) 10 Research Service.4 The editing process was designed to assure as much accuracy and completeness as humanly possible. The process can be briefly summarized as follows. There were three main phases to the editing process: the field edit, the office edit, and the machine edit. The Field Edit was designed to do three things: a. Check completeness and sense of responses while respondents and enumerators could still be reached. b. Code certain items not already precoded for enumerators. c. Prepare entries for key punchers by seeing that all were legible and uniform as to decimals, etc. The Office Edit. This also had three main objectives: a. To prepare all schedules for the key punchers in a uniform manner. b. To code and preserve data not already coded. c. To bring all problems to the attention of the directors of the study. The work was performed by two groups of people: a. The office edit staff who edited the questionnaires and did the coding. b. The Problem Review staff who resolved the questions and set the codes. 4U. S. Department of Agriculture, Economic Research Service, Editinnganual: Rural Life Survey (Washington, 1966, mimeo). Stati The pr°068sin For the n the hYPOt of COHtir developed 11 The Machine Edit. The machine edit prepared the data for analytical manipulation. It had two phases: a. The consistency checks.’ b. Imputations It involved three groups of people: a. The computer staff performed the mechanics of the edit. b. The office edit staff searched the questionnaires to determine what was right. c. The Problem Review staff determined the actions to be taken and provided any additional information the computer needed for its operation. This part of the editing process had three main functions: a. To check the accuracy of the kevaunghng and to see that all cards had been put on tapes. b. To check and measure the internal consistency of the responses and the completeness of the replies. 0. To impute the missing information where this was essential to the analysis and then to measure the effect of the imputation. 4. Statistical Analysis of the Data The survey data was placed on computer tape for statistical processing on Michigan State_University's CDC-3600 computer. For the most part, cross tabulations of the data which tested the hypotheses of the study were done using the fAnalysis 0f Contingency Tables" (bivariate frequency distributions) developed by the Michigan State University Computer Institute for Socj cedures chapter: D. Hype Th: members Corn Be them pr retrain jobs wh househc T} test a1 categop retrai: for re‘ mElp'or ' will b in the 1. Hy thumF 1 that c quantj \ I II ’ n I uniVe: Janna. 12 for Social Science Research.5 Where other routines or pro— cedures were utilized, they are described in the following chapters, as are the specific details of each cross tabulation. D. Hypotheses to be tested The general hypothesis of this study is that certain members of rural, poverty—level households in the Eastern Corn Belt possess characteristics and attitudes which make them prime candidates to undertake job retraining. This retraining would, in turn, qualify them for employment in jobs which would provide income sufficient to raise the households above the poverty category. The sub-hypotheses which this study is designed to test are sub-divided according to the study's three main categories; education and retraining needs, potentials for retraining based on measures of retrainability, and potentials for retraining based on stated interest in retraining. The major hypothesesare listed in this section. All hypotheses will be discussed in detail, with rationalizations for them, in the chapters where they are tested. 1. Hypotheses regarding education and retraining needs: Assumptions Two basic assumptions will be made here. The first is that education and training data on the questionnaires can be quantified into a meaningful, usable measure of respondents’ 5Alan M. Lesgold, "Analysis of Contingency Tables: Act II," Technical Report No. 14 (East Lansing: Michigan State University, Computer Institute for Social Science Research, January, 1968). occupatior U. S. Depn basic edu: for those. now being these req‘ level req level res lbpothese a. holds in signifies SHE age b. holds in Which are w°rkers ; l3 occupational skill level. The second is that, by use of the U. S. Department of Labor's Dictionary 9£_Occupational Titles,6 basic educational skill level requirements can be established for those occupations in which low income individuals are now being trained under various Federal programs and that these requirements can serve as a proxy for a "minimum skill level requirement," a standard against which RLS poverty- level respondents can be judged. Hypotheses a. The worker members of rural, poverty—level house- holds in this region possess levels of education which are significantly below those levels possessed by persons of the same age in the general population of the region. b. The worker members of rural, poverty-level house- holds in this region possess levels of job skills or training which are significantly below those levels possessed by workers in the region's non-poverty-level households. c. The levels of training possessed by worker members of rural, poverty-level households in this region are below the "minimum skill level requirement" for entry into jobs for which low income individuals are now being trained under various Federal programs. 6U. S. Department of Labor, Selected Characteristics of Occu ations (Physical Demands, Working Conditions, Training— Time; 1966- A Supplement to the Dictionary of Occupational "l---——--— —- Titles, 3rd edition (Washington: U. S. Government Printing Office, I§66 . d among v region 2- Hy: other hssump I econom retrai reSpon Federa large educat rural Purpos reElior allowj in oil Small North Hlpotp holdS Which will j 14 d. There exists a considerable need for job retraining among worker members of poverty-level households in this region. 2. Hypotheses regarding "Potentials for Retraining" based on other studies of retraining programs: Assumptions It is assumed here that sufficient data on the socio- economic characteristics of participants in Federally sponsored retraining programs are available for comparisons with RLS respondents. It is also assumed that, although most of these Federal programs have been conducted in urban areas, a large percentage of the program participants were raised and educated in rural areas and are, therefore, similar to our rural population with regard to these variables. For the purpose of this study it will be assumed that no basic regional differences in these characteristics exist, thus allowing the use of studies of retraining programs conducted in other regions. This is necessitated by the relatively small number of such studies completed to date in the East North Central States. Hypotheses a. The "worker" members of rural, poverty-level house— holds in this region possess socio-economic characteristics which make them.potentially retrainable. b. The younger the poverty-level worker, the greater will be his potential retrainability. c. The more formal education possessed by a poverty level wor d. ployment, retrainin 3. Hypot on statec' Surn interest should or better j< The differenr and to tj Program f0r thes a11 reSp trdining any dOWn training Overstat The that res in a rel economi‘ Ohanges Hypothe: a. bi p0Ve: 15 level worker, the greater will be his potential retrainability. d. The less the amount of time spent in recent unem— ployment, the greater will be the worker's potential for retraining. 3. Hypotheses concerning "Potentials for Retraining" based on stated interest in retraining: Survey respondents were asked to indicate their interest in participating in a free retraining program should one be given locally which would qualify them for better jobs. ‘ The subjective nature of such responses and the possible difference between the responses to a hypothetical question and to the actual offer of participation in a retraining program are recognized. _No attempt will be made to adjust for these factors, however. It must.be assumed here that all respondents perceived the question concerning a free training program in the same way. It is also assumed that (any downward bias in respondents' abilities to recall past training experiences are offset by tendencies they have to overstate the number and the extent of such experiences. The major assumption underlying these hypotheses was that respondents weighed the possibility of participating in a retraining program in a manner consistant with an economic decision making rule, that is, in terms of expected changes in future income streams. Hypotheses a. There will be a relationship between the severity Of poverty and extent of interest in retraining, those responds interest higher 1' b. responds low inc< less in' c. interes Mom in retr as pote d. relatic those I intervj ploymer retrail recent 0f the S will r amount levels aVaila 16 respondents with relatively lower incomes being more interested in retraining than poor respondentswith relatively higher incomes. b. There will be a relationship between age of Arespondents and their interest in retraining, among all low income respondents,with older workers showing relatively " I less interest than younger workers. 0. There will be a relationship between degree of interest in retraining and potential costs of the retraining program; among poverty-level workers expressing an interest in retraining, progressively less interest will be expressed as potential program costs increase. 6. Among all low income respondents there will be a relationship between unemployment and interest in retraining; those respondents who were unemployed at the time of the interview and those who experienced relatively more unem- ployment the previous year will have more interest in retraining than those respondents with relatively less recent unemployment and those who were employed at the time of the interview. Several additional hypotheses will be tested. They will relate interest in retraining to such variables as amount of prior vocational training received, educational levels, problems encountered in finding employment, and availability for employment.‘ II. RURA A. Rur Lar and expe date bar host of primarij farmers incomes Upper c rather It been ge that we commodj assist; POVertE T1 Execut: Chargd \ 7‘ United (Degem 8 this II. RURAL POVERTY AND RETRAINING - LITERATURE REVIEW A. Rural Poverty Larson said (in 1967) that TStudies of the rural poor and experience with public programs aimed at rural poverty date back more than three decades in the United States."7 Most of these Studies and public prOgrams, however, were primarily concerned with the economic viability of farms and farmers and,_thus, with maintaining farm prices and/or incomes, or with poverty pockets such as Appalachia or the Upper Great Lakes where poverty was regional in scope rather than being isolated in regions of relative affluence. It is only in recent years that a national concern has been generated for individual victims of rural poverty and that we have faced up to the fact that existing agricultural commodity programs have not been, and cannot be, of much assistance in alleviating conditions andcauses of rural poverty.8 This national concern was reflected in the President's Executive Order No. 11306 of September 27, 1966, which charged the National Advisory Commission on Rural Poverty 7Olaf-F. Larson, "Discussion: Rural Poverty in the United States", Journal of Farm.Economics, Vol. 49, No. 5 (December, 1967), p. 1235. 8James T. Bonnen, "Rural Poverty: Programs and Problems" Journal of Farm Economics, Vol. 48, No. 2 (May, l966).pp. 452-65 and James T. Bonnen,"The Distribution of Benefits from selected U.S. Farm Programs,“ in National Advisory Commission On Rural Poverty, Rural Poverty in the United States (Washington: U. 8. Government Printing Office,l968Y,pp.46l-505. 17 I To curr rura and prob For ment The che commis: concer: T ll mil detail Povert elsewh comple that t reside of pov rUral 18 To make a comprehensive study and appraisal of the current economic situations and trends in American fural life, as they relate to the existence of income and community problems of rural areas, including problems of low income,~the status of rural labor, including farm labor, unemployment, and underemploy— ment and retraining in usable skills... The charge continued in more detail. The findings of the commission are reported, with recommendations, in The PeOple Left Behind, and the volume of technical studies prepared for the commission presents 32 illustrations of the new concern for rural poverty.10 The Rural Poverty Commission estimated that there were 14 million rural poor in the United States in 1965. The details on the incidence and distribution of this rural poverty have been very thoroughly presented and analyzed elsewhere, and require only brief mention here. The most complete rural poverty study was that of Bird, who noted that half of all the poor people in the United States resided in rural areas in 1959, and that the incidence of poverty was more than twice as great in rural as in non- rural areas.11 The 1967 report on manpower of the U.S. 9National Advisory Commission on Rural Poverty, The People Left Behind (Washington: U. S. Government Printing Office,l967b p. vi (emphasis supplied). 10National Advisory Commission on Rural Poverty, Rural Epyerty in the United States-(Washington: U. S. Government Printing Office, 1968). 11Alan R. Bird, Poverty in Rural Areas of the United States, Agricultural Economic Report No. 124 (Washington: U. S. Department of Agriculture, Economic Research Service, November, 1964). Departmen poverty a was noted presently rural poo poverty : extent a: As : alleviat Poverty, veoqraph Yet, in re(lions, were com Other tr. 19 Department of Labor also presented a concise summary of poverty and employment conditions in rural America.12 It was noted by Tweeten that different classification schemes presently in use give quite different pictures of the current rural poverty situation, but that by any measure, rural poverty is extensive in this country; he discussed the 13 extent and causes of and potential cures for rural poverty. As noted above, most national programs designed to alleviate rural poverty have been focused upon regional poverty, that is, concentrationsoof poverty in large geographic regions including the Upper Great Lakes region. Yet, in comparing this region with the Ozark and Appalachian regions, Tweeten noted that a higher percentage of all farms were commercial farms in the Upper Great Lakes than in the other two regions (58 percent as compared to 46 percent). He concluded, on the basis of the farm operator level of living index, that "the Upper Great Lakes was not a serious area of rural poverty in 1959."14Yet Michigan and Wisconsin had over 200,000 rural families with net cash incomes under $3,000 in that year (see Table 1, Chapter I). 12U. S. Department of Labor, Manpower Report of the Egesidents and A Report on Manpower Requirements, Resources, ggilization,~and~Training (Washington: U. S. Government Printing Office,-April, 1967),pp. 101-121. . 13Luther G. Tweeten, Ruraleverty: Incidence, Causes, . Egg Cures, Processed Series P—59OR (Stillwater: Oklahoma State University, Experiment Station, July, 1968). l4Ibid., pp. 3—4. B. Rt Tb concer this s supper school farme: serio One c yet t the a regic this farms fami] great Stror Shrp 0n ti in I“ 3180 POVe POVE 20 B. Rural Poverty in Regions of Commercial Agriculture The apparent bias of rural anti-poverty programs to concentrate only upon regional poverty is reflected in this-statement by Tweeten: "Progressive farmers can support favorable attitudes and institutions such as good schools in commercial farming areas in which only a few farmers are poor. Thus area-wide rural poverty is a more serious problem than poverty interspersed among plenty."15 One cannot deny the seriousness of area-wide poverty, yet two fallacies of this statement stand out. First is the assumption that "only a few farmers are poor" in any region which is considered a commercial agriculture area - this is not always the case. Secondly, the focus on only farmers in such areas overlooks all the rural nonfarm families, and it is precisely among this group where the greater share of rural poverty exists today. As Booth has strongly argued, the "farm problem", defined as one of asurplus labor and inadequate income for farmers, is well on the way to being solved; the substantial poverty problem in rural America is no longer connected with farming.l6 Statistics on median income levels by counties have also helped foster a complacent attitude toward rural poverty problemsin commerical agricultural regions. Data 15Ibid., p. 39. 15E. J. R. Booth, "The Economic Dimensions of Rural Poverty," American Journal of Agricultural Economics, Vol. 51: No. 2 (May, 1969),pp. 428—42. on the cash i 1959 u the E2 of onl Numbe famil famil State East Centr m Indie Illir Michj WiSc( Soon m Seen Kent MisS Tennl TEXa SOur St is 1 Sent polls Vari 21 on the number of counties in each state in which the median cash incomes of rural families was less than $3,000 in 1959 would indicate that there is little rural poverty in the East North Central States (Table 2). The small total of only 33 counties with low median incomes in this region Table 2. Number of counties in which median cash income of all rural families, only rural farm families, and only rural nonfarm families was less than $3,000 in 1959. East North Central and Selected Southern States Total No. All Rural Only Rural Only Rural State~ of Families Under Farm Families Nonfarm. Counties 3,000 Under $3,000 families under _f $3,000 East North Central Ohio 6 l 5 - Indiana - - - - Illinois. 18 5 13 — Michigan 3 - 3 - Wisconsin 6 ~ 6 - Southern Alabama 57 47 10 - Georgia' 115 96 19 — Kentucky 9 2 7 3 1 8 1 ‘ Mississippi 75 68 6 1 Tennessee 78 67 ll — Texas 115 82 26 7 ¥ Source: Alan R. Bird, Poverty in.Rural.Areas of the United States, U. S. Department of Agriculture, EconomIc Research Service, (Washington, November, 1964), table 4, pp. 9-10. is in sharp contrast to the number of counties in selected Southern states with median rural incomes below this crude Poverty line. But the median tells us nothing about the Variance in cash incomes. Particularly for the rural nonfarm w...— 1.." ' vrq‘ group media famil The e that media rural devel that be 9: tuoi: 0f pr but Econ of t agri manp tol- wit] SOC (De 22 ”Cup with only six East North Central counties in which idian cash incomes were below $3,000, it ignores all .milies whose incomes still are less than this poverty line. e evidence presented in Chapter I, Table 1, demonstrates at much more rural poverty exists in this region than dian income levels would seem to indicate. Ruttan was among the first to recognize that our ral development program must go beyond the regional Velopment approach. He drew attention to the possibility at the total number of rural families in poverty might jgreater in regions of commercial agriculture than in the Jical rural poverty regions even though the incidence poverty was clearly lower in the commercial regions.l7 Prior to the inauguration of the present set of studies the Economic Development Division and the Office of nomic Opportunity, there were no comprehensive studies the exact nature and extent of rural poverty in commercial iculture regions. Rural Needs for Job Retraining The awareness of the need for deliberate national Dower policies and programs to provide employment skills marginal workers developed in the early 1960's along r the awareness of the apparent contribution that education, l7Vernon W. Ruttan, ”Agricultural Policy in an Affluent ety," Journal of Farm Economics, Vol. 48, No. 5 ember, 1966» pp. 1100-20. or in to ca econ; in Cb State belie an e cont and nigr priv assu the assu must man labc Hat) He 1 who 80c the His V0] 23 r investment in human capital as it became fashionable ; call it, had made to the overall growth in the American :onpmy. (The theory of human capital investment is discussed a Chapter III). The massive outmigration from rural areas in the United :ates over the two decades 1940—1960 had fostered the dief that all that was needed to eventually bring about ‘equilibrium condition with respect to rural labor was ntinued high levels of aggregate demand in the economy d additional net rural outmigration. If rural-urban grants lacked employment skills, it was assumed that ivate industry would provide the necessary training to sure their employability as it had done in the decade of a forties with its tight labor market. It was also sumed that necessary adjustments in agriculture which st be made concurrently with changes in the capital— u ratio would automatically follow the decline in the nor force. In a 1960 address to the American Economic Association, haway noted that both of these assumptions were faulty. pointed out the problem faced by unskilled migrants found themselves to be marginal members of the nonfarm iety, and that "mere out—movement of labor from agriculture 1 necessary, but not sufficient, condition to achieve :e adjustments" in agriculture.18 Hathaway called for l8Dale E. Hathaway, "Migration From Agriculture: The orical Record and Its Meaning," American Economic Review, 50, No. 2 (May,l960), p. 387. publi probl the a the I areas rural in it who a to t] Bishv urea are job in r repg Stat in 3 You Tra in R 24 lic policies to cope with these non-self—correcting clems. The economist who led much of the early thinking in 1area of human capital was also among the first to identify need for job training as a critical problem in rural as. Schultz stated that a major cause of poverty, 11 and urban, was a "long standing disequilibria rooted Lnadequate investment in particular classes of people, are therefore poor." 19 He was referring specifically 3he poor quality of rural education. As early as 1960 top also argued that "many farm people will likely need tly increased training and skills before they can earn asonable income from nonfarm employment."20 By 1967 national leaders were well aware of the need for retraining, as well as for improved public school systems, Jral America. This awareness was reflected in the 1967 :t on manpower of the U. S. Secretary of Labor which ad: Educational and training facilities available to rural residents have been shockingly deficient in both quantity and quality as compared with those in urban communities. A revolutionary upgrading of the rural labor force is called for. Called for also, and equally important, is the establishment of educational and training facilities for rural youth, so that they 19T. W. Schultz, "Public Approaches to Minimize Poverty," 3 Fishman (ed.), Poverty Amid Affluence(New Haven: Jniversity Press, 1966), p. 176. 20C.E. Bishop, "Increasing Mobility of Labor Through .ng Programs"(1960), in R.J. Hildreth (ed.), Readings 'icultural Poiic (Lincoln: University of Nebraska 1968 , p. 439. nations calling Federal one age provide skills 0- R1 S: ment A. hot (M: Expenst retrai PrOgra Person of the more f \ 2 2 M 4 8part Train: Vanel Secret 25 will not be forced to enter the laborforce under the.same handicaps from which their fathers suffer.21 The Rural Poverty Commission also reflected this tional awareness when it recommended, in addition to lling for close coordination and administration of all deral manpower development and training programs under e agency, "That adequate job training opportunities be ovided for (rural) workers to maintain and upgrade their ills and to qualify for better jobs."22 Rural Areas' Share in Retraining Programs Since the passage by Congress of the Area Redevelop- rt Act in 1961 and the Manpower Development and Training : (MDTA) in 1962, a considerable amount of effort and >ense have been directed toward developing programs of :raining for unemployed workers, largely under MDTA rgrams. The history of these acts and the number of sons retrained are well documented.23 Also, evaluations the benefits and costs of such prOgrams are becoming e.frequent (see Chapter V for a discussion of these). [_— 21U. S. Department of Labor, Manpower...(l967), p. 119. 22National Advisory Commission on Rural Poverty, HPeople Left Behind, p. 34. 23See Sar A. Levitan, Federal Manpower Policies and rams to Combat Unemployment (Kalamazoo: The W. E. hn Institute for Employment Research, 1964); and U. S. rtment of Health, Education and Welfare, Education and ning:...(Washington: U. 8. Government Printing Office, ous years); this is the annual report of the HEW etary to Congress on MDTA training. However, in prime conducts There is and a re approach tvo othe The Rure new prog The discoura POVerty "Such at training rural pc S°me pot on the u if Prng define} forCe . ll: 26 wever, very few retraining programs have been established primarily rural areas, and even less research has been uductéd into the effectiveness of such retraining. are is one study of rural retrainees in Eastern Kentucky 1 a report of the Concentrated Employment Program (CEP) sroach in pilot form being conducted in Minnesota and bother locations by the Federal Employment Service.24 Rural Poverty Commission itself stated that "Few of the programs have had a major impact on rural America."25 The evidence from other sources appears equally couraging. Discussing the failure of public anti— erty programs to reach the rural poor, Clawson stated, 3h efforts as have been made to develop rural manpower ining programs have found it difficult to locate the 11 poor in need of such training."26 And in discussing a potential effects of alternative Federal policies the welfare of rural people, Martin thought it "doubtful rograms now authorized (in 1966) will arrest the growing ciency in educational achievement of the rural labor e."27 24John S. McCauley, "Manpower Development in Rural 5," Employment Service Review, Vol. 5, Nos. 3 and 4 :h—AprIl, 1968), pp. 10—15 ff. 25National Advisory Commission on Rural Poverty, Deople Left Behind, p. vii. 26Marion Clawson, "Rural Poverty in the United States," 1al of Farm Economics,-Vo1. 49, No. 5,(December,l967), 330. 27Lee Martin, "Effects of Alternative Federal Policies lfare of Rural People," Journal of Farm Economics, 48, No. 5 (December, 1966» p. 1275. rural only Tweet all t nonfz of L2 P9106 Sign; Perc and . perc. year Most hand DTOg Labc Halt more Wt: 27 What has been the actual participation record of the rral poor in Federal manpower retraining programs? Looking rly at trainees who were farm workers prior to training, eeten noted that they comprised less than 2.5 percent of l trainees in 1966.28 However, when rural farm and rural nfarm workers are both considered, the U. S. Department Labor data indicated that they comprised roughly 19 rcent of the total number of 1966 trainees.29 The more gnificant facts revealed by this data were that over 70 rcent of all rural traineés were under 35 years of age, 1 only 22 percent of the institutional trainees and 15 :cent of the on—the-job trainees had less than nine rrs of education - meaning that those rural workers at in need of retraining because of age and educational rdicaps were the least likely to be included in retraining rgrams. The 1969 manpower report of the U. S. Department of or estimated that 20 percent of all MDTA trainees from 3 to 1968 were rural residents. The report stated that though recent manpower programs have been concentrated e in urban than in rural areas, they have had some impact ide the cities."30 In 1968, there were 13 Concentrated 28Tweeten, Rural Poverty..., p. 49. 29U. S. Department of Labor, Manpower Report...(l967h 275. BoIbid. (1969), p. 112. Employmer to 63 in It v as yet rt poverty the occu reviewin Tweeten and too rural pr Effectit mployment Programs operating in rural areas as compared 0 63 in urban areas. It would appear that rural poverty problems have not 5 yet received adequate attention from Federal anti- overty programs, particularly those designed to upgrade he occupational skills of-the rural labor force. After eviewing the-broad array of Federal anti-poverty programs, eeten concluded that "A combination of too many programs nd too little funds have made Federal efforts to relieve ural poverty imaginative and promising, but often in— ffective."31 31Tweeten, Rural Poverty,.., p. 63. III. time of It to p the perc soci are POW I. LABOR THEORY APPLIED TO RURAL POVERTY AND RETRAINING Introduction Any socio-economic problem worthy of a researcher‘s me and effort must be capable of being analyzed in terms received economic theory. Stated differently, if it is pass the tests of external consistency and workability, e received economic theory must be applicable to those rceived problems which command the time and effort of cio-economic researchers. In-thls chapter theoretical concepts of labor economics e briefly examined and related to problems of rural verty and to contemporary concepts of manpower retraining. Marginal Productivity and Bargained Wage Theories. The two prevailing labor theories today are marginal oductivity theory and the bargaining theory of wages. t contributions do these theories make to an understanding the causes of and the possible remedies for conditions rural poverty? Modern labor theory had been primarily concerned with laining the level of wages and changes in this level. er perfect competition, with the assumptions of fully loyed resources, perfect mobility of factors of production, fect knowledge of prices and wages on the part of all cerned, and rational economic motivations, marginal ductivity theory "postulates the theory that, over the g run, each factor of production will tend to receive uneration which is equal to its marginal contribution to 29 the pror be equal In less tht legal m tendenc labor e or negc Product Fr Wages t for cor change: Produc to pay aPPI‘oa run Va mdin wa96s, than I \ The p haunt he production process."32 Thus the wages of workers should e equal to their marginal productivity. In recognition of the fact that labor markets are less than perfectly competitive due to such factors as Legal minimum wages, union wage contracts, and the apparent :endency of large corporations to not maximize profits, .abor economists have formulated the theory of bargained r negotiated wages to replace, or to expand upon, marginal roductivity theory. Four basic criteria upon which collectively bargained ages depend, according to Lester, are: (1) wage scales or comparable work in other plants or industries; (2) hanges in the cost of living; (3) changes in labor roductivity; and (4) the company's or industry's ability 0 pay wages, usually measured by profits.33 A pragmatic proach such as this helps to explain some of the short— n variations in wage levels. Even so, labor productivity d increases in it are utilized when bargaining for higher ges, although it is average worker productivity, rather an marginal productivity, which is usually considered. explain the level of long-run wages, Morgan noted that 32Chester A. Morgan, Labor Economics (Homewood, Illinois: e Dorsey Press, 1966), p. 61. 33Richard A. Lester, Economics of Labor (New York: cmillan, 1964), p. 286. marg most wage wage very rele More that its Rec per met Th 31 marginal productivity theory still appears to give the most valid answers. He cited the ratio of manufacturing wages to value added by manufacturing and the ratio of wages to gross national product, ratios which have varied very little over time, as historical indicators of the relatively close relationship between wages and productivity. Morgan concluded that "While the so—called 'bargaining theory' is perhaps the best explanation to date of the short run wage-setting mechanism in the industrial economy, it is also probable that the marginal productivity theory is the most valid explanation thus far of the basic factors determining long—run wage levels and tendencies.34 Does each labor factor in fact receive wages equal to .ts marginal contribution to the production process? regional differences in wages paid for identical work performed for the same firm and continuing wage differences f from ten to fifteen percent between firms in.local etropolitan labor markets seem to indicate it does not.35 he bargaining theory of wages, recognizing the generally norganized nature of labor in rural areas, can explain Job of the causes of low wages in rural areas, and much E the short run variation in regional and local wage levels. >wever, data on relative labor earnings by industries, rd changes in them from 1948—66, indicate that many of these 34Morgan, Labor Economics, pp. 73-74. 35Lester, Economics of Labor, p. 267. variation they rais ductivity Man} of margir because ' useless. theory ti social g from anx testing by empir For John to whic assumpti that “tr the mair appears theory. r L.” ' 1 f-u’ -, 2T; 7A;—:Z_-£._g=;;\.;_:'r; :f‘w—m _ v—~ir»‘“—~' .____.¥ 32 riaticns appear to be of a long—run nature and, therefore, ay raise questions about the validity of marginal pro— :tivity theory.36 Many writers have assailed the "unrealistic" assumptions marginal productivity theory and have concluded that, :ause the assumptions are unreal, the theory itself is Bless. In discussing recent developments in economic aory that have significant implications for broader :ial questions, Johnson noted that emphasis has shifted >m anxiety about the reality of the assumptions to :ting the "robustness" of the conclusions of a theory empirical estimation or theoretical investigation. 'Johnson, the "robustness" of a theory meant the extent which its conclusions survive under changes in the “37 He concluded umptions from which it is derived. t "the results have almost invariably been to confirm main propositions of neo—classical theory."38 It ears appropriate, therefore, to accept marginal productivity >ry on the basis of its performance, but at the same a to question the assumptions underlying the theory. 36Hyman P. Minsky, "Adequate Aggregate Demand and the itment to End Poverty," in National Advisory Commission ural Poverty, Rural Poverty..., pp. 572-74. 37Harry G. Johnson, "The Economic Approach to Social tions," Public Interest, No. 12(Summer, 1968), p. 71. 38Ibid, p. 72. i Obvi assumptir economic full empi ship beta poverty : and Schul farm lab! that, wi releasin and prov industri under ot would run of forms a tights rural pr The all worl selutior 0f the Obvious deviations from reality of the theoretical umptions have received considerable attention from nomic theorists. The inaccuracy of the assumption of 1 employment requires little comment here. The relation- p between the degreevof unemployment and the extent of erty in our economy is clearly discussed by Minsky,39 i Schuh has demonstrated the relationship between the 'm labor force and the total economy.40 Minsky noted it, with the technological revolution in agriculture .easing a very large number of workers from this sector i providing a relatively elastic supply of labor to the Lustrial sector, "a rise in aggregate demand that might, er other circumstances, lead to inflationary pressure ld under these circumstances lead to faster absorption formerly rural population into the urban society. Thus, ighter urban labor market would go far to eliminate al poverty.41 The lack of perfect knowledge of wages and prices by workers is also clearly recognized. Recommended ations to this problem generally call for an improvement :he job information and placement services provided by 39Minsky, "Adequate Aggregate Demand...," pp. 562-80. 40G. Edward Schuh, "Interrelations Between the Farm r Force and Changes in the Total Economy," in National sory Commission on Rural Poverty, Rural Poverty..., L70—184. 41Minsky,"Adequate Aggregate Demand...flp. 572. the prese example 0 would aff by the Na Nume mobility have alsc imperfect some men' the soci: productit Seve on occupr associate upon incv market i: (postulat. reductio Whi do not f on the s conclusi 34 ,present state and Federal employment services. A good mple of recommended changes in these services which lld affect rural poverty victims in particular is presented the National Advisory Commission on Rural Poverty.42 Numerous labor market imperfections which inhibit the >ility of labor and thus cause or aggravate poverty 7e also been detailed. Batchelder lists among these perfections the misallocation of educational resources, he men's Unwillingness to change, family obligations, a social minimum wage, discrimination, and rising nductivity.43 Several recent studies have examined available data occupational mobility to determine the variables :ociated with mobility and the effects of mobility tn income. Of concern here is whether or not the labor 'ket is functioning in accordance with theoretical tulates and what implications its functions has for uction of rural poverty. While it is unfortunate that the several studies not focus on the same groups in the labor force nor :he same questions about these groups, so that the :lusions of the studies are really not comparable, 42National Advisory Commission on Rural Poverty, People Left Behind, pp. 28—31. 43Alan B. Batchelder, The Economics of Poverty (New : John Wiley and Sons, l966),pp. 83-95. i each does market. Two of geogra States we Lansing a mobile on However, of mobilt to disapl to inclut 1957 and male wori earnings nommobil dustry 0 that sim and Morg moved in in that industry non-move The from reg \— 4 4: Geoyrapk Vol. 2 , 451 Mobilit} M2 35 ‘ > fiéviffib?iz_g¥;:§;. ;, teach does shed some light on the functioning of the labor ‘market. Two studies which focused on the effects on income 10f geographic labor mobility between regions of the United 44 45 States were those of Lansing and Morgan and of Gallaway. J t Eansing and Morgan found that income levels of geographically t Wobile workers are less than those of nonmobile workers. however, by comparing income levels of comparable groups h bf mobile and non-mobile workers this relationship tended I go disappear. By extending the regional mobility analysis %0 include an analysis of interindustry mobility between i957 and 1960 as well, the Gallaway study showed that mobile hale workers who did not change industry of employment had earnings slightly higher (about three percent)than similar tonmobile workers, and mobile workers who also changed in— ustry of employment also had even slightly higher earnings hat similar nonmobile workers (about five percent). Lansing nd Morgan also found that the income levels of those who Dved into a region were lower relative to those already 1 that area; again by adjusting in-movers for change in .dustry Gallaway found them to have higher income than n-movers in seven of the nine regions. These two studies show conflicting effects upon income 3m regional mobility of workers, depending upon whether 44John B. Lansing and James N. Morgan, "The Effects of graphic Mobility on Income," Journal of Human Resources, . 2, No. 4 (Fall, 1967): pp. 449—460. 45 Lowell E. Gallaway, "The Effect of Geographic Labor Llity on Income: A Brief Comment, "Journal of Human >urces, Vol. 4, No. 1 (Winter, 1969), pp. 103-109. * industry It appea arises 1 workers yet only doubtfui mobile y studies A centrat the eco study b by Hath agricul Ga agricul apparer oultur back In 4 Labor, 1967) , 36 yindustry of employment changes are also considered or not. Jt appears, however, that a major weakness in both studies arises from the fact that only a small percentage of mobile orkers actually move across regions in the United States, et only regionally mobile workers were considered. It is oubtful if any valid generalizations can be made for all obile workers from the limited sample used in these tudies. A second type of mobility study is that which con— entrates on mobility of labor from particular sectors of he economy. In relation to rural poverty, a mobility 460f hired agricultural labor and two 47,48 ‘tudy by Gallaway >y Hathaway and Perkins of both hired and self-employed Lgricultural workers are of particular interest. Gallaway found considerable outmigration of hired gricultural labor from agricultural employment which was pparently responsive to income differences between agri— llture and other sectors. However, relatively large ick migration into agriculture led to relatively small 46Lowell E. Gallaway, "Mobility of Hired Agricultural bor," Journal of Farm Economics, Vol. 49, No. 1 (February, 67), pp. 32-52. 47Brian E. Perkins and Dale E. Hathaway, The Movement Labor Between Farm and Nonfarm Jobs, Agricultural >eriment Station, Research Bulletin No. 13 (East Lansing: migan State University, 1966). . 48Dale E. Hathaway and Brian E. Perkins, "Occupational ility and Migration From Agriculture," in National isory Commission on Rural Poverty, Rural Poverty..., 185—237. fin vie‘ sector. as refl costs 0 artific younger older w hypothe willing the la- iOb ch. conclu in the H large agriou relati becaus have 1 Prior mobil‘ Came in hi 3130 \ $1,92 37 (in view of income differentials) net outmigration from this sector. The large income differentials49 were explained as reflecting the combination of costs of movement, including costs of retraining for other types of employment and artificial barriers to labor mobility. Gallaway noted that younger workers predominated among the outmigrants whereas older workers predominated among the inmigrants. He hypothesized that voluntary mobility (a job change the worker willingly elects to make) characterized the former, whereas the latter could be considered involuntary mobility (a job change which the worker is forced to make). The conclusion reached was that dynamic equilibrium existed in the agricultural labor market. Hathaway and Perkins found the same conditions of large gross but relatively small net outmigration from agriculture for all agricutural workers. They found no relationship between off—farm mobility and income, apparently because the most successfully mobile were more likely to have been multiple jobholders and have higher incomes prior to changing employment than were the less successfully mobile. Occupational mobility and income gains were most common for the young, whites, farm wage workers, living in high income counties located close to SMSA's. They also found that most farm workers do not move far when 49Gallaway, "Mobility...," p. 35. The 1960 level of mean estimated earnings of hired agricultural labor was $1,922; the 1960 all—industry average was $4,924. — changing Southerr tance- n‘u‘ nobility nost."5‘ market t noted t1 nigrati: is reduw of job , success It sector simple employm mowing 38 :hanging jobs, the only exceptions being the young and the Southern Negro and there was no evidence that long—dis- tance migration pays economic dividends. In general,"the mobility process works less well for those who need it most."50 Viewing the situation in the agricultural labor market as one of dynamic disequilibrium, Hathaway and Perkins noted that measures are not needed to speed the rate of out— igration from agriculture. What is needed, they stated, '5 reduced unemployment in the overall economy and programs f job retraining to better equip those who are not uccessfully mobile for employment in the nonfarm economy. It thus appears that workers in the agricultural ‘ector do respond consistantly with predictions from a :imple model of expected income gains in moving from farm mployment to sectors of higher expected incomes. Back— oving is also a consistant choice in View of actual income xperience. Is this market characterized by barriers to >bility? The large gross movement both ways would suggest ,at it is not. However, if mobility is viewed in terms successfully remaining in the new occupation on a Ig-term basis and earning above poverty—level incomes, en there are apparently major barriers to mobility from agricultural sector. No mobility studies to date have focused their attention hose workers who are residents of rural areas but who 50Hathaway and Perkins, "Occupational Mobility...,"p. 212. — are not studies sector barrier be ass all run Ti that 1: do not product employ] theory been m; and co: And it charac produc an und 39 are not employed in the agricultural sector. The above studies illuminate the situation which exists among farm sector workers and.cIeafilyillustrates the need for removing barriers to their occupational mobility. It can probably be assumed that similar conclusions would be reached for all rural residents with poverty—level incomes. The above discussion has supported the contention that labor market conditions in the United States presently do not coincide closely with the assumptions of marginal productivity theory. A situation of less than full employment led to the development of the complex Keynesian theory of macro—economics. Numerous prescriptions have been made for improving the flow of labor market information, and certainly much improvement has already been made. and it appears that numerous barriers to labor mobility still :haracterize this market. Does all this mean that marginal )roductivity theory can contribute nothing substantial to m understanding of rural poverty conditions? It does not. he assumptions are made for a system in static equilibrium; S long as dynamic adjustments are continually being made n an economy, even if perfect competition were to exist, 1e failure to attain an equilibrium implies that static ssumptions will not hold. The real test for marginal productivity theory is .ether or leads to policy prescriptions which will, when rried out, improve the functioning of the labor market d in so doing reduce the extent of poverty—level incomes. — The follow theory has C. Human An a1- variation attacking theory, i marginal This appr because w Whereas t entry-lev in a part recognize between 6 occupatic due to ti marginal homogene differen Rec capaciti approach economic PIOCES S 40 he following section indicates that on this basis the heory has made a major contribution. - Human Capital Investment An alternative approach to the question of long-run Variation in wage rates between industries, as opposed to attacking the assumptinns of marginal productivity labor :heory, is to apply the theory directly and examine the marginal productivity of workers in various industries. Phis approach leads to the conclusion that wages differ oecause workers differ in their marginal productivity. Whereas this difference has long been recognized between entry-level workers and trained or experienced workers in a particular occupation, it has not been generally recognized that such productivity differences may exist aetween entry-level or marginal workers in the different bccupations and industries. This lack of recognition was due to the additional assumption or simplication of marginal productivity theory that treated labor as a romogeneous factor of production rather than recognizing lifferences infilabor skills. Recognizing that workers do possess different human apacities has led to what Johnson considered "a new pproach to the economics of labor - more broadly, the conomics of the role of human beings in the productive rocess — based on the concept of 'human capital'."51This 51Johnson, "The Economic Approach...," p. 73. concept equipme ductive through vestmer Th extensi Vaizey that si import: econom: and to and it: dentia.‘ 1960 mi 41 ancept envisages workers as particular types of capital guipment employed in the production process whose pro— uctive capacity is developed by a process of investment, hrOugh education and on—the-job training, and this in— estment yields a return over the worker's lifetime. The history of this human capital approach has been :xtensively documented andneeds only brief mention here. 'aizey summarized the View of classical economists, noting :hat since Adam Smith economists have been aware of the .mportance of human resource development.52 Yet modern economists tended to concentrate upon physical capital and to neglect the human factor. This apparent neglect and its implications were noted by Schultz in his presi— lential address to the American Economic Association in .960 when he said:53 The failure to treat human resources explicitly as a form of capital, as a produced means of production, as the product of investment, has fostered the retention of the classical notion of labor as a capacity to do manual work requiring little know~ ledge and skill, a capacity with which, according to this notion, laborers are endowed about equally. This notion of labor was wrong in the classical period and it is patently wrong now. The development of the concept of human resources ad been well summarized by Harbison and Myers,54 52John Vaizey, "What Some Economists Said about iucation," Chapter I, The Economics of Education (London: Lber and Faber, Ltd.,l962)- 53Theodore W. Schultz, "Investment in Human Capital," e American Economic Review, Vol.5l, No. l (March,l96l),p,3. 54Frederick Harbison and Charles A. Myers, "Concepts Human Resource Development,"Chapter I, Education, Manpower 3 Economic Growth (New York: McGraw Hill, 1964). fl Tweeten,55 smarizec‘ the human observed 9 level, ad to increa: Ginzl consideri: of econom in :econom the chang human fac chology w growth an . from a. de to increa education three maj the deve lngton : Service , 56E (New Yo 42 veeten,55 and others. In general, the studies which they ummarized found that education, through enhancement of he human factor, had made a major contribution to the bserved growth in national economies. On the individual .evel, additional education contributed considerably :0 increased expected lifetime earnings. Ginzberg has noted that a number of economists are :onsidering the human resource factor as the key determinant >f economic progress, and he considers this "a revolution in-economic thinking that may yet exceed in significance "56 This new emphasis on the the change wrought by Keynes. uuman factor in production, the emergence of a new psy— chology which emphasizes the dynamic facets of human growth and development, and a transformation of the economy from a dependence on unskilled, physically strong laborers to increasing dependence on men of general and specialized iducation and skill have been, according to Ginzberg, the ihree major forces responsible for the modern concern for ihe development of human resources. Some studies have attempted to measure the contribution o lifetime earnings of various amounts of formal schooling. 55Luther G. Tweeten, The Role of Education in Alleviating ural Poverty,-Agricultural Economic Report No. 114 (Wash— Igton: U. S. Department of Agriculture, Economic Research ervice, June, 1967). 56Eli Gdnzberg, The Development of Human Resources Iew York: McGraw Hill, 1966), p. 3. fl All stt benefit in mak: noted 4 sonclu at all 10 per 50-per studie educat the on man hmd the is a high l both i 43 .1 studies presented only general approximations of the snefits of education and numerous methodological problems 1 making the computations remain unresolved. Yet Ribich >ted that “practically all of the studies arrive at the >nc1usion that the payoff rate for continued education : all levels is remarkably high - something in excess of ) percent for college education and perhaps as high as ) percent for increments at lower levels."57 Comparative :udies of the payoffs from vocational versus general lucation have been limited, the work by Ribich being is only one which attempted to make such comparisons 1 a national level with a standardized methodology. He >und that "vocationally oriented training, at least in re form of recent manpower training programs, exhibits higher rate of payoff than does general education."58 The evidence on the role of education in enhancing ith national economic output and individual lifetime .come streams appears very conclusive. Most economists, erefore, automatically ascribe to education a major sponsibility.in alleviating family poverty level incomes. y Pt is the nature of the relationship between education h poverty? Wolfbein claimed that about 40 percent of l . . . . p decline in poverty among families in the United States i 57Thomas I. Ribich, Education and Poverty (Washington: a Brookings Institution, 1968), p. 9. 58110101., p. 97. fl betweer of edut period. of the six otl Thurow poverty among 1 eight 1 comple' had c0] improv. of the cautio improv low ed worker 44 between 1950 and 1964 resulted from the increased level of educational attainment of household heads during the period.59 Also, by means of a more precise specification of the relationship between poverty and education (and six other explanatory variables) in a regression model, Thurow found education to be significantly related to poverty. The 1963 incidence of poverty was 44 percent among households with family heads having less than eight years of schooling, 10 percent for heads having completed high school, and under 5 percent when the head , i had completed four years or more of college.60 Thus, improvements in education are, according to Thurow, one of the most effective ways of eliminating poverty. While ‘ :autioning that accelerating the decline of poverty by improving educational levels would be difficult because low educational attainments were concentrated among older workers, he concluded that investment in human resources, iogether with equal rights for Negroes and the push towards ull employment, could potentially make significant re- uctions in the number of families living in poverty. Tweeten has made the most comprehensive attempt to elate the findings on human resource development to rural 59Seymour L. Wolfbein, Education and Training for Full ployment (New York: Columbia University Press, 1967), 102. 60Lester C. Thurow, "The Causes of Poverty," Quarterly ournal of Economics, Vol. 81, No. 1 (February, 19675,p.46, poverty determi 1959 ti 57 pen years c where . with TI 10 per as a w in the ’I betwee hOWeve secti( and i1 is not to Ea: reall; \ 45 poverty conditions, discussing social as well as economic determinants of rural poverty. He stated that Education has a two-fold effect on rural poverty. First, it increases skills of persons, potentially raising farm management levels as well as increasing suitability for nonfarm jobs. But equally important may be the second effect of education — enhanced motivation and aspirations for improved earnings and living standards, and changed attitudes more consistent with frictionless assimilation into a new environment. Estimates by Bird, cited by Tweeten, showed that in 1959 the incidence of poverty among farm families was 57 percent in families whose heads had less than eight years of schooling; this dropped to 31 percent in families where the hpad had twelve years of schobling. Comparison with Thurow's data, which showed rates of 44 percent and 10 percent respectively for these two groups in the nation as a whole, gives an indication or rural—urban differences hn the relationship between education and poverty. ‘ The studies cited indicate a very definite relationship etween education and poverty. Caution must still be followed, owever, in applying the relationships found in cross- ection studies of the entire population to those families nd individuals now in poverty. Years of schooling attended 5 not the only factor which causes those with more education 0 earn higher incomes, and the studies themselves have eally presented only tentative estimates of returns to uman capital investment. In addition, most studies measured 61Tweeten, The Role of Education..., p. 21. average invest] 0: to sew antipo Ribich estima progra owner school l0b re all by cost : PrOgr effor incom in re iWes is u] inVe: dOes 00mm Cure 46 average, rather than marginal, returns to this form of investment. One attempt has been made to evaluate the returns to several forms of educational investment designed as antipoverty programs. In the study cited previously, Ribich compared benefit-cost ratios (which could be estimated on the basis of very limited data and few programs) of programs aimed at dropout prevention and compensatory education at the secondary level, pre- school training, increased per—pupil expenditure, and job retraining of unemployed workers. It was found that all but the job retraining programs had very low benefit— cost ratios, usually less than unity. The retraining programs fared much better, apparently because their (efforts were focused on individuals known to have low incomes.62 So for marginal contributions of education in relieving poverty conditions, the typeoof educational investment made appears to be of significant importance. . Conclusion Since the major emphasis of job—retraining programs '8 upon raising worker productivity, the human capital 'nvestment extension of marginal productivity labor theory oes appear to present valid conclusions and policy re- ommendations in relation to the causes of and potential ures for conditions of rural poverty. 62Ribich, Education ... , p. 97. Tc margins an inc: complet Bluest< States emplOy in par While famili Percen preser l retraj tithe: itsel But t 0ther emPlo Indu: ' 2 47 To the extent that factors other than low worker marginal productivity cause rural poverty conditions, an increase in worker productivity will not, by itself, completely solve the rural poverty problem. For example, Bluestone noted that in 1963, over 25 percent of United States households in poverty were headed by a fully employed individual.63 'This poverty resulted, at least in part, from their being employed in low—wage industries. While Thurow noted that the incidence of poverty among families with a fully employed head dropped from 12.2 percent in 1956 to 6.9 percent in 1963, this still re— presented eight and one-half million families.64 Raising the productivity of these workers through retraining would have some effect on their income levels, either through upward mobility in the low—wage industry itself or through movements to higher paying industries. but the low—wage industries would presumably still employ bther workers at poverty-level wages in the absence of full Fmployment. ‘ 63Barry Bluestone, "Lower-Income Workers and Marginal industries," in Louis A. Ferman, et. al. (eds.) Poverty ln America (Ann Arbor: University of Michigan Press, 1968), ’p. 273-302. 64 Thurow, "The Causes . . ., pp. 77—78. ,5, A. Fami The: household emphasis RLS stud: dwelling more that Separate Jus S0 a hou “0m to both 1% (Fl of two ( 1nirriagg stated, chil(ire: living , SurveYS A180 in is One others \ 6: PP- Al- IV.ECONOMIC AND SOCIAL CHARACTERISTICS OF THE SURVEY POPULATION A. Family Composition65 There are several alternative ways of organizing a household for statistical studies depending on the primary emphasis of the study. The basic interview unit in the RLS study was the household unit (HU), not just an occupied dwelling itself, for an occupied dwelling may contain more than one household depending on whether they live teparately or together. 1 Just as a dwelling can contain one or more households, so a household can contain one or more consumer units. a consumer unit (CU) was the general term used to refer :0 both families and unrelated individuals. A family Eli (FU), according to the Census definition, consists lf two or more persons living together and related by blood, arriage or adoption. In this study, unless otherwise tated, the family unit also included any unrelated hildren under sixteen years of age and any foster children iving with the family, even though the current population lrveys treat a foster child as an unrelated individual. .50 in this study, an unrelated individual (UI) or unit one over fifteen years of age who lives alone or with hers unrelated to him by blood, marriage, or adoption. 65This section draws heavily from USDA, Economic :earch Service, Rural Life Survey: Editing Manual, Al —A3 I F0 basic u the gen units a used ir (H-WMUl taxpayt study, to tho This a heads husban i.e., Pament requii had u but t were decis Who ( femaj rely r9th: sPou pQVe unit 49 For many purchasing and taxing transactions the basic unit is the marital unit. Marital uni; (MU) is the general term applied to both husband—wife marital units and to other marital units, and it was extensively used in RLS tabulations. The husband—wife marital unit (H—WMU) is the husband—wife couple of Census or the joint taxpayer of the Internal Revenue Service. In the RLS study, unless otherwise stated, the H-WMU was restricted to thOSe couples living together at the date of the survey. This agreed with Census in that it included husband-wife heads of sub—families (a sub—family is a family group, husband-wife-child or parent-child, living with relatives, i.e., couples, parents with children, or children with parents) but differed from IRS both as to date and to the requirement that they be in the same household. A H-WMU mad the combined income or assets of both husband and wife nut the classifying personal characteristics of the unit were those of the husband unless otherwise stated. The other marital units (OMU) were the one~person ecision making and income units. It comprised all those ho could not be classified as H-WMU. It included Census smale and other male heads, other relatives, and un- elated individuals. It included IRS taxpayers filing eturns as singles, heads of households, or surviving louses. This study of needs and potentials for escape from verty through retraining focused. alternately on consumer its and on their individual members. The basic unit for determ: was th: determ for es charac the co potent their Detail was oi 0f age alysi: or 501 indiv house Bu wally holds the s PeIce cent samp “Gig "l". the :termining the poverty status of all RLS respondents is the consumer unit, so needs for escape from poverty were :termined by consumer unit poverty status. Potentials >r escape from poverty, however, were determined by the Laracteristics and attributes of individual members of re consumer units. The study determines the retraining itentials of individuals and then estimates the effect :eir retraining will have on raising consumer unit income. ,tailed information relative to labor force participation . is obtained for all consumer unit members above 15 years 3 age. Whenever "individuals" are the group under an- gsis, it must be noted that it is this particular group some sub-sample of it that is referred to, and not all dividuals (including children under 16) in the survey useholds. Farm and Non—Farm Households For purposes of sample design, households were origi— .1y identified as either farm operator or non—farm house- ,ds. ’ There were twice as many non-farm as farm households,sg sampling procedure then selected approximately a 0.6 cent sample of farm operator households and a 0.3 per— t sample of non—farm households. To make the entire >1e proportional, the non-farm households were given a rht of "2" and farm operator households were weighted This permitted generalizations from the sample to area as a whole on the same basis for both farm and non- households or for the sample as a whole if farming status C. De Ti consum expres receiv requir income Cor Pow posit if th ratio non-p staty r€er anal} Vari; Unit the of r refi :atus was-not distinguished. Delineation of Poverty Categories The basic poverty code used in this analysis was the >nsumer unit poverty status. This poverty status was :pressed in terms of the ratio of consumer unit income fceived to that unit's income required, with income :quired being determined by consumer unit "poverty level" ycome less a modification for home grown food. That is, Consumer Unit _Income Received Poverty Status _Income Required , Consumer Unit Income -Consumer Unit _ Consumer Unit Home Poverty Level Grown Food Modification The value of the poverty status ratio would be sitive whenever there was any income received and negative the consumer unit received negative income, with the tio 1.00 being the dividing line between poverty level and n-poverty level incomes. Throughout the Rural Life Survey study, the poverty atus ratios have been broken downiinto categories which Flect varying degrees of poverty. This breakdown permits .lysis of the influence of the numerous socio—economic iables on the degree of poverty suffered by consumer ts rather than just lumping together all units below "poverty line." The ratios used throughout this analysis retraining potentials, utilizing the degree of poverty (nement, were arbitrarily classified as follows: <0.80 - deep poverty 0.80 <1.00 — poverty 1.00 €1.20 - near poverty 1.20 nsumer Unit Home Grown Foodeodification= In making final adjustments in the level of Consumer lit income requirements, Orshansky assumed that farm amilies earned income in kind from food that was home :own. The farm poverty level was, therefore, somewhat fbitrarily set at 70 percent of the non—farm poverty level. 1 Instead of accepting and applying the arbitrary 30 per— int adjustment factor of Orshansky for income in kind of .ral farm families, the Rural Life Survey study attempted i adjust incomes as accurately as possible by the actual i l punt of food requirements that were met by home-grown food r both farm and non-farm consumer units. To do this, ch respondent was asked to specify the percentage of air total 1966 needs of fruits, vegetables, grains, milk, 3 meat and poultry products which were raised by household nbers or given to the household by employers, friends, Latives, or welfare agencies. This percentage of food requirements which was home— >wn was converted to a home-grown food component by tiplying it by the dollar food requirements for in— iduals by age and sex types and summing for all in- iduals in the Consumer Unit and, as in the case of arty—level described above, by adjusting for family size. The non-farm poverty level (the left hand term in the ninator of the CU poverty status equation) was then ced by the value of this food in kind to give the sumer unit income required." This measure of income Lred can be considered a "modified Orshansky" requirement, the 0r: units 1 grown I consum the ex presen povert The 01 concey limite adee ent. The out are all of 195 all is ave the re] iHCOI Twirl requ tran and .e Orshansky criteria being modified for all consumer tits by actual amounts of food requirements met by home :own food. It must be cautioned that this procedure for determining >nsumer unit poverty status presents only one measure of 1e extent of rural poverty in this subregion. This is not :esented here as the only or the "true" measure of rural >verty, since any poverty line is necessarily arbitrary. 1e Orshansky poverty measure is plagued with several anceptual shortcomings. Bonnen has discussed the major .mitations of the Orshansky measure, noting that . It defines poverty entirely in terms of a nutritionally adequate low—cost food budget. No other human need enters into the definition in any operational manner. The income numbers which are set out as poverty—line cutoffs are not meaningful income aggregates since they are simply arbitrary food—cost figures multiplied in all cases by an assumed food—budget—income “multiplier" of three. This multiplier of three is derived from a 1955 food—budget study which found that one—third of all consumer income was spent on food... But not only is the multiplier statistic a decade old; it is also an average for the entire range of income rather than for the lower end of the distribution, which would be the relevant statistic. In addition, the reduction made in required farm tomes because some percentage of farm family food re— rements are home-grown also assumes that farm families uire the same percentage less than non-farm families of isportation, housing, medical care, and all other goods service. In so doing, the Orshansky measure substantially 67Bonnen,"Rural Poverty...," pp.462—463. reduces t the power The these lim measure i doing it there is E. The 1. By mining c Survey p units an With a F The Consumer am°n9 fa consume] 0f POVeJ perCent less the attaChe. with an rate Wa units. sintile reduces the number of farm families that would be found in the poverty catagory by other measures. The poverty measure used in this study contains all of these limitations. It does improve upon the Orshansky measure by using actual amounts of homegrown food, but in so doing it also extends whatever understatement of poverty there is to the non—farm households as well. E. The Rural Life Survey Population 1. Extent of Poverty—--Standard Criteria ‘ By applying the procedures described above for deter- nining consumer unit poverty status to the Rural Life Survey population, the following percentages of consumer units and individuals were found to be in poverty, that is, with a poverty status ratio below 1.00. The date in Table 3 indicate that sixteen percent of all :onsumer units were in poverty in 1966, with the percentage mong farm units only slightly higher than for nonafarm onsumer units. However, among family units only, the rate f poverty was over fifteen percent for farm and only ten arcent for non—farm consumer units for a total rate of ass than twelve percent. The rate of poverty among un— .tached individuals68 was very high, as shown in Table 3 th an overall rate of nearly fifty percent; the poverty te was slightly higher for non—farm than it was for farm its. Since unattached individuals represented only about 68Unattached individuals are persons living alone With persons to whom they are not related; they are gle member consumer units. twenty Mum] raise 1 than tn for all P. than b econom and no econom person SPecia may be effect emplo} used : from ; Rural Physi 57 twenty percent of the total survey population, the effect of nearly a fifty percent poverty rate among them was to y raise the total poverty rate for consumer units from less than twelve percent for family units to sixteen percent for all consumer units. Perhaps a more meaningful breakdown of consumer units than by farm-non-farm status is the breakdown by potential economic viability. All consumer unit heads under age 65 and not physically disabled were classified as potentially economically viable. This group no doubt includes some >ersons not retrainable and not potentially viable without :pecialized programs. As stated in Chapter I, individuals ‘ my be afflicted with non—physical disabilities which ffectively prevent them from being retrained and gainfully . mployed. .The measure of potential economic viability sed in this study, and the retraining potentials derived rom it, will tend to overstate the actual potentials of the ural Life Survey poor. Unfortunately, no measure of non- hysical disabilities was available in the survey data. The date in Table 4 are arranged by age, health, and amily status. It is apparent that poverty was much more revalent among the aged and the disabled, as would be Hence we» msv me~ ee-ma as-mm mmv museum cummaoz fl canmfl> hfipamm echemmnm _ smmammmwwe iii , suamomnome. owumm wEooaH 004 BoHom muss: Hoanwnoo zo>psm oHHq Amman \o only fort survey cc and the t years of consuumer of the 4' Individu The in the R these in in Table Head 1y forty-two percent of the poor.consumer units in the rvey could be considered potentially economically viable, ed the heads of fifty-six percent of those were over 44 ars of age. The number of potentially viable poor nsumer units represented less than seven percent (6.7%) the 4766 total consumer units in the survey. dividuals There were 10,533 individuals over age 15 included the Rural Life Survey. The percentage distribution of ese individuals by consumer unit status is presented Table 5. No separate breakdown of persons in poverty Table 5. Number and Percent Distribution of Survey Population by Consumer Unit Status Total 1967 Rural Life Survey IndividuaIS‘ Consumer Unit Status Number Percent Heads 4218 40.0 Wives of heads 3885 36.9 Unattached individuals 548 5.2 Other relatives 1882 17.9 Total 10533 100.0 : these four groupings is given, because the poverty rate : wives and other relatives will very closely approximate rate for family consumer units discussed above. The povertj cussed L Table survey the si This 6 consum relatf 00118111 appiie are The seven The 1 Power Pover Peres Over the 1 rate Was. QrOu POVe the 61 arty rate among unattached individuals was also dis— sed above. Looking at all individuals together, the data in le 6 indicate that just under thirteen percent of the fey individuals were in poverty in 1966, compared to sixteen percent for consumer units mentioned above. 5 difference between poverty rates for individuals and sumer units lies in the fact that all wives and other atives have the same poverty rate as the head of their sumer unit. Therefore, the family rate of 11.8 percent .ied to the ninety—five percent of the individuals who members of family units offsets the effect of the forty- :n percent poverty rate among the unattached individuals. large majority of the survey population, and of its :rty level individuals, are members of family units. As was the case for family consumer units, the farm rty rate for individuals of fifteen and one—half ent exceeded that for non—farm persons, which was just eleven percent. Among potentially viable individuals, otal poverty rate was just over eight percent, with the among farm individuals at twelve percent, while it nly half that for non—farm individuals. In the age over 64 years, one-third of all persons were in ty regardless of their place of residence. Looking at the percent distribution of poor individuals, ate in Table 6 indicate that fifty—two percent of all MCOHUDDHHUMHQ UGQUHmvnH USN «WDUNUW QUCUUHWUN TEN SUHNWEITWQ N0 UWWUEUUHTN NS KUHQ>QN CH WHWDUH>HUCH NW>H3W am GHQ—WU... 62 e/Jaeiiii N. s . iii m ow o.ooH o.m . . N m m NH k.wN e.em N.NH coupes: mom Hmuou mo coausn .eeeeeeeeeeeeeeeeeeen -enumes a [III-Iii m.o¢ H. . «m o ooh N.NH o.“ e.m~ e.HN N.HN o.e sauce o.ke . . O me o ooh H.mH o.w m.Hm H.©H e.Hm m.k auaw-aoz H.Nm m.~e o.ooH H.w . . . . . eeeeeeeeeeei e m 0 am m om e om a as some Moon mo floausn -ueumao N eeeeeeeeeeweeeeewe m.~m . . . m m m NH o oe w.km a.m~ w.e w.“ o.e swoop 4 mm m.e e.HH s.Hs m.om 9.0m N.k m.e k.e sneu-:oz H om H.NH m.mH 0.9m m.m~ e.wN s.mH m.HH e.oH some max enemas sauce mom mev mom ee-ms se-mN va suede oumwfioz H oHan> lemom tofibmwfio _ iguana aunmee-aws mQOHusanumHQ uswuuwm can .msumum ooawcwmom tam :uHmwmucw< mo kuno>0h EH wqmnca>aon4 xv>u=o HDTT oeumm oEooaH OOH Boaom mHmDpH>Han No>usm owed Hmpcm mom a poor i forty- units twenty lative 1 diffs: catege aged . peroe where total sixte Sing; 3.9! (not POVe inco Stud thro p°Pu 63 Jor individuals are potentially viable, as compared to arty—two percent potentially viable among poor consumer rits (Table 4). The difference lies largely with the under venty-five group which consists primarily of other re- etives living with poverty level families. In the overall survey population there were very slight .fferences in the percent distribution by ageihealth Ltegories. The only large differences were among the viable, red 25—44, where the percent of non-farm persons was ten ercent more than farm persons, and those aged 45—64, lere the difference was reversed. Eighty percent of the wtal survey population were potentially viable, and only xteen percent were over 64 years of age. 2. Extent of Poverty for Retraining Purposes nce no apparent differences exist between the family with ”99 income ratio (in poverty) and one with a ratio of 1.01 ot in poverty) and since being in a state of "near erty" still leaves a lot to be desired, all persons with ome.ratios below 1.50 were included among the "poor" in this dy of needs and potentials for escape from poverty ough retraining. Thus, one—fourth of the survey ulation (26.3%) was included in the group below the 0 income ratio as shown in Table 7. Of this group, rly two—thirds (62.6%) were considered to be potentially nomically viable, that is, under 65 years of age and lthy. Those below the 1.50 income ratio who were viable resented one-fifth of all potentially viable respondents g, l n :uAUHNkrOnH 1H0 UWHMQQ: >fl~ SOs-HUNHu-CCAF eh @HDNH xviaviuuu uni 64 [II II," iiiii w.N¢ @.Hm . o ooh N ma e.m e.mN N.mN N.mN N.N sonuseepumaa in!!! 0 cm o.HN TillaflamfllllTlllJlllllllllllll scoouofi m Ne e.am H.Nm e.ON N.NN a.wn 0m.av eltlllwillllllllefllllllllllnIIIIlllllIlllllllllllllllllllllllle HmuoH muHCD HoEJmcoo on [Twill . i No 0 ooH N a e.m m.NN m.HN m.wN N.NH consonan6maa mewllfiflMfllllllTllllfllllll . uuoouofi 0N m 0N N no e.me e.wa N.aH e.HN e.wn em.nv flflflhlll.m.w Illlldllllillll. HmuOH o w a m w.o~ m.oH m.o m.m H.w m¢.HION.H mesmtumm m.m . . . . a m a e o NH o m N.w e.m N.q w.m eH.H-oo.H mupo>om Hmoz w. . . . . m m N N g m NH 5 HH o.m m.N m.N m.N mm.eow. muuo>om m NN m.m ©.w m.mN o.oH m.ON m.o o.m ¢.¢ ow.v kuuw>om moon mom e231, H305 3M 8v SM 3v? meVmN va 0331 meMHMB . — GHQNH> wEoofiH easemmna _ samwmmmele aunmee-mw< maesea>aesH oeumm oEoocH omA BoHom mamDpN>NtGH mo>u5m omwa Hausa moma maceusnfluumfla uncoucm was .wcumum nuamomeow< can :muuo>om mo common: >n coaumasaom ko>u3m mo owmufioouom ii .i Tiii||||||l and one Te indivi Only t as the very c major downs by far as, at very e was e: to no age 9 Among of ty Was j Perce item Ihas divi iron the Were 65 d one-sixth (16.5%) of the entire survey population. Table 7 also presents the percentage breakdown of those dividuals below the 1.50 income ratio by "degree of poverty." 1y total percentages are given for consumer units inasmuch the gonsumer Unit percentages by degree of poverty ry closely approximated those for individuals. The only jor difference was in the under 25 age group. No break— wns of the under 1.50 income ratio population were made farm—non—farm status or by type of consumer unit status , again, this group was distributed/on these dimensions ry similarly to the group under 1.00 income ratio which 5 extensively described above. Among the potentially viable individuals, it is interesting note that approximately one-fifth of each of the three 3 groups is included in the group below 1.50 income ratio. ang those individuals over age 64, more than one—half the total population (individuals and consumer units) 1 t included in this measure of poverty, and over forty cent of all persons under age 65 who were disabled had ome ratios below 1.50. 3. Selected Characteristics of the Retraining Study Poor smuch as this retraining study focused upon all in- iduals with income ratios below 1.50, it will be this up of "poor" which is referred to in the remainder of study unless otherwise specified. It was noted above that sixty—three percent of the poor e considered to be potentially economically viable, that is, un potent other percer disabi throue Educa- to me mi Incom .80 1.00 1.20 iota Tote Toya Male Fems Yea] 11; 66 , under age 65 and not seriously disabled. They would be tential candidates for job retraining programs. The ier thirty-seven percent of the poor, the thirty-two scent over age 64 and the five percent under 65 but sabled, could only expect to be lifted from poverty rough some form of social welfare program. ication Concerning the education levels of the poor in relation median years of schooling completed for the general Table 8 Percentage Distribution of the Survey Poor by Relationship to Educational Medians of the Regional Population, by Degree of Poverty and Total by Sex. 7 Rural Life Survey Individuals with Income Ratios Below 1.50 Relationship to Educational Median ome ‘ > i0 Specia1* < — , < .80 4.5 39.8 47.5 8.2 < 1.00 3.4 45.4 46.3 5.0 < 1.20 3.3 41.3 47.7 7.6 <’ 1.50 3.2 39.0 49.2 8.6 el< 1.50 3.7 40.7 47.9 7.7 11 Non-poor 1.0 28.4 56.8 13.8 Ll Sample 1.7 31.6 54.5 12.2 s < 1.50 4.3 45.8 43.4 6-5 *Includes those with no schooling or with less than five S due to handicaps. By degree of poverty, chi—square— 89 with 9 d.f. (sig. only at .30 levelk popula percer the me educae signi media the p howev surve eight equal media 8. 1 diffe by a male medi Post POO: trai Pope ten str den dif Des 67 opulation in this region, it was found that forty—one ercent were below the median, forty-eight percent were at he median, and only eight percent exceeded the median ducational level for their age group. There were no ignificant differences in the relationship to educational sdians when the poor were subdivided by degree of poverty; 1e percentages were fairly uniform across all categories. owever, the poor did differ very significantly from the lrvey non—poor in this respect. For the non-poor twenty- Lght percent were below the median, fifty—seven percent yualled the median, and fourteen percent exceeded the edian. The percentage relationships are presented in Table Looking at males and females separately, no significant .fferences appeared for either sex group when subdividing ' degree of poverty. However, ten percent more of the .1es, as compared to females, were classified below the dian, with nearly fifty percent of the males so classified. st High School Training It was found that less than four percent of the survey Or had received any type of post high school technical aining. While only eight percent of the total survey pulation had received any training of this tYPer the dif' rences between the poor and and the non—poor were still riking as the percentage for the non-Poor was more than uble that for the poor. There were also highly Significant fferences in the relationship between degree 0f poverty and 5t high school technical training, with nearly seven percen' hardsh compar categc preser In Incom Ratio .80 1.00 1.20 Tota Tota Surw Occ 1ev Whl 68 ercent of those in the highest poverty category, i.e., ership (1.20-1.49), having received some training as empared to less than three percent in the other three .tegories. The data on post high school training are esented in Table 9. Table 9. Relationship Between Degree of Poverty and Post High School Technical Training for Survey Poor; Percent Distributions 67 Rural Life Survey Individuals Below 1.50 Income Ratio Post High School Technical Training come. tiO None Some < .80 97.8 2.2 80 - .99 98.4 1.6 00 ‘ 1.19 97.3 2.7 20 ’ 1.49 93.2 6.8 tal < 1.50 96.3 3.7 tal non—poor 90.7 9.3 rvey Total 92.2 ”7.8 Among poor, chi square = 37.095 (with 3 d.f., sig. .001 level). :upational Skill Index A rough measure of respOndents' occupational skill 7618 was developed based upon the type of occupation in -Ch they worked during 1966. Those who worked in more than 01 skill ' were c skills S highes loves degre Pests degre skil] all e Ploye Pero for Tron Skil Perc in i come 69 an one type of occupation were credited with the highest ill level in which they worked. The various occupations re classified on a continuum from lowest to highest illed as follows: Skill Level Occupations Code 1 ghest 4 Professional and technical employees and self—employed; non—farm managers and officials 3 Craftsmen 2 Clerical; sales; self employed farmers; operatives; non—farm non—technical self employed vest 1 Service workers; laborers; armed forces The percentage distributions of the survey poor, by free of poverty, are given in Table 10. As would be ex— :ted, there was a highly significant relationship between free of poverty and skill level, even with such a rough .11 level measure. Also, the very large majority of . the poor, irrespective of degree of poverty, were em- eyed in the two lowest skill level occupation groups. The ‘centages in the two lowest groups were about ninety—percent the total poor and for all but the highest poverty level up, which had eighty—five percent in the two lowest 11 categories. This compared to only seventy-four cent in these two categories for the non-poor individuals the survey. The larger percentage of the under .80 in— e ratio group in skill level category two was due to the t that all respondents with negative income ratios were membe was t as WC a up Incc @ 70 embers of farm families. Another interesting point revealed by the data in Table 10 as that degree of poverty was strongly related to employment, 3 would also be expected. Table 10. Relationship Between Degree of Poverty and an Occupational Skill Index for Survey Poor Who Held Jobs in 1966 and Totals by Sex; Percent Distributions. 67 Rural Life Survey Individuals Below 1.50 Income Ratio Continuum of Occupational Skill Levels lowest +————->highest ycome Percent of Total £597 1 2 3 4 With 1966 Job < .80 22.6 70.3 4.2 3.0 37.2 80 - .99 34.5 53.6 11.3 0.6 38.1 00 — 1.19 28.1 60.3 6.7 4.9 43.8 20 - 1.49 27.7 57.4 9.9 5.1 54.4 tal (1.50 27.3 60.9 7.9 3.8 i 44.2 ‘1 .es (1.50 20.0 66.2 10.6 3.2 ‘ 69.4 iv talesgram. One approach is to compare the present values of a future income stream which can reasonably be expected )m the worker's present occupation with the income ream he can reasonably expect from the occupation for .ch he will be retrained. Two ways of making this nparison are discussed in this chapter. The second approach is to conduct a benefit—cost ilysis based on the actual experiences of a sample of verty—level persons who have undergone job retraining. ‘this way the net benefits to workers from retraining 11d be estimated, and the addition of these net personal 1efits to original income levels of poverty-level workers 11d indicate the potentials for their escape from poverty 1ough job retraining. The results of several benefit- 1 t studies of retraining programs and their implications raising poverty level incomes are also examined in this 1 Expected Future Income Streams It would be relatively simple to determine if it would 1 1 1a worker to undergo job retaining if data were avail— g 0n the present value of future income streams he could 1 sonably expect to earn OVer his lifetime from various 1 7, alternat alternat race and the pres The rough a} five “5] broad 0- was rep Venkare values four a] move w} made f1 tribut Time s unempl 0r est eS’Cima 78 alternative occupations. Such precise data, covering many alternative occupations for workers by age, education, race and other pertinent variables are not available at the present time. There are two studies available which present "very rough approximations" of the desired data, one limited to five "specific" occupations and the other covering ten broad occupational classifications. One of these studies was reported in the Ph.D. thesis of Venkareddy69 and in Venkareddy and Johnson.70 This study estimated present values of future income streams of males in farming and in four alternative occupations into which farmers most frequently move when taking non—farm employment. These estimates were made for the brpader purpose of projecting the age dis— tribution of farm operators in the United States to 1970. Time series data on (average) wages, interest rates, and unemployment rates from 1917 to 1962 (from published sources pr estimated from regression equations) were utilized in estimating these present values. 69Chennareddy Venkareddy, "Present Values of Expected Future Income Streams and Their Relevance to Mobility of Farm Workers to the Non- -farm Sector in the United States, 1917— 62" (unpublished Ph. D. dissertation, Michigan state University, 1965» 70Chennareddy Venkareddy and Glenn L. Johnson, "Pro— jection of Age Distributions of Farm Operators in the United States Based Upon Estimates of the Present Value of Income, " American Journal of Agricultural Economics, Vol.50, No. 3 (August, 1968» pp. 606-620. incom study only large mate: as " 40y beli it a inci mob the This approach to estimating present values of future 1ncome streams has many limitations to its use in the present 1tudy. Venkareddy and Johnson estimated the present values only for twenty-five and forty-five year old workers, .argely because of the effort involved in making the esti- 1ates. The educational level of workers was not included as , variable in the study. The two ages were arbitrarily selected 1 15 "representative" of persons in the two age groups 15 to 10 years and 40 years and older. The younger category was 1 1elieved to include most persons who enter farming to make 1.t a lifetime work, whereas the older group was believed to 1nclude "many of those who leave farming." The farm labor 1obility studies cited in chapter three would indicate that he latter of these conditions has not held in recent years. In addition, where previous studies indicated that farm orkers most often move to employment in (1) building trades, 2) manufacturing, (3) service industries and (4) trade, he authors selected for consideration the sub—occupations of l) "helpers and laborers" in building trades, (2) "laundries" a service industries, and (3) "retail trade" under trade, 5 well as "hired agricultural labor" instead of farm >erators simply because wage rates were available for rese groups, with some exceptions, for the period 1917—1962. .e authors made no attempt to justify these sub—occupations being representative of the type of work actually en— ged in by occupationally mobile farm workers. appea earn rath1 line rate farm of t betw difl dif 8110 fut For Sol dm r111 80 Projections of wage rates for these occupations also appear very shaky. First, it was assumed that all workers earn income until their death, approximately at age 72, rather than until the presently accepted retirement age of 65. Secondly, estimates to the year 2007 were made as a linear function of current year and past year annual income rates. But, if the hoped for adjustments in the size of the farm labor force take place there should also be a narrowing of the difference, or a change in the linear relationship, between agricultural wage rates and those in other sectors. Finally, the assumptions that unemployment rates do not differ for the two age groups considered, and that "the differences between the capacities, skills and training of a 25 year old and a 45 year old worker are not significant enough to effect any difference in their expectations of future income streams Up to nl years"71 (n1 ranges from 25 to 27 remaining years of life) appear to be quite unrealistic. For example, the 1960 census data on median years of schooling Completed for these age groups raises strong doubts about the latter assumption, particularly for both rural farm and rural non—farm residents (Table 13). These data show a difference of three years in median years of schooling completed between the two age goups which include I! 71 Venkareddy, "Present Values..., p. 21. the tw valu cape earr 9st: acc1 “Pa Va] o‘er 81 the two ages used in this study. Table 13. Median Years of Schooling Completed in United States, Selected Age Groups Residence Age Group 25—29 45—49 U. S. Total 12.3 10.3 U. S. Rural Farm 11.5 8.7 U. S. Rural Non—farm 12.0 p 9.0 Source: U.S. Bureau of the Census, U.S. Census of P0 ulation: 1960. Detailed Characteristics. U. S. Summary. Final Report PC (1) — ID (Washington: U.S. Government Printing Office, 1963), Table 10. The second study which estimates the present value of expected future income,streams was done by Miller and Hornseth.72 The development of this set of estimates was motivated by the concern of the legal profession for placing a pecuniary yvalue on a human life, or on a person's impaired earning 1capacity. Using cross—sectional data on average (mean) 1959 earnings contained in the 1960 Census, the present value of estimated lifetime earnings was calculated for males according to age, education level, color, and major occ— upational group. The ten major occupational groups listed in the 1960 Census occupational index were used, and lifetime 72 Herman P. Miller and Richard A Hornseth, Present Values of Estimated Lifetime Earnings, U.S. Bureau of the Census Technical Paper No. 16 (Washington: U.S. Government Printing Office, 1967). earni rates est: ear: Wer Eve has _ -_ - (——~._ .—- i“, fix 1.? welsh-Lfiflme $1. > 1 82 earnings were estimated for various combinations of discount rates and annual productivity increase rates. Miller and Hornseth correctly projected the working life of a man to his retirement age, rather than to his expected mortality. Their use of cross—sectional data again assumed that present (1959) relative income positions would remain the same in the future. However, such is likely to be the actual expectations of individuals pre- sently making occupational decisions. The use of this assumption inherent in cross—sectional data appears justified for the purpose of the Miller and Hornseth projections, while its acceptance by Venkareddy and Johnson in projecting the decline in farm operator numbers cannot so easily be justified. Miller and Hornseth noted that estimates made from life-history data which trace a man's earnings through his working life might, if such data were available, provide better estimates of present values. Even then,future projections would only be made on the asis of past trends. The Miller and Hornseth study avoided the thorny 1'3 1 1 problem of selecting specific discount rates and rates ‘ 1 D f annual productivity increase, leaving such decisions to fhe user of the data, but compounding the problem of 1 fietermining what the most probable present values might be. 1 The major drawback in the use of their estimates, yowever, is the lack of disaggregation in the occupational 1ategories. For example, the category "professional techn occur the c opera iflCll and : diff cate sent prok rate of- Sell the can hit per du< Wm Mi ra 83 technical, and kindred workers" includes such diverse occupations as physicianSeand—surgeOns and clergymen; the category “farmers and farm managers" includes owner operators and share croppers; the category "sales worker" includes the diverse occupations of real estate brokers and newsboys. Certainly the expected future earnings differ significantly for the occupations noted in each category, yet all are averaged together in this study. Comparisons between estimated present values pre— sented by these studies are difficult to make, and are probably quite meaningless. One would need to know the rate of interest used by Venkareddy and Johnson before one of the four rates used by Miller and Hornseth could be selected in such a comparison. The maximum rate considered by the latter was 5.0 percent, while the rates used in the former varied from 6.95 percent in 1921 to 4.48 per- :ent in 1947. For 1959 (the only year in which any comparison might be made) the Venkareddy and Johnson rate was 5.41 bercent.73 Also for comparisons, the annual rate of pro— uctivity increase implicit in Venkareddy and Johnson ould have to be known before the proper rate used by 1 1iller and Hornseth could be selected. ‘ Nevertheless, if one heroically assumes a discount 'ate of 5.0 percent and a productivity growth rate of 3.0 ercent to be "reasonably" close to those used by Venkareddy L___i l 1 1 73Venkareddy, "Present Values...," p. 48. and L comp: but whic the sect estf leV1 fac 399 oat pre woi dis so Wn 84 and Johnson, and that the occupational categories are comparable, it is apparent that the two methods arrive at widely divergent estimates of present values for all but farm laborers (Table 14). One is left wondering which type of study, time series or cross—sectional, gives the "better" estimates of present values. The cross- sectional study of Miller and Hornseth was able to make estimates for workers Of all ages with differing educational levels, and it avoided many of the methodological problems faced in generating time series data. The lack of dis- aggregation of income data into more realistic occupational categories limits the usefulness of this study in determining present values of expected future income streams for workers retrained in specific occupations, but with sufficient disaggregation of Census data this method could be rather economically utilized to make the desired projections. Whereas the Venkareddy and Johnson study did make estimates of present values for relatively specific occupational 1categories, it was plagued by lack of data and numerous methodological problems. _It also appears, from Table 14, Fhat this approach tends to greatly underestimate these present values. A final weakness of both approaches for the purpose Of this study,along with the occupational aggregation problem, is that both used average incomes of all workers in each occupational category, Retrained poverty-level workers would most likely earn only marginal incomes upon 85 Table 14. Expected Lifetime Earnings Compared — 1959 (in thousands of dollars) Categories: (Miller and Hornseth) (Venkareddy and Johnson)* Education age age Levels 26 46 25 45 Operatives & Kindred Workers (incl. laundry & dry cleaning) Laundries all ed. levels 116 64 55 42 8 years 114 66 12 years 130 72 16 years 145 84 (braftsmen, Foremen & Kindred Construction (incl. carpenters, electricians painters, machinists, toolmakers mechanics) all ed levels 138 75 142 102 8 years 128 74 12 years 150 84 Manufacturing 16 years 202 118 142 84 Farmers & Farm mgrs. (incl. tenants & share croppers) Farming all ed levels 87 44 57 43 8 years 79 44 12 years 104 57 16 years 165 90 Farm Laborers & Foremen all ed. levels 49, 26 8 years 56 32 12 years 81 46 16 years 89 48 * All education levels are implied. beginnii earn be from th explain study h level n such as employs structo contin1 86 beginning employment and would probably always tend to earn below the occupational average. This would result from the fact that there are several variables which explain the incidence of poverty, as pointed out in the study by Thurow.74 Even though training raised a poverty- 1evel worker's skill level, other socio—economic variables such as his race, formal education, percentage of time employed, his place of residence and its industrial structure, as well as motivational variables, would continue to exert downward pressures on his income level. C. Retraining Benefit—Cost Analyses When considering studies of the effectiveness of job retraining programs, it is essential to know what the basic objectives of these programs have been. Evaluations which are made for the benefit of public-policy makers can be expected to focus on the issues and objectives 1 . which are considered most important by these policy makers. 1 The Federal retraining legislation with the most clearly defined set of objectives is the Manpower Develop— ment and Training Act (MDTA) of 1962, the nation's first large—scale attempt at retraining. The MDTA objectives are (l) to increase the nation’s output; (2) to reduce the aggregate level of unemployment; (3) to reduce the govern- nental costs of unemployment; and (4) to reduce the burdens of unemployment for specific groups of the unemployed. 74Thurow, "The Causes of Poverty." The s coulr gram beca othe thos most were atts prow 521 to COS sin Spe ans Ie1 PO‘ Fe Ce 1e statistics on retrainees in the early years of MDTA :uld not be used to determine with any accuracy the pro- ram's effectiveness in reaching objectives one and two ecause of general overall growth in the economy due to ther factors. It appeared that the program had some success n reducing government costs of unemployment, at least for hose workers retrained, but it was quite apparent that the ast disadvantaged members of the ranks of the unemployed are not being retrained. Amendments to MDTA in 1963 and 1965 ttempted to deal with this failure of the program by roviding more adequate retraining allowances and, in effect, y lowering entrance requirements by providing for up to 2 weeks of basic education prior to actual skill retraining. No studies of significant scope, with results generalizeable 3 the total population, have been made of the benefits and )sts of retraining specifically rural workers. In addition, ince reduction of the number of families in poverty was not becifically one of the goals of MDTA, none of the program ralysis studies deal specifically with the potential for straining to enable workers' families to rise above averty levels. In keeping with the objectives of the MDTA, most 1deral assessments of the program focus upon (1) the per- ntage of MDTA course completers who are placed in employment; ) the increase in per hour earnings for retrainees ter training compared to pretraining earnings, and (3) the rcentage of time employed after training as compared to pretraini 1969 rep: Welfare placemen‘ the manp cessful. report 0 on perso employed rePorted report 5 level 01 cent sh< Within 1 a decli: two-sir trainin earning \ 75 E . % Repert MDTA (y 196911! 71 "The II u I w 88 pretraining employment. This focus is summed up in the 1969 report of the Secretary of Health, Education and Welfare (HEW) which states, "according to all available evidence—— placement in jobs, increase in earnings, length of employment —— the manpower institutional training programs have been suc— cessful."75 The earnings evidence appears in a December 1968 report of the U.S. Department of Labor.76 Based primarily on persons completing retraining in 1965 and 1966 who were employed on the latest reporting date, 70 percent of whom reported pretraining and posttraining earnings data, the report shows an increase of 20 percent in the average earnings level of employed graduates. Of those reporting, 60 per— cent showed advances in earnings, 25 percent moved laterally within the same earnings interval, and 15 percent experienced a decline in hourly earnings. In the latter group, nearly ytwo—fifths still earned at least $1.75 per hour after re— 1training. The majority of those reporting pretraining earnings below the $1.25 per hour (1966)Federal minimum, 75U.S. Department of Health, Education and Welfare, Education and Training: A Chance to Advance, 7th Annual Report to the Congress on Training Activities Under the MDTA (Washington: U.S. Government Printing Office, April, 1969), p. 1, " 76U.S. Department of Labor, Manpower Administration, The Influence of MDTA Training on Earnings," Manpower Eval- uation Report No. 8 (Washington, December, 1968» pp. 7-11. nearly earning after 1 report T mobili were e sucoes clearl placer adequ ments Gradu the r haVe retr: 1969 that ing Were 89 nearly one—third of the total, did move upward in the earnings scale. Only one-eighth were found in this category after training. The following three tables summarize the report findings. The significant aspect of this report on earnings mobility of retrainees is that it reports on only those who were employed on the last reporting date. The evidence on the success of retrainees in finding employment is not completely clear, largely due to early inadequacies in gathering job placement data and later attempts to correct these in— 1 adequacies. Early reports tended only to record job place~ ments immediately upon course completion. Data for 1965 graduates indicated that 71 percent were employed, with the rate for males being 77 percent,77 and these figures have been widely quoted as indicators of the success of 1retraining. Better follow-up reporting was reflected in the 11969 HEW report, which stated that "recent studies‘indicate 1that 85 percent of those who completed institutional train- }ing obtained jobs (some time) after training and 75 percent 1were employed at the time of last contact."78 The rates are ____________________ 77 U. S. Department of Labor, Report of the Secretary ~0f Labor on Manpower Research and Training Under theififiTA_ %(Washington, 1966» p. I8. 78 U. S. Department of Health, Education and Welfare, 1Education ... , p. of ME Straigh' Tabl 90 Table 15. Pretraining and Posttraining Earnings of MDTA Institutional Graduates in 1965 and 1966 (Percentage distribution) Employed graduates Straight time average hourly earnings Before After training training Total: Percentl 100.0 100.0 $0.50 to $0.74 6.6 1.7 $0.75 to $1.14 20.9 7.8 $1.15 to $1.24 4.3 2.1 $1.25 to $1.49 23.3 21.4 $1.50 to $1.74 14.3 17.4 $1.75 to $1.99 8.1 13.2 $2.00 to $2.49 12.0 20.4 $2.50 to $.299 6.3 10.6 $3.00 and over 4.1 5.5 Median earnings $1.44 $1.73 l . _ Percent distribution based on 79,836 employed Graduates reporting pretraining earnings and 95,542 reporting posttraining earnings. NOTE: Detail may not add to 100 percent due to rounding. Source: U.S. Department of Labor, Manpower Administration, The Influence of MDTA Training on Earnings," Manpower Egaigation Report No. 8. December, 1968. Washington, D.C. ae. _- Posttrair MDTA Ins‘ Straight average earnings training 91 Table 16. Posttraining Earnings Compared With Pretraining Earnings of MDTA Institutiénal Training Graduates in 1965 and 1966 Straight time average hourly earnings before chang es in earnin S Number of employed graduates experiencing training Total graduates No Change Increases Decreases reporting Tota1:Number 73,161 16,650 44,476 12,035 $0.50 to $0.74 4,807 488 4,319 ————— $0.75 to $1.14 15,400 2,370 12,780 250 $1.15 to $1.24 3,198 99 2,872 227 $1.25 to $1.49 17,117 4,399 11,620 1,098 $1.50 to $1.74 10,470 2,267 6,385 1,818 $1375 to $1.99 5,908 1,146 3,208 1,554 $2.00 to $2.49 8,763 3,195 2,576 2,992 $2.50 to $2.99 4,566 1,430 716 2,420 1 $3.00 and over 2,932 , 1,256 ---- 1,676 Source: U. S. Department of Labor, Manpower Administration, Evaluation (Table 2). Report N9: 8, December, 1968. 'The Influence of MDTA Training on Earnings,"Manpower Washington, D. C. WNEHEHG¢ Pr ( ((((((( UWGH m umSomcmz . :chmuw o ou ou Op On cu ou cu ou . ouowwb cam oo.m om.Nm oo.Nw mN.Hm om.Hw mN.Hm mH.Hm mk.ow om.ow mmaHauma HmuoH mauso: own . Illlllllllllllllll .cflcfimut chwm uuo>m oEHu M wwcwcumo mHuco: ommuc>m oEHu ucmwmuum uzwfimuum _ / even. bet The of earni since nc comparat certains to retr: the ove: wage le to unio answer retrain Tr early 1 trainix Control Ti Studie relati Proced 61961101 The Ss fewer 3901131 below then 1' the s: Pays even better for OJT Trainees. The major shortcoming of these Federal evaluations 3f earnings and employment gains of retrainees.is that, since no data is presented on the economic behavior of :omparable control groups of workers, it cannot be as- :ertained from them how many of the gains are attributable :o retraining and how many to "normal" factors such as :he overall increase in economic activity, Federal minimum rage legislation, and regularincreases in pay rates due .0 union—management contracts. Thus, alone they fail to .nswer questions concerning potential income gains from straining. The need for comparable control groups Was recognized arly by persons conducting benefit-cost analyses of re- raining programs, and most attempted to incorporate various ontrol groups in their evaluations. There have been basically two types of benefit—cost tudies completed to date. The first can be considered elatively large scale, based on some type of sampling rocedure, and requested specifically by one of the Federal gencies responsible for administering manpower programs. 1e second type has been of much smaller scale, focused on awer retrainees in relatively local regions, and often ponsored by non—governmental organizations. The discussion elow will focus first on two of the larger type studies, 1en attention will be turned to three works which evaluated 1e smaller studies, to attempt to determine if retraining 1Y5 for poverty-level workers. Befor is worthwl about the: A 001 administr 1968, tha benefit-c 0f the "n for the r impossibl by examir training availabl. EXCEPt a Bee sta lin ful act ex; int 0t} Sample 1 program \ 79 10 1W1? 9° 121. 81 94 Before examining the existing benefit—cost studies, it is worthwhile to note the comments of two research groups about the possibilities for conducting such studies. A committee appointed by Congress to examine the administrative practices of training programs noted, in .968, that one reason it did not undertake an economic >enefit-cost evaluation of the various programs was because of the "non—existence of benefit—cost evaluation models For the various programs.”9 The committee even found it .mpossible to determine actual impact on trainee incomes 1y examining a statistical sample of trainees' post— raining employment history because the "crude records" vailable precluded even the selection of such a sample xcept at undue cost. The committee, therefore, recommended: Because data on program performance and trainee status, characteristics, and accomplishments are limited or unavailable to the degree that meaning— ful evaluation of the benefits and costs of these activities appear precluded, a comprehensive and expeditious system of data collection must be put into being immediately. Other researchers for this committee planned to Dnduct a follow-up study based on a 5 percent random ample of trainees registered six months earlier in selected rograms. The proposed study was dropped due to the 79U. S. Senate, Subcommittee on Employment, Manpower, 1d Poverty of the Committee on Labor and Public Welfare, nployment and Training Legislation — 1968 Background iformation Supplement (90th Congress, 2nd Session) Washington: U. S. Government Printing Office, June, 1968), 80Ibid.,p. l6. “inadeqi records not cle unknown T1 be disc Corpor: purp051 to dev were m non-tr Nevert quoted devel1 about train benef betwe durat 95 "inadequacy of statistical data on trainees"; few program records were adequate for follow—up purposes, dropouts were not clearly identified, and post—training status was often unknown.81 The first of the large sample benefit-cost studies to be discussed was completed by the Planning Research Corporation for the Manpower Administration.82 The primary purpose of this study,in1view of the above criticism, was to develop and demonstrate methodology, and reservations were made about the quality of the data. Specifically, non—trainee control groups were not used for comparison. Nevertheless, both the U. S. Department of Labor and HEW quoted the findings of the study in support of manpower development and training programs.83 From a sample of about 2,000 institutional trainees and 650 on—the—job trainees who enrolled in training, estimates were made of benefit-cost ratios of training based on differences between pretraining and posttraining hourly earnings and on duration of employment after training. —————_______________ 81%. , p, 268. 82Planning and Research Corporation, Cost Effectiveness éggiysis of On the Job and Institutional Training Courses, A Report to the U. S. Department of Labor, Office of Manpower Policy, Evaluation and Research (Washington: By the Corp— oration, 1967). Presigzgt :ndDgzgg:gegfi fignLgsgi’Rgagigzggnf:pog:s§5rdg: ‘fiEITIEaEion and Training (Sashingtgn: U. S. Government I , Printing Office, January, 1969), p. 219; U. S. Department ‘Of Health, Education and Welfare, Education...,p.,63. Thi cost rat to soci1 the Fed on-the- only on enrolls were 2. trainii hourly traine gains pretra l evalua abili1 the U preju benef decre assig valus clea] This study found the average net Federal benefit- ost ratio, defined as the direct and indirect benefits 0 society (exclusive of increased taxes paid) compared to he Federal investment per trainee, to be 3.28 to l for n-the-job training and 1.78 to l for institutional training, nly one year after training. This was for all program nrollees; when only completers were included, the ratios ere 2.13 to l for OJT and 1.09 to l for institutional raining.84 It was found that OJT trainees received higher ourly wages and were unemployed less than institutional rainees, yet the latter experienced greater percentage ains in both earnings and employment because of poorer retraining performance. The value of such a study lies more in its comparative valuation of alternative training programs than in its bility to measure gains from retraining. Nevertheless, 1e U.S. Department of Labor concluded that, "even without rejudging the number of years for which the differential enefit would last, or whether it would tend to increase, acrease, or remain constant, or without arbitrarily SSigning a discount rate by which to calculate present alues of benefits, the desireability of both programs is Lear."85 The only large sample study which included a "comparable" k“ 84U.S. Department of Labor, Manpower Report...,p. 219. 851bid., p. 219. control r Opinion 1 the Offi This stu ability persons courses researcl be inte provide about t "contrc 1966, r Wherea: Contai: time p PIESen retrai 1 Other in W81 frOm 1 were sisni \ Insri Vol. 97 control group was conducted by Main of the National Opinion Research Center of the University of Chicago for the Office of Manpower Policy, Evaluatibn, Research.86 This study was based on interviews with a national prob— ability sample of about 1,200 trainees and 1,060 other persons who were unemployed about the same time the training courses started. Since it is impossible in this type of research to obtain a true control group, each trainee to be interviewed was selected first and then asked to provide the names of three acquaintances who were unemployed about the time his training course started. One of these "controls" was then interviewed. Interviews took place in 1966, more than a year after job training was completed. Whereas this procedure had to rely 0n recall data, and it contained certain methodological problems such as differing time periods during which work histories were compared, it presents the most thorough and objective analysis of 1 retraining programs to date. Main found that the secular growth in the economy and other, unknOWn, factors enabled controls to obtain increases in weekly wages which were not significantly different from trainees' increases when ten background variables Were controlled for, even though trainees did experience Significantly greater increases in family income (60 percent N“ .86Earl D. Main, "A Nationwide Evaluation of MDTA Institutional Job Training," Journal of Human Resources, V01- 3, (Spring, 1968), pp. 159—70. ‘1 p -‘._’2 —.:‘ L; 1":3Y:’§1§»_’—7¢;H7;.9——-k._,_(1_ kl vi , still e income, trainin betweer (the di in the not he to obt in fan b cost 1 net ps the o] if it under Main to in COSt Born: West ment 1196 Beer E e Bens Pro‘ 1171 98 still earned less than $80 a week). The increases in family income, estimated at about $10 per week, resulted because training did.have an effect on employment, estimated to be between 11 and 22 percent of the period after training (the differences depending on what variables were controlled in the analysis). Thus, while MDTA training probably did not help people get better paying jobs, it did help them to obtain more full-time employment and, thereby, increases in family incomes of approximately $500 per year. No attempt was made in this study to compute benefit- cost ratios. What the Study does present is a measure of net personal benefits from retraining, which is essentially the only-measure needed in the present study to determine if it pays the individual with poverty level income to undertake retraining. The measure of benefits obtained by Main provides others with one carefully estimated variable to include in their benefit-cost calculations. There have been only three basic smaller benefit- cost studies of retraining programs by Page in Massachusetts,87 Borus in Connecticut,88 and.”Somers and Stromsdorfer in West Virginia.89 There were also additional variations on 87David A. Page, "Retraining Under the Manpower Develop- lent Act: A Cost- Benefit Analysis," Public Policy, Vol. 13 (1964): pp. 258- 267. 88Michael E. Borus, "A Benefit-Cost Analysis of the Iconomic Effectiveness of Retraining the Unemployed," Yale Iconomic Essays, Vol. 4, No. 2 (Fall, 1964), pp. 371-429. 89Gerald G. Somers and Ernest W. Stromsdorfer, "A I-enefit— —Cost Analysis of Retraining," Seventeenth Annual I'I‘OEeedin- of the Industrial Relations Research Association Maoison: Univer31ty of Wisconsin Press ,1965), pp. 172-185. the ori studies brief 1 s. as wel retrai earnin more i contrc studis in the emplo: concl- a son COSt Hard. PP- 5 From 0f t1 Wm Strol the 1 1% (Mad ’0 ,_. NO / H. \I 99 the original studies for the two latter ones.90 Since these studies have been analyzed extensively elsewhere, only a brief summary of their analysis need be presented here. Somers discussed the West Virginia benefit-cost study, as well as several others which estimated the effects of retraining in.terms of increased employment and increased earning.91 He noted that all benefit-cost studies found more favorable relationships for trainees relative to control groups and that regression analyses in two of the studies found retraining to be.a major explanatory variable in the improved income.positions, largely through better employment records of formerly unemployed workers. Somers concluded that "the retraining of unemployed workers is a sound social investment."92 Mangum also briefly summarized the three-benefit- cost studies.93 While noting that such small samples of 90See Michael E. Borus, "The Cost of Retraining the Hard- -Core Unemployed," Labor Law Journal (September 1965), pp. 574- 583; Michael E. Borus,_11 Time Trends in the Benefits From Retraining in Connecticut," Twentieth Annual Proceeding 9f the Industrial Relations Research Association (Madison: Univers1ty of Wisconsin Press, 1968), pp. 36:46; Ernest W. Stromsdorfer, "Determinants of Economic Success in Retraining the Unemployed: The West Virginia Experience," Journal of Egman Resources, Vol. 3, No. 2 (Spring, 1968):PP. 139-168. 91Gerald G. Somers (ed.), Retraining the Unemployed (Madison: University of Wisconsin Press, 1968),pp. 7-13. 92Ibid., p. 7. 93Garth L. Mangum, MDTA: Foundation of Federal Manpower Policy (Baltimore: The John Hopkins Press, 1968), pp. 125- 27. train neces sugge shou] conci doub- pr0g: meas indi subs arr: fro Thi wid 0011 in pr di At tr trainees from particular locations and occupations are not necessarily representative of the entire—program, and suggesting that all benefit—cost analyse's conclusions should be treated with restrained skepticism, Mangum concluded that "There appears to be little reason for doubting the value of the manpower development and training "94 He believed that, by any reasonable rogram in general. measure, the program's benefits, both in society and the individuals retrained, have exceeded its costs by a substantial margin. The three benefit~cost studies under consideration arrived at.widely divergent benefit-cost ratios, ranging from 137.3 in the Connecticut study to 6.1 in Massachusetts. This "substantial margin" of benefits over costs, and the wide divergence between the estimates, raises doubts concerning the findings of these studies. Ribich has attempted to resolve these gross differences 95 n estimated benefit-cost ratios. He noted that the rimary reason for the discrepancies was the considerable ifference in the basic assumptions made in each study. fter developing a set of computational rules for the reatment of program costs, program returns, and the interest ate to be used in discounting present values of expected uture incomexstreams, Ribich recalculated the three 941bid., p. 127. 95Thomas I. Ribich, Education and Povergy (Washinton: The rookings Institution, 1968) pp. 38-50. benef assu: rati< and retr conc wort unaI of] cos est gre 101 :nefit-cost ratios by applying the rules and uniform :sumptions to each study. The resulting benefit-cost itios were 10.1 in Connecticut, 4.2 in Massachusetts, .d 15.0 in West Virginia.96 While the Ribich corrections still indicated that training is‘a socially profitable undertaking, questions ncerning the profitability of retraining poverty—level rkers (the original focus of this chapter? are still answered. Borus did make some estimates of the effects -retraining the hard—core unemployed, using the data his earlier study.97 While he found that program sts might be doubled for this type of retrainee, he timated that the hard—core groups would experience eater benefits as compared to others, even in excess these increased costs. Conflicting opinion was presented by Somers, however. noted that previous labor force experience appeared to a major predictor of post-training employment success, ase with more prior unemployment and who were in un— illed occupations being the least likely to find post- aining employment.98 This conclusion was reached on the sis of the earlier studies which examined programs not 96Ibid., p. 49,Tab1e 3. 97See footnote no. 90. 98Somers, Retraining..., p. 10. desig Else1 stat1 tom con age ill whi jot mar ma 0 Amid—S esigned specifically for the hard-core unemployed. lsewhere, using U.S. Department of Labor data, Somers tated:. In summing up the position of the disadvantaged, it can be said that their employment and earnings after retraining are not as favorable as the employment and earnings of other retrainees; but the labor— market position of the disadvantaged is considerably’ enhanced by their retraining, as compared with their own pretraining experience and as compared with disggvantaged workers who have not been re— trained. In a Michigan Study of MDTA retraining, Nosow also found that the most striking results were for persons :onsidered "marginal" to the labor force - the very young, Lged, poorly educated, nonwhite, female, and vocationally .ll-prepared.100 When trained specifically for a job in rhich labor was in short supply, marginal workers obtained obs and lost their marginal character. Successful labor arket outcomes, in terms of successful job placement, was ade possible for these people because of job retraining, osow concluded. The Main study examined above included programs ose emphasis was focused more on the hard-core group, 1though only about 50 percent of the enrollees could be classified. The results of more recent emphasis in the 99Gerald G. Somers, "Our Experience with Retraining and location," in R.A. Gordon (ed,);Toward A Manpower Policy ew York: Wiley, l967),p. 226. 'T— 100Sigmund Nosow, "Retraining Under the Manpower Deve10p- nt Training Act: A Study of Attributes of Trainees Assoc- ted with Successful Retraining,‘l East Lansing. Michigan ate University, School of Labor and Industrial Relations anuary 1968) (mimeo). pp. 3(1) - 3 (4). larger Sector educat more c level careft appear patins D. 1 in ea were expec would upon Prime anal: iner. of f Sect the time 103 larger cities with the JOBS (Job Opportunity in the Business Sector) program, and other programs including basic education training prior to job-retraining, will provide more conclusive evidence on the results of training poverty- level workars. Published studies of those programs which :arefully calculate benefit-cost ratios have not yet appeared. However, early reports from businesseSpartici- >ating in these programs are quite optimistic. J. Empirical Estimates of Potentials for Increased Earnings Through Retraining With the limited data available on estimated increases .n earnings obtained as a result of job retraining, attempts 'ere made to project the potential income increases xpected to accrue to low income RLS respondents who ’ould undergo retraining. These projections were based pon the three employment-income variables which received rimary attention in the several benefit-cost studies nalyzed above - increases in straight time hourly earnings; ncreases in family incomes; and increases in the extent f full-time employment. . Full Time Employment Increases The study by Main101 which was analyzed in the preceeding ection found that increases in family income were primarily 1e result of increases in percent of time employed full ime by training completers as compared to controls. lOlMain, A Nationwide Evaluation.... BY aPP durati survey at est group 1 the e in th in Ta to th Weeks Time Week Full Time \ 1-10 11.1 ful form 1119 M.D me Oct the 104 applying the findings of the Main study regarding ration of full time employment after retraining to the rvey low income group, it should be possible to arrive estimates of increases in full time employment this oup can expect to receive through retraining. Unfortunately, no directly comparable measures of e extent of full time employment existed for the groups the Main study and the survey respondents. The data Table 18 incorporate a conversion of the Main data the scale utilized in the Rural Life Survey, the number Table 18. :ks Employed Full Time for Respondents Having Some Full -me Work in the Reporting Period,lby Respondent Class Percent Distributions Main Study Groups RLS—Viable 1.50 Income Ratio, 1ks . by Income Class Egg. Completers Control .80 .99 1.19 1.49 1.50 .0 6.9 18.9 8.4 9.4 7.2 10.8 9.3 '40 34.1 52.2 37.7 33.0 27.6 24.7 29.6 59.0 28.9 53.9 57.6 65.2 64.5 61.1 lWeeks full time was translated from "percent of months l-time between training date and interview" for the Main ups; the years ranged from 1964 through 1966. The report- year for the RLS was 1966. Source: Earl D. Main, A Nationwide Evaluation of .T.A. Institutional Job Trainipg Programs (Chicago: iSnal Opinion Research Center, Univergity of Chicago), oberrl966, Table IV. 5, p. 79. weeks of full time work during the preceeding year. On basis of this limited data, it is apparent that no near cup and hav the apt IOC _105 leaningful estimates of increases in duration of full time :mployment can be made for survey respondents who would indergo retraining. The survey poor appear not only to rave had more favorable full time rates than the controls in :he Main study, but they also had rates which closely approximated those of Main's completers. This left no room for estimates of potential increases in duration to >e obtained through retraining. It was thought that the difficulty with this attempt :0 measure potential employment duration gains resulted from the non-rigid definition of full time employment atilized in the RLS study. Respondents were simply asked now many weeks they were employed on specific jobs in .966, and how many1of those weeks were worked full time. Indoubtedly many respondents, particularly self employed farmers, reported weeks of full time employment in excess 1f the number in which they actually worked a full forty our week. Since we have no way of knowing the extent f this over—reporting of full time work, there is no ay that the two studies can be reconciled to arrive t the desired estimates of expected increases in full ime work duration. . Increases in ConSumer Unit Income The result of increases in percent of time employed 111 time after retraining, according to the Main study, 1s to increase family incomes by an average of $500 r year for MDTA training course completers as compared to controls. the exten consumer increase As < status W1 to incom Consume Poverty By addi recalcu to read would 1 the re: ified the be 10W ir half 1 were 3 than recei arbit the 3 unit $500 lift 106 :rols. A second approach to empirically estimating extent go which retraining would conceiveably lift :umer units from poverty was to simply add the $500 :ease in income to each consumer unit's income. As detailed in Chapter IV, consumer unit income :us was determined by the ratio of income received .ncome required, that is rsumer Unit ihlncome’ReCeived 'erty Status Income Required = "Consumer Unit Indome Consumer Unit;_Consumer Unit Home Poverty Level GroWn’Food Modification adding $500 to the numerator of this fraction and ilculating consumer unit income status, it was possible teadily determine the effect this increase in earnings .d have on poverty level incomes. Table 19 presents results of this transaction with consumer units class— ed by income status "before" and "after" receiving benefits of retraining, subdivided by age and various income categories. The data reveal that while over ' of the potentially economically viable consumer units raised one category in the income ratio scale, less one fifth of the entire group below 1.50 income ratio ived sufficient increases to raise them above this trary poverty line. When the poverty line is set at 1.00 income ratio, just over one fifth of the consumer s would be lifted from poverty with the addition of to consumer unit income. Is it reasonable to conclude that job retraining would only twenty percent of the poverty level consumer units out of p characte: consumer the char The most status. was cla: certain been 01 noted i were cl data is which I of dif Main's ratios all < 1 lifted l RLS n the d of RL throu 3. T311311 the 1 107 of poverty? Probably not, because the "pretraining" racteristics of the Rural Life Survey low income sumer units appear to be considerably different from characteristics of Main's controls or his completers. most noticeable difference, of course, is their income tus. Whereas all of the RLS subsample considered here classified as being in poverty, it is reasonably tain that not all of Main's MDTA graduates would have n classified in poverty prior to training. It was ed in Section C that about one—half of the Main graduates e classified as "hard—core" unemployed. However, no a is given on annual incomes in the Main study from ch poverty rates might be derived. A second factor difference between RLS low income consumer units and n's respondents was the reporting of negative income ios by a number of the RLS farm consumer units (6% of (1.00). None of these had any possibility of being ted above even the 1.00 income ratio. Again, the general lack of comparability between the respondents and the Main study respondents precluded drawing of definite conclusions about the possibility {LS poverty level families being lifted from poverty >ugh job retraining. Increased Hourly Earnings From the U. S. Department of Labor data presented in es l6 and 17, the following items were computed: (1) mean posttraining hourly earnings for each pretraining In 19.61 In001 <0 0.81 1.01 108 Table 19. Income Status of RLS Consumer Units Before and After Addition of $500 "Retraining" Income Percent Distributions 7 Rural Life Survey Consumer Units Below 1.50 Income Ratio Before Training After Trainin Income Status ome Status Age of CU Head No Change Improved 0.80 2 45 72.3 26.8 45< 65 62.4 37.6 Total 67.1 32.9 0 - 0.99 < 45 36.2 63.8 45< 65 11.1 88.9 F7 Total 23.9 76.1 0 - 1.19 ‘ < 45 21.0 3 79.0 45< 65 18.3 ; 81.7 Total 19.6 g 80.4 o — 1.49 < 45 59.3 [ 40.6 45 <65 32.7 g 67.3 Total 49.8 : 50.2 a1 Lifted < 45 52.7 f 47.3 Step 45< 65 37.1 . 62.9 Total 45.4 f 54.6 11 Lifted < 45 79.2 20.8 flm wuowdxo ammouoma w>aomou EH ommouocH o>wooou o>Hooon on omoSu zuHB oEHu uawwmuum t HmudH cu wwuoowxm cu pouoomxo uoz wouoodxm mucopsommom manHmuuoum J. .J .13)... 434 naumun oEoonH pom mmchumm mHusom wsHsHmsuumom oHumm mEoocH oo.H BoHom muHCD Hwfismcoo mo>usm was wanHmuuonm :uoBuom as msoaum o 113 meu Hmuoy He -- -- 8.2 He H.HN «SN «2 -- -- Na NS 3&3 MN m N.@ om.N ON m.om mo.H we wN ¢.mN H.HN ea m¢.Nmnoo.Nw mH o H.HA w¢.N m N.Nm ¢Q.H MN NN m.¢m m.om om om.Hw-mN.Hw mH N H.HN mN.N w N.wm N¢.H HN mm o.Ho N.Nm sm ¢N.Hw-0m.Hw mH N H.HN No.N o N.wm aN.H oH mm o.No N.wm Hm m¢.kumN.Hw NN NH N.Nm qN.H m m.wm mm. m we N.¢w o.wm mm «N.kumN.om a w Nam 3.3, H N.$ 8.0...” N HN new mac mm £63 M+muH ..aum h H w.mnm o m Qnm muowmxo ommouoca.o>Hmuou :H ommouosH o>Hooop o>HsooH ou wmosu fiuHB 08H“ uanmuum w HmuoH ou venommxm ou wauowaxm uoz venomoxm mummwsommmm wanHmuuwum OHumm oEoosH 114' number of consumer units in poverty. This measure . the advantage of comparability between U.S. Department Labor reports and the RLS data. It also implicitly orporated changes in full time employment rates, the ‘iable which Main found to be the major cause of increases family income. These estimates of changes in poverty es with job retraining also assume that labor market ditions for RLS retrainees would be the same as con— ions for MDTA graduates in the Labor Department survey. VI. RETRAINING NEEDS AND POTENTIALS OF LOW INCOME PERSONS n Introduction To obtain a measure of the needs for retraining of the {ural Life Survey's low income level individuals, it was .ecessary to obtain a measure of their present level of kills and abilities and to compare this level with a min— 5 mum standard training requirement. Two recently published supplements to the U.S. Depart- L ent of Labor's Dictionary of Occupational Titles were tilized to develop a generalized "basic training re— uirement." The requirements for average performance on the obs for which unemployed workers have been retrained under DTA training programs were used as the base for determining he minimum training standards. Present levels of education, in terms of years of :hooling completed, were reported by each respondent and mess were used as a measure of general educational avelopment. The level of vocational training of respondents as determined by quantifying the amount of vocational iucation, on—the—job training, or other specialized vocational raining which they indicated they had participated in. Those low-income respondents who were found to be in zed of job retraining were then classified according to 1eir expected ability to successfully complete a job ztraining program. 115 116 . Development of Basic Training Requirements The purpose in developing a basic training requirement easure was to have a standard against which RLS poverty- evel respondents could be compared to determing the extent f their need for retraining. Such requirement must be xpressed in terms of a minimum skill level which, when ttained, should enable the poverty—level worker to earn an ncome sufficient to lift his family above the poverty level. Before such requirements are developed, it is necessary 3 define the type of requirements that are being established. ine has explained that educational and training requirements ay be thought of in terms of functional or performance re- 1irements, employer or hiring requirements, or the education- L attainment of workers at a point in time. The functional aquirements approach, defined as "the requirements determined 1 objective job analysis as necessary and sufficient to :hieve average performance in the specific tasks of the >bs,"102 was followed in this study. Although "average performance" is nowhere clearly de- .ned, estimates of physical demands, working conditions, and raining time have been made for each of roughly 14,000 >b titles identified by the U.S. Department of Labor in 102Sidney A. Fine, "Use of the Dictionary of Occupational .tles to Estimate Educational Investment," Journal of .man Resources, Vol. III, No. 3 (Summer, 19685, pp. 5‘660 103 :s Dictionary 9f Occupational Titles (DOT). The raining requirements were expressed in terms of a measure 5 general educational development (GED) and a measure of >cational training time required, the specific vocational reparation (SVP) scale. Three fundamental skills which people are supposed to :quire from general education were delineated: reasoning, 1thematics, and language. The requirements for each of these :ills on each job as described in the 293 were then determined, .sed on a seven point scale of development. This general ,ucational development is described as education of a general ture which does not have a recognized, fairly specific, cupational objective. Ordinarily such education is obtained elementary school, high school, or college. It also derives om experience and individual study."104 The SVP scale is a measure of the amount of time required learn the techniques, acquire information, and develop the cility needed for average performance on a job. 103U.S. Department of Labor, DictiOnary of Occupational tles, Vol I, Definitions of Titles, and Vol II, Occupational ass1fication,3rd, ed. (Washington: U.S. Government Printing fice, 1965 . 104U.S. Department of Labor, Selected Characteristics _Qgcupations (Physical Demands, Working Conditions, Training ES),A Supplement to the Dictionary of Occupational Titles, 1. ed. (Washington: U.S. Government Printing Office, 1966), A-5. 118 It includes training received through vocational educa- tion, apprentice training, in-plant and on—the-job training, and essential experience in other jobs. Each job title listed .n the DOT hasbeen rated in terms of functional training :equirements on both the GED and the SVP scales. When considering poverty-level respondents' needs for retraining, basic training requirements in terms of the IVP meaSure were utilized. Potentials for successfully :ompleting retraining programs were based upon the GED measure .nd corresponding skill levels deemed attainable by individuals 'ith given levels of general educational development. The procedures utilized in developing the basic training equirements and respondents' SVP and GED levels are de— cribed in Appendix A. A list of over 900 job titles was btained and weighted by total MDTA enrollments in these ccupational classifications.105 After recording all GED and VP measures for the 900 job titles separated roughly ccording to male and female occupations, the median SVP evels were determined (Appendix Table A—2). The "median 105U. S. Department of Labor, Manpower Administration, egister of Projects Approved Under the MDTA Through June, 2E; (Washington: by the Department, 1965). This was the 11y volume of the Registry which was published; it is elievéd that MDTA programs were developed sufficiently by Jne, 1965 so that job titles added to the approved list after lat date would have little effect on the overall dis— ribution of enrollments. nge" SVP established in this manner was SVP 5-7 for ies and SVP 4—5 for females (see pages 210through.214 : a complete description of the procedures followed). The lian range SVP levels were then designated as the minimum indard training requirements. It was found that the occupational categories utilized developing the training requirement could be classified :0 four broad skill-level groups as follows: SVP l and 2 = unskilled occupations SVP 3 and 4 = semi—skilled SVP 5 and 6 = skilled svp 7 and 8“ = highly skilled Thus, the median range for male occupations consisted of .her skilled or highly skilled occupations and for females included semi-skilled and skilled occupations. Respondent SVP and GED Levels The development of SVP and GED levels for respondents presented in Appendix C. In the Rural Life Survey, each pondent indicated the type of vocational program, if any, t he participated in while attending high school and the unt of vocational training received, if any, in the past years. Following weighting procedures outlined by Fine,106 high school vocational education and the portion of college cation considered to be vocational training were converted )_________________ 1 106Fine, "Use of the Dictionary...," p. 368. |_ .1- 1weeks of specific vocational preparation. This value was in added to weeks of post high school vocational training, in the total was converted to an equivalent SVP value “ording to the DOT scale. An arbitrary adjustment in the jue given to vocational training was made according to the (ent to which the training was utilized in recent employment. Each respondent's level of educational development, as Jsured by years of schooling completed, was utilized as basis for his GED level. Although it was recognized t the number of years of school attendance was not a per- it substitute for a measure of a respondent's general ed- itional development, it was used because no better .ernative was available from the survey data. The con- sion of years of schooling to the GED scale was made in 107 The con— ordance with the translation made by Eckaus. sion categories appear in Appendix Table C-l. Retraining Needs Based on Basic Training Requirements The median required skill levels have been developed in as of the vocational preparation required for average per— nance on the type of jobs for which workers have recently 1 retrained under the MDTA. The measured skill levels of potentially economically viable respondents described 1e were compared to these required skill levels and were 107Eckaus, R.S., "Economic Criteria for Education and .ning," Review of Economigg and Statistics, Vol. 46, 2 (May, 1964), pp. 181-190. 121 assified as less than, equal to, or greater than the re- ired levels. Table 24 contains the percentage distributions ‘ respondents' SVP levels in relation to the required levels ’ income ratios and separately by sex. For males with income ratios below 1.50, ninety—two ercent were classified as having skill levels below required :ill levels. There were no significant differences between 1e different low income classifications for males. The :rcentage of females below the required levels ranged from .ghty-four to ninety percent, with an overall average of .ghty-six percent. The difference between the several low 1come groups for females was statistically significant. Looking at all potentially viable survey respondents, the .fferences between the lower and higher income groups for 1th sexes were highly significant. The total percentages :low the required levels were eighty-six percent for males .d seventy-five percent for females. If respondents had dertaken any training which affected their measured SVP vels, they were more likely to be classified above the re- ired level than equal to requirements. It appears that the SVP measure probably underestimates spondents' actual skill levels, as it records only formal cupational training which was received in the last ten ars. Nevertheless, in comparison to the non-poor, those :h income ratios below 1.50 were significantly less well 5 on the skill level index. Also, rates of eighty—five: 122 Table 24. Relationship Between Potentially Viable Respondents' Skill Levels and Median Required Skill Level, by Income Ratio and by Sex; Percentage Distributions 1967 Rural Life Survey Individuals* Skill Levels Related to Median Levels Required Income Males % of Females .77 % of Ratio ( = > ' oH tuoz :uHB Hoom Hmuoa How Hmu0H pooz suHS uoom Ho>oH Qmw wuHmEom mHmHusmuom monz Qmw oHumm eaooaH om.H sonm mHmseH>HeeH N6>H=m 6HHH Hausa NeaH mGOHuanHumHQ unwouom mxom Nb .mtooz msHsHmHumM pqu mucoCsOQmom quosH 30H mo meHusouom wsHsHmHqu .wN mHan HBOH 8H 8H H.HN 8% 8H 8H H88. 6 504/ 3 m6 6 m m.m.q O.mfi o.Nm .EmeOHQ wGH/_/ mime N.Mm w.._...q m _nsHmuuou sH mmdUUSw q ¢.Nm fuom HmHuaouom toow H.HN q COHumocpo onmb no ouHcng Nma .oHnm H“ m m.¢H m.mH m.¢H nsHmHuuH NHnHmmom\ c.0N H.HN A“ H.HN m omo N m.o uschHmmsm ou coHu m.H N . . H.H w.o smocto onmH HHHB wN.H H.N H oso U NHso uHHmsHmuumM / 0.0 H wsHaHmuuom pom Hoom wcHsHmHuom uoom wsHsHmHuom Mom Ho>oH tooz HHHB uoom Hmuoe Mom HmHOH puuz nuHs Hoom Ho>oH emu monEom mHmHucuuom mchz mmw OHHmM waoosH m¢.H opon< mHmspH>HpGH Nu>usm oHHH Hmusm NomH deHuanHumHQ uncouom onw an .mpooz wsHsHMHuom nuHB musopsoamom oEoosHHHusmHm Ho mHmHusuuom msHsHmuuuM .mN mHHmH 139 of basic educational instruction, forty—two percent of the males and one-third of the females in the low income category would be classified as in need of dual education— training programs. The actual need for such programs undoubtedly lies somewhere within the range of these extremes of three percent and forty—two percent. The remaining majority of the low income respondents in need of training were found to have good potential for successful retraining, with fifty-eight percent of the males and fully two-thirds of the females so classified. The percentages of the total low income population re- presented by those in need of training are also presented in Table 28. Over half of the total low income individuals were classified in need of and with good potential for completing a job retraining program. Table 29 shows the degree of retraining potentials among the survey respondents with income ratios above 1.49. Although a smaller percentage of this group appeared in need of dual type education-training programs, roughly sixty percent of the total group was estimated to have need of and good potential for completing retraining programs. In all, eighty—five percent of the non-poor males and seventy-two percent of the females were classified as being in need of retraining. In conclusion, it can be said that only a small per— centage of the survey poor who need retraining appear to have no potential for successful completion of job retraining Programs, less than five percent of both sex groups. 140 Approximately one-third of both sex groups were considered to be borderline cases, their GED levels being low enough to cast doubt on their ability to complete a retraining program. It can be reasonably concluded that a majority . of the poor who need retraining, nearly sixty percent of them,have excellent chances of successfully completing regular MDTA type retraining programs. These conclusions were based upon a measure of the general educational achievement of each respondent. This is certainly not the only variable which determines one's potential ability to succeed with job retraining. Level of educational achievement has been shown to be closely related to extent of vocational training received, however. With the inconclusive evidence concerning the relationship between retraining success and other socio-economic variables which was discussed above, use of the GED measure does 1 l 1‘ 1 1 1 . facilitate the development of reasonable approximations Of potentials for successful retraining. VII. RETRAINING POTENTIALS BASED ON INTEREST IN RETRAINING A. Introduction It was concluded in Chapter III that the "human capital" modification cf marginal productivity theory provided an apprOpriate theoretical framework for the analysis of rural poverty problems. It was noted that, other things being equal, some individuals have relatively low marginal productivities because relatively little has been invested in them as human capital in terms of occupational or vocational training. The rate of return on the actual educational investment in them may be acceptable, but because so little has been invested in them their labor incomes are low and they are in poverty. The solution to the poverty-problem, from this analytical framework, is to raise the marginal productivities of such individuals by increasing the level ofinvestment in them as human factors. This increase in the quality of the labor input should, in turn, increase the returns to that factor. If sufficient investment is made, the individuals will earn labor returns sufficient to lift their families from poverty. The potential ability of job-retraining programs to lift rural households from poverty levels depends largely upon the willingness of workers in these households to Participate in the programs. This study now turns to an analysis of respondents' interest in participating in job— retraining programs. 141 142 B. Interest in Retraining l. The Decision Rule RLS respondents were aSked if they would be interested in taking a "free training course given locally which would qualify them for a better job." Enumerators were instructed to emphasize the three conditions regarding the training course — that it would be free, given locally, and would qualify the respondent for a better job in one or more. aspects; higher pay rate, less seasonality, more convenience, more permanence, etc. Ignoring the hypothetical nature of the retraining offer, the respondents' interest in participating in a retraining program would be theoretically determined by their evaluation of the net discounted value of the future income streams attributable to the improvement in their skill levels as compared with future income streams expected without job retraining. The decision to accept or reject participation in a retraining program has been expressed symbolically by Chesler as follows:119 T l R -_ _ B M1 = E P (Y: - Y ) (——--—-7— . C Where M represents the discounted present value of the future income stream from the start of training (t=o) to 119Herbert A. Chesler, "The Retraining Decision in Massachusetts: Theory and Practice," in Gerald G. Somers (ed.), Retraining the Unemployed (Madison: The University of Wisconsin:Press, 1968), pp. 149-170. 143 the time of retirement from the labor force (t); YR represents income.from post—training employment, and Y is income earned from work without being retrained; r'is the individual's internal rate of discount. Also, the symbol P denotes the probability that the individual will be capable of actually working in any given year or period of time, t; and C-denotes the Opportunity costs of training to the individual such as potential earnings foregone while in_training, the surrender of leisure time, the possible relinquishment of seniority and other rights accumulated on previous jobs, and the uncertainty of ob— taining employment upon completion of retraining. If respondents behave consistantly with the above deciSion rule, they would be interested in undertaking retraining if the value of M was greater than zero and they would be indifferent to the retraining offer if M equalled zero for them. If M were negative, non-partici- pation would be the rational choice. In general, the lower the present level of income earned by respondents the larger the value of M and, therefore, the greater likelihood that they would express an interest in taking retraining. This is also assuming that the retraining opportunity costs for lower incOme individuals are either the same as or less than those for relatively higher income individuals. It was clear that individuals with incomes signifi- cantly above the poverty line would express little interest 144 in job retraining. For most of them, the skill levels attainable through government sponsored retraining programs would be lower than skill levels already attained and the value of M to them would be negative. But for those respondents whose income levels were below or barely above the poverty line, the value of M would, in most cases, be positive. The hypotheses selected for analysis in this chapter were based upon the variables in the above decision rule. By examining these variables in detail in relation to respondents' expressed interest in participating in job retraining programs, the extent of concurrence with the decision rule on the part of low income respondents will be determined. This will permit the drawing of conclusions- about the economic decision making of these individuals and will indicate the importance of these variable in future considerations of retraining policies and programs in rural areas. The focus of attention in this chapter is upon all the potentially economically viable low income individuals in the survey inasmuch as such a small percentage of them possessed adequate levels of specific vocational preparation. Ninety-two percent of the males and 86 percent of the females had SVP levels below those required. It was assumed that those few skilled respondents were in need of some upgrading of skills that were either obsolete or in surplus in their labor market. There were 123 individuals 145 (7 percent of the subgroup) for whom the question concerning interest in retraining was non-applicable because they were still attending school or college or their records were incomplete. The total number of responses in some tables will vary due to various control variables applied. All hypotheses were tested separately for males and females in reflection of the types of retraining courses offered under MDTA programs which are often designed primarily for either males or females. Also, in VieWCQf the differing societal expectations of the work roles of males and females in the family unit, it was expected that considerable differences in response to the retraining query would exist between the sexes and should be controlled for in the analysis. Although the original question concerning interest in retraining allowed a conditional interest response (“if it doesn't interfere with my current work"), responses were classified here as "interested“ or "not interested" only. The conditionally interested responses were grouped with "interested" responses because, first, the survey editors believed that interviewers did not usually probe deeply enough to elicit the conditional affirmative response. Secondly, when the two responses were not grouped several cross tabulations could not be statistically evaluated because of low theoretical frequencies in several cells of the contingency tables. 146 '2. Hypotheses It was hypothesized here that poverty-level workers would have reasoned consistantly with the above economic decision rule, therefore: Hypothesis (a) - There will be a relationship between the severity of poverty and extent of interest in retraining, those poor respondents with relatively lower incomes being more interested in retraining than those with relatively higher incomes. This hypothesis follows from the fact that lower current income levels will give lower values to Yt' the expected future income stream without retraining, in the decision rule and this will lead to larger values of M. The data for testingtflfixshypothesis are presented in Table 30. By the chi—square test of independence, this hypothesis was not supported by the survey data for either male or female respondents. In fact, the small but non-significant relationship which did exist for males was in the opposite direction than was hypothesized. That is, as poverty status increased, the degree of interest in retraining among males also increased. For females, the relationship was in the direction hypothesized but it also was not statistically significant. Two features of the data in Table 30 do stand out and merit additional comment. First, it is apparent that there did exist a considerable interest in participating in free job retraining programs on the part of the poor 147 Table 30. Interest in Retraining in Relation to Poverty Status, By Sex; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable | Males Females 1 1 Poverty Retraining Interest St t l' % of ‘ ”4% of“ a us 1* NI* Total 1* N1* Total 1.20 c 1.50 50.3 49.7 41.4 37.4 62.6 41.1 1.00 4 1.20 44.1 55.9 20.5 28.0 72.0 17.7 0.80 4 1.00 39.3 60.7 12.0 38.8 61.2 13.7 < 0.80 46.1 53.9 26.1 36.5 63.5 27.5 Total 46.6 53.4 100 35.7 64.3 100 *I = Interest; NI = no interest. Chi—square at .10 level with 3 d.f. = 6.251; calculated chi-square for males = 4.007; for females = 5.053 (neither is significant). segment of our open-country population. Although those respondents in the lowest income categories did not express more interest in retraining than those in slightly higher categories, forty-seven percent (nearly one—half) of all the males and thirty—six percent of all the females did express interest in retraining. Among the total survey population one—third of the respondents expressed an interest in participating in a job retraining program, so the rate for the potentially viable low income level females closely approximates the total survey rate, while the males in this low-income subgroup expressed considerably more interest than did the non—poor in the survey. Thus while it appears that between the different low income categories 148 the respondents did not react as expected, as a group compared to other survey respondents the low income males did behave consistantly with the decision rule in expressing greater interest in retraining than the non-poor. The greater than one-third rate of interest in re— training among low—income females is the second point which deserves comment. In View of the fact that only twenty-one percent of the potentially viable low income females were employed at the time of the survey (see Table 34) and under half of them expressed an interest in retraining, approximately one—fourth of the low income females were not working but were interested in participating in job retraining programs in preparation for obtaining employment. Hypothesis (b) - There will be a relationship between age of respondent and his interest in retraining, among all low income respondents, with older workers showing less interest than relatively younger workers. This hypothesized relationship should appear because the older the worker the shorter the time period in which he can expect to earn the Y: income and the smaller the value of M. In addition, costs may be greater for the older worker in terms of seniority and other benefits foregone by changing jobs. This hypothesis was tested with the data in Table 31. Here, the chi—square test of independence indicated that older respondents were less interested in retraining than 149 Table 31. Interest in Retraining in Relation to Age,,by Sex; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable Age Males Females Levels Retraining Interest Retraining Interest 1* NI} '3’”be I? NI '7 %' Of Total) lTotal < 35 44.6 55.4 41.2 37.8 62.2 45.7 35 - 44 56.9 43.1 27.6 41.5 58.5 24.5 45 - 54 46.8 53.2 18.8 32.1 67.9 14.8 55 - 64 30.4 69.6 12.4 23.3 76.7 15.0 Total 46.6 53.4 100 35.7 64.3 100 *Chi—square for males = 18.791 (with 3 d.f., sig. at .001 level); for females = 13.572 (with 3 d.f., sig. at .01 level). younger workers among both males and females; the relation- ship was in the expected direction and highly significant for males, and as hypothesized and significant for females. Nearly seventy percent of the respondents in this sub- group were under forty—five years of age, and the interested rate among them was nearly fifty percent for males and about forty percent for females. The lower rate of interest among the under 35 age group was undoubtedly due to the influence of the very young, unmarried respondents who faced less economic responsibilities and who had had fewer negative experiences in the job market because of their lack of skills. It is significant to note that one—third Of the males over age fifty—four, and nearly one—fourth 150 of these older females, also expressed an interest in job retraining. Hypothesis (c) - Among all low income respondents the amount of prior occupational training received, and the extent of its use, will influence the degree of poverty level respondents' interest in retraining; those respondents with relatively more training and who used it relatively more in recent employment will be less interested in (additional) retraining than those with relatively less prior training or who used their training relatively less. This relationship should exist because of two influences. First, workers with more prior training will have higher expected Y than workers with less or no training.-Secondly, t workers who have had vocational training but still earn poverty-level incomes are likely to place a lower expected value on YE than workers who have not had prior training. These two factors will both tend to reduce the value of M for respondents with relatively more prior training as compared to those with less training. b Since the SVP value for respondents which was cal— culated in Chapter VI incorporated both the amount of prior training and the extent to which it was used, the SVP scale was used here as the independent variable. The data for testing this hypothesis appear in Table 32. The SVP levels of respondents were grouped initially according to the broad skill level categories identified in Chapter VI; unskilled, semi—skilled, skilled, and highly skilled, 151 Table 32 Interest in Retraining in Relation to Amount of Prior Training, by Sex; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable ' SVP Males Females Level Retraining Interest Retraining Interest % of % of I NI Total I NI Total Special 44.0 56.0 86.1 34.7 65.3 86.6 1 - 4 70.2 29.8 6.4 50.0 50.0 2.9 5 — 6 60.0 40.0 2.7 60.0 40.0 0.6 >6 55.6 44.4. 4.9 38.6 61.4 9.9 Total 46.6 53.4 100 35.7 64.3 100 For males, chi—square = 14.923 (with 3 d.f., sig. at .01); females could not be statistically tested. plus the "special" category for those whoSe SVP level was essentially non—computable and presumably equal to zero. It was found that the frequencies were so low in the two lower skilled categories that SVP levels 1—4 were grouped together in the final analysis. The hypothesis was initially tested with the inclusion of the "special" category containing all those for whom no SVP was computable-— thPSe with no formal vocational training. For both sex groups the persons in this category had the lowest interest in training rate, so low, in fact, that with over eighty- SiX percent of all respondents in the "special" classifi- cation, the hypothesis had to be rejected. For females 152 no judgment of the significance of the relationship between SVP and retraining interest could be.made due to low theoretical frequencies in several cells, while for males the relationship was significant but not in the hypo- thesized direction.' A further test of the hypothesis was made by excluding the "special" category and looking only at respondents with a measureable SVP level. Males and females were combined due to the relatively small sample size. As- shown in Table 33, the relationship was then in the hypo- thesized direction. The level of significance was only .05, however, again due in part to the small sample size (222 total cases). Perhaps of must value here is the knowledge that less than fourteen percent of all potentially viable low income respondents had received training which could be included in this study's measure of vocational training. Over one— half of those with measureable skill levels were classified in the highly skilled category. Hypothesis (d) - Among all low income respondents, those who were unemployed at the time of the interview and those who experienced relatively more unemployment in 1966 will have more interest in retraining than those respondents with relatively less 1966 unemployment and those who were employed at the time of the interview. This relationship results from the effect of unemploy— ment on expected income; the more an individual has been 153 Table 33 Interest in Retraining in Relation to Amount of Prior Training for Respondents with Measureable SVP; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable Retraining Interest SVP % of Level I NI Total ' 1 - 4 f 63.0 37.0 32.9 5 - 6 1 60.0 40.0 . 11.3 >6 i 43.5 56.5 55.8 Total 51.8 48.2 100.0 1 Chi square = 7.290 (with Z.d.f.sig. at .05). recently unemployed the more likely he is to View Yt as smaller, and YR as larger, than will the person with t relatively less unemployment. These two elements both lead to a larger value of M for workers with greater unemployment.‘ As hypothesized, interest in retraining in relation to weeks of unemployment in 1966 was considerably greater for both males and females with relatively more weeks of unemployment. However, the number of individuals who were considered to be unemployed at all in 1966 was so relatively small (only 9 percent of the males and 5 percent of the females) that no statistical tests could be reliably made. 154 The problem arose due to the use of "official" definitions of unemployment. An individual was classified as unemployed only if he was either "not working but looking for work" or receiving unemployment compensation, or both, at some time during 1966. All workers who were unemployed but not eligible to receive unemployment benefits, and all underemployed, were excluded from this calculation. It is doubtful that ninety-One percent of the males whoSe families had poverty level incomes experienced absolutely no unemployment during the year. No meaningful conclusions could be drawn from the data on 1966 unemployment. Whereas the RLS questionnaire obtained extensive information regarding the recent employment experience of respondents, it did not enquire directly if the respondent was presently employed. ‘It was necessary therefore, to determine present employment status (i.e. at the time of the interview) by screening answers to two other questions. After each 1966 job, if any, was described, respondents indicated whether they were still working at that job or not. If no 1966 job was stillhheld and if no new 1967 job was described, the respondent was considered unemployed at the time of the interview. This unemployment measure may have overstated the Percentage of workers unemployed as compared to U.S. Department of Labor figures because it included persons who were not working and not looking for work at interview time, that is, those who had withdrawn from orlhad not yet 155 entered the labor force. On the other hand, the under— employed were again classified as employed by this measure. Although this was in many respects an unsatisfactory. measure of unemployment, it provided a rough estimate for the purpose of this study of the number of unemployed persons who might be available for employment and who might be interested in participating in job retraining. As shown in Table 34, the relationship between current employment status and interest in retraining was statisti- cally significant, but not in the hypothesized direction. Table 34 Interest in Retraining in Relation to Employment Status at Time of Interview, by Sex; Percent Distribution 1967 Rural Life Survey l Employ— Male 1 Female ment Retraining Interestl Retraining Interest Status NI Total I NI ' Total (% of) (% of) Employed 52.2 47.8 84.6 4515 54.5 21.1 Unemployed 15.8 84.2 15.4 33.0 67.0 78.9 Total 46.6 53.4 100 35.7 64.3 100 Chi square for males for females chi square = .001); at .01). 51.477 (with l d.f., sig. at 9.898 (with l d.f., sig. Over half of the employed males, and nearly as high a Percentage of the employed females, expressed interest 156 in retraining. The rate of interest among unemployed males was quite low. By way of explanation, it is probable that many workers who were classified as employed but who-were actually underemployed were very interested in the possibility of obtaining job training. On the other hand, many of those considered to be unemployed were not in the work force at all and therefore they were not interested in retraining. It would certainly appear that many employed low income workers, by any measure of employment status, would be intereSted in job training which would enable them to obtain "better" jobs. Hypothesis (e) - Among all low income respondents for those who looked for some type of a job during the past three years, those whose.biggest problem in finding employment resulted from their lack of skill, education, or training will express more interest in retraining than either those who experienced no problems or those whose problems were not related to such personal inadequacies. This relationship is the result of the effect of problems due to personal training inadequacies upon an individual's view of benefits to be gained from correction of those inadequacies. In the theoretical decision rule framework, this would lead to a larger value assigned to 1 Y5, and thus to larger M for respondents having had this type of difficulty finding employment. The data relating to this hypothesis appear in Table 35. Only twenty—three percent of the low income males reported having looked for work in the past three years, 157 Table 35 Relationship Between Difficulties in Finding Employment and Interest in Retraining, by Sex; Percent Distribution Major Male Female Problems in Retrainingplnterestj Retraining Interest Finding % of % of Employment I NI Total I NI Total No Problem 64.3 35.7 30.7 \46.9 53.1 31.4 No Personal Inadequacies 70.2 29.8 45.6 62.5 37.5 39.2 Personal Inadequacies 74.1 25.9 23.7 85.0 15.0 29.4 Total 69.3 30.7 100 64.2 35.8 100 L . -— N 1 (158) (70) ((228) (131) (73) f(204) and only thirty—one percent of the females so reported. Of those who sought employment, about one—fourth reported that some personal training inadequacy was the major problem they encountered in finding a job. The rate of interest in retraining for respondents with these problems was somewhat higher than for those facing different types of problems and considerably higher than for those who reported no problems finding employment, thus supporting the hypothesis. In view of the extent of low income problems among this population it is rather surprising to find such a small percentage of respondents who reported having sought employment in recent years. Among those who did seek employ- ment the overall rate of interest in job retraining was 158 considerably higher than for the total subgroup; it was nearly seventy percent for male and sixty-four percent for female employment seekers as compared to forty-seven and thirty-six percent respectively for the total subgroup. Hypothesis (f) - There will be a relationship between the respondents' interest in retraining and their availability for employment in a (different) job during the following year, those declaring their availability being more interested in retraining than those who would not be avail— able. Clearly, the worker who sees himself available for a different job in the near future will be more likely to view retraining as one way of obtaining a different job. This will lead to larger values of Y: for him relative to the person who does not consider himself available. In addition, because of his apparently weaker attraction to what work, if any, he is now doing, the available worker will face smaller opportunity costs from participating in retraining relative to his non-available counterpart. Both of these factors will tend to increase M for the person available to work at another job. Two measures of respondent availability were utilized to test this hypothesis. The first measure was taken from the respondents' indications of their availability in the following year and this is tested with the data in Table 36. The second measure represented an attempt to determine the respondents' availability in relation to 159 Table 36. Interest in Retraining in Relation to Stated Availability, by Sex; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable Availability ' Male Female Retraining Interest Retraining Interest for Job % of % of I NI Total I NI Total Available 80.8 19.2 36.1 75.6 24.4 26.7 Not Available 27.4 72.6 63.9 21.6 78.4 73.3 Total 46.7 53.3 100 36.0 64.0 100 N (343) (392) (735) (316) (562) (878) Chi square for males = 193.479; for females = 217.711 (both sig. at .001). how well their 1966 job or jobs provided reasonably stable employment and, particularly for female respondents, the extent of family obligations which might keep them out of the labor market. This calculated measure of availability and its relationship to interest in retraining is given in Table 37. Those respondents who considered they would be avail— able for alternative employment in the succeeding year were Very significantly more interested in retraining than those not considering themselves available, as was hypothesized. Not too surprisingly, in View of the low level incomes received by respondents, over one—third of the males and over one-fourth of the females declared that they would be available for a job, or another one. Inasmuch as the data 160 Table 37. Interest in Retraining in Relation to Potential Availability, by Sex; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable Potential Male Female Retraining Interest Retraining Interest Availability % of % of I NI Total I NI Total Not expected to be 50.1 49.9 68.8 36.1 63.9 47.9 Available Can reasonably expect .35.6 64.4 27.3 30.5 69.5 39.6 available Expected to be 62.1 37.9 3.9 50.5 49.5 12.5 Available Total 46.6 53.4 100 35.7 64.3 100 N (345) (395) (740) (316) (570) I (886) Chi square for males = 15.035 (with 2 d.f., sig. at .001, but negative); for females = 14.714 (with 2 d.f., $19. at .001). in Table 34 revealed that eighty—five percent of the males were classified as employed at the time the survey was taken, it is clear that many of the employed were prepared to change jobs and the vast majority of those prepared to change were also interested in receiving retraining to assist them in obtaining alternative employment. When a determination of respondents' availability for alternative employment was made on the basis of recent work history and family obligations, the non—availability 161 rate among males was very close to the rate of stated non-availability-- sixty-nine percent as compared to sixty-four percent respectively, (Table 37). Again, nearly one—third of the males were determined to be potentially available for other employment.' However, the relationship between this measure of availability and interest in retraining was not in the hypothesized direction, although it was highly significant statistically. While those males definitely expected to be available had the highest rate of interest in retraining, those "reasonably expected available" had the lowest rate. It was-clear that the calculated measure of availability did not identify as available those males who declared that they would be available for alternative employment. As Stated above, the measure of declared availability for males was highly significantly related to interest in retraining. For female respondents, the scale of calculated availability was in the hypothesized direction and highly significant. Even then, the percent interested in retraining among those definitely expected to be available was twenty-five percent less than among those whoHstated that they would be available. The apparent weakness of the calculated availability code was that it identified as reasonably expected available respondents who, although not participating fully in the work force, had relatively less interest in undergoing retraining so that they could 162 be-more fully employed. Hypothesis (g) - There will be a relationship between the poverty-level respondents' attitudes toward their present jobs and their interest in retraining. Respondents who like their present jobs and would recommend this work to others will have relatively less interest in retraining than workers who hold more negative views concerning their present job. Two forces serve to cause the hypothesized relationship. First, if one likes his work he is receiving some amount of psychic income from it, so Yt will be relatively higher for him than for those who like their work less. Secondly, switching (hypothetically) from employment one likes to a presently unknown occupation becomes an added cost of retraining. Both of these forces would tend to lower the value of M for persons who like, relative to those who dislike, their present employment. This relationship was tested with the data in Table 38. Among male and female respondents classified as "employed", there was found a highly significant relationship in support of this hypothesis. Workers' interest in job retraining is closely related to their attitudes toward present employment, presumably even when they receive poverty level wages from that employment. It is also surprising to note that over sixty percent of the employed low incomemales, and over fifty percent of the females, reported that they liked and would recommend to others their present job. 163 Table 38. Interest in Retraining in Relation to Attitude Toward Current Job, by Sex; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable Attitude Male Female Toward Retraining Interest Retraining Interest Current Job~ % of % of I NI . Total I NI Total (1) like and recommend 46.9 53.1 61.7 45.3 54.7 51.7 (2) like only 50.0 50.0 16.5 16.0 84.0 17.2 (3) just a 1 job 59.2 40.8 12.8 .21.4 78.6 9.7 (4) would change if 80.0 20.0 9.0 80.6 19.4 21.4 possible Total 52.0 48.0 100 45.5 54.5 100 N (289) (267) (556) (66) (79) (14§T For males, chi square - 20.828 (with 3 d.f., sig. at .001); for females, chi square = 27.485 (3 d.f., sig. at .001) O Hypothesis (h) - There will be a relationship between degree of interest in retraining and potential costs of the re- training program. Among those poverty-level workers expressing an interest in retraining, progressively less interest will be expressed as potential program costs increase. With a free training program the value of C in the decision rule for respondents derives primarily from the several opportunity cost items mentioned above (p.143). As the potential dollor costs of the program increase, the 164 size of C increases, reducing M and decreasing the extent of interest in retraining. The percentage changes in retraining interest with increasing potential costs of retraining are presented in Table 39. Table 39. Percentage Changes in Retraining Interest With Increasing Potential Costs of Retraining, and Percent of Total Poor viable Population, by Sex 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable Potential Males Females Retraining No. and Percent Interested in Retraining Costs I %I %Change—Cum. I %I %Change-Cum. (1) Free 345 100 - - 316 100 - - (2) O < $100 232 67.2 —32.8 -32.8 187 59.2 —40.8 —40.8 (3) $100< $500 66 19.1 -48.1 —80.9 26 8.2 -51.0 -9l.8 (4) 2$500 20 5.8 —13.3 —94.2 4 1.3 - 6.9 —98.7 Percent of Total Poor Potential Males Females Retraining No. and Percent " ’ _ No. and Percent Costs Interested Interested I %I %Change—Cum. I %I %Change—Cum. (1) Free 345 46.6 — - 316 35.7 — — (2) 04 $100 232 30.1 -l6.5 -16.5 187 19.4 —l6.3 -l6.3 (3) $100‘<$500 66 8.5 -21.6 —38.1 26 2.7 —16.7 —33.0 (4) 2$500 20 2.6 — 5.4 -43.5 4 0.4 — 2.3 —35.3 It was very clear that the rural poor responded consistantly with the decision rule to hypothetical increases in the cost Of a hypothetical retraining program. Only respondents who answered affirmatively to the "free" training offer were asked 165 the succeeding questions, so it is not known if some respon- dents would have expressed interest in a retraining program with a positive dollar price tag. Such is possible among people who might view the free program as a form of "charity" or "socialism". For the most part, however, respondents lost interest very rapidly as potential training costs increased. The percentage of the total low income viable population interested in training at the different price levels, and percent changes, are also presented in Table 39. Hypothesis (i) - There will be no relationship between the poverty—level respondents' education level, in relation to the median education level for their age, and their interest in retraining. Respondents whose educational level is below the median will be as interested in retraining as those with educational levels equal to or greater than the median for their age group. This hypothesis was made on the assumption that the concentration of workers with relatively high educational levels in Federal MDTA training programs to date has been more the result of admission screening practices of program administrators than of less interest in retraining on the part of the more poorly educated workers. By using education in relation to the median for the respondents age level, as opposed to years of schooling completed, age is controlled 166 for in this relationship.120 The data testing this hypothesis appear in Table 40. Table 40. Relationship Between Education and Interest in Retraining, By Sex; Percent Distribution 1967 Rural Life Survey Individuals Below 1.50 Income Ratio and Viable Education Male Female in Relation 'Retraining Interest Retraining Interest to Median I NI % of I NI % of Total Total <; 50.9 49.1 52.0 36.9 63.1 41.5 : 41.9 58.1 42.8 34.7 65.3 53.3 > 42.1 57.9 5.2 39.1 60.9 5.2 Total 46.6 53.4 100 35.8 64.2 100 N (343) (393) (736) 7316) (566) (882) Neither relationship is significant above the .10 level. As hypotheSized, there is no significant relationship between the two variables for either sex group. Respondents with below average educational levels, as measured by years of schooling completed, did not view their lack of educational achievement as a barrier to participation in job retraining PrOgrams. 120The median levels were those computed for Ohio males from the 1960 Census. The medians for all other residents of the ENC states—W§?€_not significantly different from those for Ohio males. The procedure used to develop these medians was described in Marvin E. Konyha, "Educational Medians in the East North Central States, 1967," Michigan State University, East Lansing, 1969 (mimeo). 167 C. Retraining Potentials This analysis of variables related to interest in participating in job retraining programs has indicated that considerable potential does exist for enrolling low income rural people in retraining programs in this sub- region. Nearly one-half of all potentially economically viable males, and over one-third of the females, indicated that they had an interest in participating in job retraining programs. Nearly seventy percent of the hypotheses tested were supported by the analysis, indicatingithat the low income rural residents of this subregion tended to evaluate the expected returns from retraining in a manner consistant with the decision rule. An extensive analysis of the relationship between retraining needs and potentials among the rural poor and of the implications of the hypotheses explored in this chapter are presented in the next two chapters. VIII. SUMMARY AND CONCLUSIONS A. Introduction Three general hypothesesprovided the investigative framework for this study. The first, and most general, was that many of the households or consumer units in the open country rural areas of the Eastern Corn Belt States were receiving poverty level incomes in 1966. This hypothesis was based upon 1960 Census data which revealed that nearly one-fourth of the rural families in this region received net money incomes of less than $3,000 in 1959, and upon a 1966 Census Bureau survey which found poverty rates in the entire North Central region of nearly fifteen percent among rural farm families and sixteen percent among rural non-farm families. The second general hypothesis reflected the recent emphasis on manpower training programs in the nation. It stated that the successful completion of a job retraining program by the worker member(s) of the area's rural, poverty level consumer units would be expected to provide sufficient increments in income (assuming available job Opportunities existed) to lift the units above the poverty line. The third general hypothesis was that worker members of rural, poverty level consumer units in the Eastern Corn Belt possessed characteristics and attitudes that would make them prime candidates to undertake and successfully complete job 168 169 retraining and to accept different, more remunerative employment. The summarization of the findings related to each of these major hypotheses and various subhypotheses, and the conclusions drawn on the basis of these findings, are presented in the subsequent sections. - B. Rural Poverty in the Eastern Corn Belt The report of the National Advisory Commission on Rural Poverty121 and numerous journal articles and Depart— ment of Agriculture reports helped draw considerable national attention to the overall nature and extent of rural poverty in the United States. Even prior to the Rural Poverty Commission's formulation, much national energy had been devoted to overcoming the problems of regionally concentrated poverty conditions, whether rural or urban in location. Prior to the inauguration of the present set of studies by the Economic Development Division and the Office of Economic Opportunity, however, there were no comprehensive studies of the exact nature and extent of rural poverty in regions characterized primarily as commercial agriculture regions. In determining the extent of poverty in this sub— region, consumer unit income was the measure utilized to determine poverty status. Consumer units consisted of 121National Advisory Commission on Rural Poverty, The People Left Behind, . . . . 170 1 either family units or unattached individuals, and they were classified according to farm or non—farm status. The poverty status of consumer units was determined through the procedure developed by Orshansky122 for the Social Security Administration with one major modification. Whereas the Orshansky procedure assumed that farm families meet thirty percent of their food needs through home-grown foods, this study determined the actual percentage of food requirements met with home-grown food for both farm and non—farm survey consumer units. The poverty status ratio was the ratio of income received to income required, with requirements based on size, age, and sex characteristics of each unit. Consumer units with income ratios below 1.00 were considered to be in poverty by this criteria. Several conceptual limitations of the Orshansky criteria were recognized. I Sixteen percent of all survey consumer units had income ratios below 1.00 in 1966, with the percentage among farm units only slightly higher than for non—farm units. However, the rate was nearly fifty percent among unattached individuals as compared to less than twelve percent for family units. On the basis of potential economic viability, the extent of poverty was found to be much greater among the aged and the handicapped as compared to the young and healthy. Thirty-five percent of the consumer units whose head was over 64 years old were in poverty, while just over nine percent of the potentially viable units, ¥ 122 H Orshansky, "Counting the Poor..., . 171 those whose heads were under 65 years of age and not handicapped or seriously disabled, were below the 1.00 income ratio. The potentially viable poor consumer units represented forty—two percent of the poor c0nsumer units in the survey and less than seven percent of the total number of survey consumer units. The poverty rate for individuals over age 15 in the .survey was just under thirteen percent, with a farm rate of fifteen and one-half percent and a rate among non— farm individuals of just over eleven percent. Fifty— two percent of the poor individuals were potentially economically viable. For retraining purposes, all individuals with income ratios below 1.50 were classified as "poor". One—fourth of the survey population was included in this group of poor individuals, and sixty—three percent of the group were classified as potentially economically viable. Those below the 1.50 income ratio who were viable represented one-fifth of all potentially viable respondents and one— sixth of the total survey population. Conclusion The data on the extent of poverty level incomes in this subregion, utilizing the definition of poverty income described in Chapter IV, confirmed the hypo— thesized existence of poverty level incomes among many of the rural, open country consumer units in the Eastern Corn Belt States. 172 C. Job Retraining — The Escape Route From Poverty 1. Labor Theory The two prevailing labor theories today are marginal productivity theory and the bargained theory of wages. The bargaining theory is built around the concepts of comparable wages for comparable work, the cost of living, labor productivity, and an employer's ability to pay (profits). This theory does an acceptable job of explaining short-run variations in wage levels. The marginal productivity theory states that, in the long-run, under perfect competition, workers' wages will equal their marginal productivity. The perfect competition theory assumes full employment of resources, perfect mobility of productive factors, perfect knowledge, and rational economic motivations. But full employment has eluded our modern industrial economy, not all workers are fully cognizant of wages and employment opportunities, and there are numerous barriers to occupational and geographical mobility of labor resources. Nevertheless, by concluding that worker wage rates differ because worker marginal productivities differ, marginal productivity theory has led to policy prescriptions which, when carried out, will improve the functioning of the labor market and reduce the extent of poverty level incomes. 2. Human Capital Investment By recognizing that workers are not a homogeneous factor, that they do have differing marginal productivities, 173 the theoretical solution was to recognize that workers were a special type of capital, human capital, and that returns to labor could be enhanced by additional in— vestment. This human capital investment takes the form of education and training. Studies of the payoff rate for education at all levels indicate this rate is very high, and it appears to be even higher for vocational training programs which focus specifically upon low income (low marginal productivity) workers. Improvements in one's education and training are recognized by economists as one of the most effective ways of eliminating poverty. Conclusion It was concluded, on the basis of the investment in human capital modification of marginal productivity theory, that job retraining has the theoretical potential for lifting rural poor families out of poverty. 3. Benefits of Job Retraining - Approaches to Benefit Estimation One approach to estimating the potentials for job retraining to lift low income consumer units from poverty is to compare the present value of the future income stream which can reasonably be expected from the worker's present occupation with that of the income stream he can reasonably expect from the occupation for which he will be retrained. The two available attempts at making such 174 comparisons proved to be unsatisfactory for generating the type of estimates sought in this study, however. The second approach to estimating retraining's potential for lifting consumer units from poverty is to estimate-the increase in earnings retrained workers may expect with the use of benefit-cost analyses. Although the application of benefit-cost procedures to retraining programs has been plagued with serious methodological problems, some estimates have been completed which indicate the benefits to society are definitely in excess of societal costs. The focus of these studies was on returns to society rather than on retraining's potential for lifting families from poverty. They did indicate, never— theless, that increases in retrainee incomes would be considerable. One study of retraining completers and a reasonably similar group of controls who did not retrain indicated that completers experienced about a $500 increase in annual family income as compared to controls. The reason for this increase was that completers experienced more full—time employment after training than controls. Empirical Estimates of Retraining Benefits Three attempts were made to apply the findings of previous benefit—cost studies to the Rural Life Survey poor consumer units to estimate their potential for escape from poverty with retraining. No estimates of expected increases in full—time employment could be made because of noncomparable 175 data by which to compare RLS respondents with those in prior studies. Application of the $500 increase in annual consumer unit income to RLS units also gave unsatisfactory results because of the clear differences in the background characteristics of RLS respondents as compared to prior study respondents. The third estimate of rates of escape from poverty combined reported increases in hourly earnings of MDTA institutional training program graduates with the percentage distribution of RLS respondents on the income ratio scale. It implicitly incorporated increases in full—time employment. This procedure resulted in the estimation that only thirteen percent of the consumer units below the 1.00 income ratio, and just over one—fourth whose income ratios were below 1.50, would be expected to remain below the respective ratio after completing job retraining. Those consumer units expected to remain below the 1.00 income ratio by this estimate represented less than one percent of all potentially economically viable consumer units, compared to six percent prior to "retraining"; the percent of the total units below the 1.50 income ratio was reduced from seventeen percent to less than five percent with this adjustment. There estimates were only intended to be indicative of the income ratio changes attainable through retraining; they are not definitive projections. Conclusion It was concluded that the benefit—cost analysis approach to estimation of retraining's potential for lifting 176 families from poverty does provide an adequate methodological framework for these estimates. Estimates based on increases in hourly wage rates, combined with income ratio distribution of the RLS respondents, indicated that a large proportion of potentially viable, rural, low income consumer units in this region could theoretically be lifted from poverty through job retraining. D. Characteristics and Attitudes Related to Retraining l. Retraining Needs — General Needs of Low Income Rural Workers Considerable evidence was presented in Chapter II to . . . . . 4 indicate that economists, including Hathaway,123 Schultz,12 and BishOp,125 clearly understood the extensive needs for vocational retraining among the thousands of workers who would be displaced from the agricultural labor force in the 1960—1969 decade. The available evidence on the extent to which rural low income workers have participated in Federal manpower retraining programs was found to be unclear. Tweeten126 noted that farm workers comprised less than two and one-half percent of 1966 trainees, yet governmental agencies 123Hathaway, "Migration From Agriculture:...," 124Schultz, "Public Approaches to Minimize Poverty,“... 125Bishop, "Increasing Mobility of LabOr...," 126Tweeten, Rural Poverty..., p. 49. estimated that nearly twenty percent of all MDTA trainees from 1963 to 1968 were rural residents. The difficulty was in the lack of adequate data regarding place of residence of trainees. But even the twenty percent estimate revealed that rural workers were highly under- represented in retraining programs in view of the fact that nearly half of the nation's poor resided in rural areas. Conclusion It was concluded, on the basis of the extensive adjustments in resource use that are taking place in United States' agriculture and the relatively low rate of participation of rural workers in Federal retraining programs, that there is consider— able need for increased enrollment of rural workers in retraining programs. Individual Retraining Needs of RLS Low Income Respondents (a) It was hypothesized that RLS poverty level respondents possessed levels of education significantly lower than those levels possessed by persons of the same age in the general population of the region. Survey data indicated that forty-one percent of the respondents below 1.50 income ratio had educational levels below the median, forty-eight percent equalled the median, and only eight percent exceeded thermedian (compared to roughly twenty-five, fifty, and twenty—five percent, 178 respectively, for the general population). The survey poor differed very significantly from the non-poor on this variable as well, and an even larger percentage of males than of females were classified below the educational median. (b) It was hypothesized that worker members in poverty level households possessed occupational skill levels signi- ficantly below those possessed by workers in the region's non—poverty households, and below the minimum skill required for entry into jobs for which low income workers were being retrained. The survey respondents were classified on the basis of potential economic viability. All respondents who were below age 65 and not physically or mentally handicapped were considered to haVe potential viability. Sixty—three percent of the respondents with income ratios below 1.50 were classified as potentially economically viable. The data in Table 24 indicate that these significant differences in skill levels related to median required skill 1eVels did exist for potentially economically viable respondents, both male and female. Ninety—two percent of the low income males and eighty-six percent of the females had skill levels below the established minimum required level. Among the non—poor, the corresponding rates were eighty— five percent for males and seventy—two percent for females. Conclusion It was concluded that there exists a very extensive need for job retraining among low income workers 179 in this subregion. 2. Retraining Potentials Based on Individual Character— istics Sufficient data on the socio-economic characteristics of participants in Federally sponsored retraining programs were not available for comparison on a number of variables with RLS respondents. Therefore, specific hypotheses regarding age, unemployment, and other characteristics could not be tested. The 1963 U.S. Department of Labor study of the extent of formal occupational training among the 60.8 million workers in the United States revealed that, for workers with less than three years of college, years of schooling completed and formal occupational training were closely related. Workers with more education had considerably more occupational training. On the basis of this relationship and the general educational development (GED) rating of job titles for which workers are being retrained under Federal manpower programs, RLS potentially viable low income respondents were classified on potential retrainability according to their educational levels. Respondents with GED levels 1 and 2, under five percent of the males and only three percent of the females, were considered to be retrainable only in programs which provided basic educational instruction as well as vocational training. Another thirty—seven per- cent of the males and thirty percent of the females had GED levels of 3 and were considered doubtfully retrainable -— 180 some would likely require dual education—training programs as well. There were still fifty-eight percent of the males and two—thirds of the females who were considered to have good potential for success in retraining programs, those with GED levels 4 and 5. Conclusion The conclusion reached from this analysis was that a substantial majority of the low income survey res— pondents who were in need of vocational training had the potential to successfully complete regular MDTA type occupational retraining programs. 3. Retraining Potentials Based On Attitudes Toward Retraining The potential ability of job retraining programs to lift rural consumer units from poverty levels depends ultimately upon the willingness of workers who need re— training to participate in the programs. Theoretically, a respondent would be interested in participating in a retraining program whenever the net discounted present value of the future income stream attributable to an improved skill level exceeded that value of the income stream he could reasonably expect to receive without retraining. Several hypotheses were formulated on the basis of this economic decison rule concerning respondents' interest in retraining. Most of the hypotheses relating interest in retraining with respondent characteristics were supported by the data. With only one exception, the non—support of hypotheses 181 resulted from the nature of the variables being tested, their being "hybrid" in that they were generated as a combination of two or more other variables. The only "normal variable" hypothesis not supported by the data was the one relating retraining interest with degree of poverty, as no significant differences appeared in this test. It appeared that uncontrolled variables washed-out expected differences. When all low income respondents were compared with the total survey population, the low income males showed considerably more interest in retraining than did all survey males. Altogether, nearly one—half of the low income males and one-third of the females expressed interest in participating in job retraining. Conclusion It was concluded that, as a group, the low income male respondents tended to react in an economically consistant manner toward potential retraining. The two hypothesis which were not supported by the 1 data stated that more interest in retraining would be I displayed by those respondents who were unemployed at the time of the survey interview and by those who were cal- culated to be available for alternative employment in the succeeding year. It appeared that the measure of unemploy— ment was the cause of rejection on the first of these hypotheses, as it included among the unemployed those not considering themselves to be in the labor market. This same methodological weakness existed with the calculated measure of availability, as again persons with no strong commitment to a current job and no family obligations were the ones determined available for alternative employ- ment. In both cases persons with no interest in becoming members of the work force also showed little interest in retraining programs and caused the hypotheses to be re- jected. All other hypotheses were supported by the survey data, and with statistical significance wherever statistical tests could appropriately be administered. Interest in retraining was greater among the younger age groups of respondents (except for the youngest,generally unmarried group), reflecting the longer expected time period for receiving increased earnings. The age—interest relation— ship was highly significant for males and significant for females. Although too few respondents reported any 1966 unemployment for statistical testing of the relationship, there was considerably greater interest in retraining demoustrated by respondents who experienced more than four weeks of unemployment in 1966 than by those with no unemployment. The rates of retraining interest were ninety Percent for males and eighty-three percent for females with longer 1966 unemployment, compared to forty-four and thirty-five percent, respectively, for males and females with no 1966 unemployment recorded. Among respondents who had sought a job in the past three years, those whose major difficulty in finding employment resulted from some personal educational or skill level deficiency (about one—fourth of those who sought work) demonstrated somewhat more interest in retraining than those with other major difficulties and considerably greater interest than respondents who reported having had no difficulties finding employment. The overall rate of retraining interest was much higher among those who had sought employment than among the total subgroup. Conclusion Respondents reacted to the potential offer of partici— pation in a job retraining program consistantly with the decision rule in relation to the variables of age, 1966 unemployment experience, and recent experience in finding employment It was noted above that respondents didn't react as hypothesized concerning their calculated availability status. When they classified themselves on availability, however, the available respondents were highly signifi- cantly more interested in retraining than those not considering themselves available. More than one—third Of the males and over one-fourth of the females reported that they would be available for alternative employment. As a corollary to stated availability, respondents whose attitudes toward their present jobs were more favorable were significantly less interested in retraining than respondents whose attitudes toward present jobs were less favorable. Concerning various potential costs of-retraining programs, interested respondents quickly lost interest in retraining as potential cost increased. Only two— thirds of the interested males remained interested at a program cost of up to $100; only onehfifth continued to be interested when costs were set from $100 up to $500; and only six percent were interested in retraining which would cost $500 or more. The percentage interested declined even faster for females as potential retraining costs increased. As hypothesized, there was no significant relationship between respondents' education levels in relation to Degree educational norms and their retraining interest. of interest was spread uniformly across all educational levels. gpnclusion It was concluded that respondents reacted consistantly with interest in retraining as related to their own estimation of their availability for alternative emplOyment, their attitudes toward their cUrrent jobs, and potential retraining program costs. Survey data indicated that respondents were equally interested in retraining across all educational levels. Conclusions of the Study (Recapitulation) 1. Conclusion Regarding the Extent of Poverty The data on the extent of poverty level incomes in this survey led to the conclusion that poverty level incomes do exist among many of the rural, open country consumer units in the Eastern Corn Belt States. Conclusions Regarding Potential Income Gains From 2. Retraining On the basis of the investment in human capital (a) modification of marginal productivity theory, it was concluded that job retraining has the theoretical potential for lifting rural poor families out of poverty by raising workers' marginal productivities. The benefit-cost analysis approach to estimation (b) of retraining's potential for lifting families from poverty provides an adequate methodological framework for these estimates. Estimates indicated that a large proportion of the rural, low income consumer units in this region could theoretically be lifted from poverty through job retraining. Conclusions Regarding Potentials for Retraining 3. Success There is considerable need for increased enroll- (a) ment of rural workers in retraining programs. (b) A very extensive need for job retraining exists among low income workers in this subregion. (c) A substantial majority of the low income survey respondents who were in need of vocational 186 training had the potential to successfully complete regular MDTA type occupational retraining programs. (d) As a group, the low income male respondents tended to react in an economically consistant manner toward potential retraining. (e) Respondents reacted to the potential offer of participation in a job retraining program in a consistant manner in relation to their age, 1966 unemployment experience, and recent difficulties in finding employment. 1 i (f) Respondents reacted with economic consistancy with interest in retraining as related to ‘ their own estimation of their availability for alternative employment, their attitudes toward their current jobs, and potential retraining costs. (g) Respondents were equally interested in re- training across all educational levels. F. Potentials for Escape From Poverty Through Retraining — Empirical Estimates Job training cannot provide an escape from poverty for all low income families in the Eastern Corn Belt. There are numerous, interrelated causes of poverty. Where low worker marginal productivity is the primary causal factor, then increases in marginal productivity through job retraining can reasonably be expected to correct the poverty status of 187 a large percentage of the consumer units. When low worker marginal productivity is only one of several causal factors, or when other factors predominate in causing poverty level incomes, then other solutions to the poverty problem must be found. Those respondents whose poverty was related to ad— vanced age or physical disabilities were considered in this study to have no potential for retraining. Respondents whose poverty was caused by low motivation and aspirations effectively excluded themselves from this analysis of retraining potentials by expressing no interest in retraining. other percentages of respondents were eliminated from the estimates of retraining potentials by the additional adjustments summarized below. The nature of the data utilized in making several of the estimatesincorporatedin this study prevents the drawing of definite predictions of retraining potentials. The following empirical estimates are believed, however, i to present reasonable approximations of the role that job retraining can be expected to play in eliminating rural poverty in the Eastern Corn Belt States. They are based on the several measures developed in this study and summarized above, and on the survey finding that three—fourths of the low income respondents lived within commuting distance (approximately forty miles) of a city of 25,000 or more pOpulation. The estimates of the percentages of the potentially economically viable low income respondents who 188 would be lifted from poverty through job retraining are as follows: Percentage who: Males Females Source of Estimate a. Need Retraining 92% 86% (Chapter VI, Section D) b. Have Retraining 88% 84% (Chapter VI, Section H) Potential o. Are Interested 41% 30% (Chapter VII, Section in Retraining B,2) d. Live within 31% 22% (above paragraph) Commuting Distance e. Can Be Expected 26% 19% (Chapter V, Section C) to Find Employment f. Earn CU Income 20% 14% (Chapter V, Section Above Poverty Level D,3) These estimates reveal that approximately twenty per- cent of the potentially economically viable consumer units with income ratios below 1.50 could be expected to be lifted from poverty through job retraining. These are considered to be relatively conservative estimates for several reasons. First, the assumption that the twenty—five percent of the respondents not living within reasonable commuting distance of urban employment would not be retrainable probably overstates this problem. People do find jobs utilizing their occupational skills in smaller cities as well. Sec0ndly, the conversion from individual retraining and employment potentials to consumer unit income is necessarily arbitrary. If it is assumed that only one worker per consumer unit is retrained, then the twenty percent estimate for retrained males may reasonably be applied to consumer units. 189 But if two or more consumer unit members are retrained, then as many as one—fourth of the consumer units could be lifted above poverty levels. Finally, with adequate» job placement efforts on the part of retraining program administrators, employment rates could be expected to exceed the 85 percent rate of past MDTA graduates. The measure of interest in retraining probably tends to overstate the actual retraining potentials. It is generally recognized that response to such a hypothetical question as the one asked concerning interest in retraining will be considerably greater than actual participation in a training program. However, more interest would undoubtedly have been shown if the hypothetical training program had included the subsistance allowance usually incorporated in MDTA programs. On balance, it was decided to make no arbitrary adjustments in the degree of interest expressed. It was implicitly assumed that all interested respondents would participate in retraining programs. G. Conclusion The-above estimates indicate that job retraining programs can reasonably be expected to lift at least one— fifth of the potentially viable poverty level consumer units above the poverty line in this subregion. While this would still leave a major portion of the rural poverty problem unsolved, any single program that can reasonably be expected to solve twenty percent of this critical problem must receive major emphasis in anti-poverty programs in rural areas of the East North Central States. .n‘. v ,w- .- "pl \ .. -._.‘ ...... .....pf‘ 46% IX . IMPLICATIONS A. Introduction There have been two significant assumptions implicit in this analysis. The estimates of labor market performance of potentially retrained rural workers were made with the assumption that labor market conditions would remain as they were in the 1965-1967 period. Should future unemploy- ment rates differ significantly from those in that period, then placement and earnings estimates for rural retrainees would also need to be revised.“ By including no estimate of non-completiOn rates for training enrollees, the study contained the assumption that all who enrolled would complete training. Such has not been the actual experience in MDTA programs. As noted in Chapter V, however, many drOp out of training courses to accept employment. For this reason, no adjustments in estimates were made for non-completion of retraining. Perhaps the most critical implications of this study are those concerning the role that job retraining can not play in eliminating rural poverty in this subregion. The thirty-seven percent of the low income individuals con- sidered to have no potential for economic viability can only be removed from the poverty category by social welfare oriented, as Opposed to work oriented, programs. They will, for the most part, require some form of income supplements. 190 191 Among the potentially economically viable respondents, approximately fifty percent were found to be in need of job retraining but expressed no interest in participating in retraining programs. While many of these were wives of men who would be retrained, half of the males also expressed no interest in retraining. This segment of the rural poor presents an urgent challenge to anti-poverty program administrators to develop innovative approaches to solving its poverty problems. Finally, innovative types of delivery system need to be developed to provide retraining and employment opportunities for those low income individuals who are not within commuting distance of employment centers. B. Implications for Present Training Programs The extensive poverty level incomes and needs for retraining among RLS respondents are clear indications that ) the present rural system of vocational training (where one exists) in the Eastern Corn Belt States has left a consider— able percentage of the rural workers ill—prepared to compete in today's labor force. Only four percent of the survey POOr, and only eight percent of the entire survey population, had received any post high school technical training within the past ten years. Elsewhere, the U.S. Department of Labor 1963 study of vocational training in the total United States labor force revealed that sixteen percent of the workers with less than nine years of schooling completed and nearly fifty percent of those with 9—11 years of schooling had received some formal occupational training (see Table 26). Although the data 192 are not directly comparable, they do serve to illustrate the apparent lack of vocational training among the survey respondents. Although the limited data available on the place of residence of enrollees in Federal retraining programs was not broken down by geographical regions, even the liberal estimate that twenty percent of MDTA participants were rural residents indicates that these Federal programs have not concentrated their efforts where the highest rates of unemployment (and underemployment) and poverty exist. Inasmuch as the total MDTA enrollment of 599,000 through 1966127 represented the equiValent of only three and one— half percent of the 1963 labor force with less than nine years of education, these Federal programs could not have been expected to make a major impact on rural training deficiencies. But they could have indicated, by enrolling and training more rural trainees, that they were focusing their efforts where the need was greatest. The extent of interest in retraining programs expressed by both poor and non-poor survey respondents indicated that rural vocational training programs have not adequately functioned in this subregion. 127U.S. Department of Labor, Manpower Report of the President and A Report on Manpower Requirements, Resources, Utilization, and Training, (Washington: U.S. Government Printing Office), April, 1967, p. 277, Table F—l. C. Implications for Redirection of Rural Retraining Efforts 1. Potentials for Escape From Poverty Through Retraining Manpower development programs have been a significant part of our overall national manpower program since 1962. Yet it was not possible to cite for this study any estimates of the potential that job retraining has for raising low income consumer units above the poverty line. The few “effici- ency" studies of retraining programs that have been completed to date have been concerned with societal costs and benefits, with comparisons of alternative retraining programs in terms of these costs and benefits, or with increases in employment or hourly earnings for trainees. None of the evaluations have addressed themselves to measuring retraining's potential for lifting people from poverty, ostensibly because the manpower development programs have not been directed toward that end themselves. It was noted in Chapter V, Section C, that the objectives of the MDTA are: (l) to increase the nation's output; (2) to reduce the aggregate level of unemployment; (3) to reduce the governmental costs of unemployment; and (4) to reduce the burdens of unemployment for specific groups of the unemployed. With the emphasis on national output and on unemployment it was not surprising that program evaluations focused on employment and cost factors. It appears that retraining evaluations have not asked the appropriate questions, therefore we do not have appro- Priate answers about potentials for escape from poverty. 194 And the appropriate questions have not been asked because the retraining programs have pursued objectives other than lifting poverty level families from poverty. The implication of this is that it is time to make the lifting of families from poverty the primary objective of manpower programs. It does seemingly little good to reduce unemployment by training a worker for employment in such a low skilled, low paying job that his family continues to receive a poverty level income, as some training programs apparently have done. There should be no question of the ability of job retraining to lift families from poverty — a worker should not be considered sufficiently trained until he can earn an above poverty income, providing he has the mental and physical capacity for such training. The importance of retraining rural workers, particularly farm workers, for employment which provides adequate incomes was illustrated by Hathaway and Perkins.128 They found that the initial income changes experienced when moving from farm to non—farm employment were an important determinant of whether an individual remains in non-farm employment. With adequate retraining, the low income rural worker (farm or non—farm) would be in a position to experience initial income changes sufficient to keep him 128Hathaway and Perkins, "Occupational Mobility and Migration...," p. 205. 195 in the alternative employment and keep him earning above poverty income. 2. Retraining for What? Inasmuch as one-third of the survey consumer units were classified as farm units, and a number of the non— farm units also received some "farm income, what are the potentials for assisting respondents to become viable farmers through agricultural training programs? A farming potential, based on the possession of some amount of farm acreage and the extent to which that acreage provided income in 1966, was developed for RLS respondents. The data on acreage potentials for respondents below 1.00 income ratio and for respondents below 1.50 income ratio appear in Table 41. While over thirty percent of the poor individuals (by either measure) had some acreage potential, half of those with acreage were already over age 64 in 1967. Furthermore, of those under age 65, sixty percent of the group below 1.00 income ratio and two—thirds of the group below 1.50 income ratio either earned no income from their land or these earnings were less than half of their total earnings in 1966. Of the respondents under age 65 with some acreage potential, only one-third with income ratios below 1.50 and forty—one percent with income ratios below 1.00 received all or a major share of their 1966 earnings from their acreage. Respondents with major earning from acreage represented only seven percent of the below 1.00 income 196 Table 41. Potentials for Viability in Farming Based on Acreage and Its Use, 1966, for Poor Respondents; Percent Distributions 1967 Rural Life Survey Individuals Income Ratio Classification Below Below 1.00 1.50 Percent of Total Population 12.8 26.3 Percent of Poor with Acreage Potential 34.3 30.9 Percent of Those with Potential Over Age 64 52.5 48.0 Those Under 65 with Acreage 1. No 1966 Income From Acres 18.1 24.3 2. 1966 Income Not Farm "Earnings" 17.3 19.1 3. Farm Income Minor Share of Earnings 23.2 22.9 Total 1-3 58.6 66.3 4. Acreage Gave Sole or Major Share of 41.4 33.7 Earnings 5. Percent with "Major Share" over age 44 74.7 75.3 Percent of Poor Under Age 45 With Acreage 6.7 5.4 Potential ratio individuals, and only five percent of those below the 1.50 income ratio. Most with acreage potential were consumer unit heads, and three—fourths of them were already over age 44 in 1967. The implication of these acreage potentials is clear. Even in this region of commercial agriculture, only a very small percentage of the rural, open country low income population has any potential for becoming viable commercial farmers through job retraining. Many with acreage had already recognized this and had turned to non-farm employment for 197 part of their income. Thus, retraining for non-farm employment would, in most instances, facilitate a process which was already well under way in 1966. 3. Implications of Retraining Needs The measure of retraining needs developed in this study was not intended to be a definitive measure, because the measure of respondents'skill levels included only formal occupational training received in the past ten years. As a rough approximation of retraining needs based on skill levels being below a required skill level, however, the measure did permit indicative estimates of retraining needs. This study found that approximately eighteen percent of all potentially economically viable respondents had need for retraining and had income ratios below 1.50.129 With a 0.6 percent sampling rate, the poor needing retraining would total some 200,000 persons. If the non-poor with apparent retraining needs were included, the total would be considerably larger, possibly from one—half to three— fourths of a million individuals. The apparent scale of retraining needs among the rural population in this subregion implies that retraining programs will have to be much more extensive in the future than they have been to date if they realistically plan to retrain all needy individuals. The need for retraining the estimated 200,000 poor workers in the region is an immediate need. Once the relatively short range programs for retraining these workers ____________________ 129Figures derived from data in Table 24. ) E z are completed, longer range training needs could be concentrated upon. In the long-run, after present skill deficiencies are corrected, there will be continued need to train new entrants to the labor force and to provide upgrading to some percentage of the previously retrained workers. Both short-run and long—run training needs require attention. 4. Implication of Retraining Potentials On the basis of retraining potentials developed in this study, retraining program administrators can be assured that sufficient numbers of rural, poverty level workers are interested in participating in, and have good potentials for completing, regular MDTA type retraining programs. On the other hand, if retraining is going to give first priority to the most needy of the rural poor, then much more effort will have to be exerted in providing basic educational instruction, as well as vocational training. Although this study focused only on educational level as a measure of potential for retraining success, past experience with retraining "hardcore" individuals indicates that a complete package of services — counseling, health, placement, follow-up - are also required to insure their successful retraining. The same would undoubtedly apply to the rural poor with the least potential for success in retraining as measured here. 5. Implications of Expressed Interest in Retraining Survey respondents expressed interest in a "free training Program given locally" which would enable them to obtain a 199 better job. It is probable that even more than forty— seven percent of the males and thirty—six percent of the females would be interested in a program which was not only free but which provided a subsistance allowance while in training as well. Conversely, there would no doubt be less interest in retraining courses which were not given locally. The data in Table 34 indicated that a large percentage of the respondents interested in retraining were employed at the time of the interview. It would be essential to provide a subsistance allowance before most of these ; respondents could actually participate in retraining. Interest in retraining was found to be relatively strong across all age groups. In providing retraining programs for members of the older age groups, administrators would have to consider their unique training and employment problems. The data in Table 10 revealed that, as would be expected, poverty was closely associated with employment. But employ- / ment rates among poor males were the same as employment rates / among all non—poor respondents. One of the major differences, / then, between poor and non—poor consumer units was the employment rate among females. This implies that retraining programs should concentrate upon retraining females and raise families above the poverty line by increasing the number of wage earners rather than by raising the skill level of the male consumer unit head. 200 6. Implications for Trainee Recruitment Under the MDTA manpower training programs the primary means of recruiting candidates for retraining programs has been through referrals from the state employment service offices. The survey data indicated that the method of publicizing programs and recruiting candidates would probably have to be altered considerably in rural areas of this subregion. As shown in Table 42, even though they were receiving poverty—level incomes, only eight percent of the low income consumer units reported that a CU member visited the employment office in the 9 past year. It is clear that none of three other public service agencies, the Cooperative Extension Service, the Social Security Administration, and county welfare offices, would be better prepared to reach those in need of training than the employment office. Data in Table 42 indicate that these other agencies had even less frequent contact with the rural poor than did the employment service. The major implication to be drawn from this is that ! totally new institutional arrangements may be required for / recruitment of rural retrainees. Alternatively, the present ( recruiting agencies will have to adopt revolutionary new techniques of trainee recruitment. D- Conclusion The major conclusion to be drawn from these implications for rural retraining programs, and perhaps from the entire StUGY. is that there must be a continuing emphasis on 201 Table 42. Use of Public Services by Low Income Consumer Units; Percent With Visits, by Income Ratio 1967 Rural Life Survey Consumer Units Below 1.50 Income Ratio Public Service Visitied Income Ratio Cooperative Social Security County Employment Extension Administration Welfare Office .80 10.8 11.3 9.5 5.4 .80- .99 6.7 8.6 4.8 15.2 1.00— 1.19 2.7 7.5 9.5 8.8 1.20- 1.49 "5.0 3.2 5.3 7.8 Total 6.5 7.2 I 7.3 I 8.3 training and retraining rural residents, low income and non—low income alike, for competitive employment in today's labor market. Regardless of the extent to which rural-urban distinctions are said to no longer be valid in the United States, it is in the rural areas that the people are "left behind," and it is into the rural areas that new institut— ional arrangements for recruiting and retraining potentially viable, poverty level workers will have to be extended. BIBLIOGRAPHY BIBLIOGRAPHY BOOKS Batchelder, Alan B. The Economics of Poverty. New York: John Wiley & Sons, Inc., 1966. Gingberg, Eli. The Development of Human Resources. New York: McGraw Hill, 1966. Harbison, Frederick, and Myers, Charles A. Education, Mappower, and Economic Growth: Strategies of Human Resource Development. New York: McGraw—Hill, 1964. Lester, Richard A. Economics of Labor. New York: The Macmillan Company, 1964. Mangum, Garth L. MDTA: Foundation of Federal Manpower Policy. Baltimore: The Johns Hopkins Press, 1968. Morgan, Chester A. Labor Economics. Homewood, Illinois: The Dorsey Press, 1966. Ribich, Thomas I. Education and Poverty. Washington: The Brookings Institution, 1968. Somers, Gerald G. (ed.). Retraining the Unemployed. Madison: The University of Wisconsin Press, 1968. Steel, Robert G. D. and Torrie, James H. Principles and Procedures of Statistics. New York: McGraw Hill, 1960. Vaizey, John. The Economics of Education. London: Faber & Faber, Ltd., 1962. Wolfbein, Seymour L. Education and Training for Full Employment. New York: Columbia Univers1ty Press, 1967. 202 ARTICLES IN EDITED BOOKS Bishop, C. E. "Increasing Mobility of Labor Through Training Programs," in Hildreth, R.J. (ed.). Readings in Agricultural Policy. Lincoln: University of Nebraska Press, 1968. Bluestone, Barry. "Lower-Income Workers and Marginal Industries," in Ferman, Louis A., Kornbluh, Joyce L., and Haber, Alan (eds.). Poverty in America. Ann Arbor: University of Michigan Press, 1968. Bonnen, James T. "The Distribution of Benefits from Selected U.S. Farm Programs," in National Advisory Commission on Rural Poverty. Rural Poverty in the United States. Washington: U.S. Government Printing Office, 1968. Borus, Michael E. "Time Trends in the Benefits from Retraining in Connecticut," in Twentieth Annual Proceedings of the Industrial Relations Research Association. Madison: University of Wisconsin Press, 1968. Chesler, Herbert A. "The Retraining Decision in Massachusetts: Theory and Practice," in Somers, Gerald G. (ed.). Retraining the Unemployed. Madison: The University of Wisconsin Press, 1968. Hathaway, Dale E. and Perkins, Brian E. "Occupational Mobility and Migration from Agriculture," in National Advisory Commission on Rural Poverty, Rural Poverty in the United States. Washington: U.S. Government Printing Office, 1968. Minsky, Hyman P. "Adequate Aggregate Demand and the Commitment to End Poverty," in National Advisory Commission on Rural Poverty. Rural Poverty in the. United States. Washington: U.S. Government Printing Office, 1968. Schuh, G. Edward. "Interrelations Between the Farm Labor Force and Changes in the Total Economy," in National Advisory Commission on Rural Poverty. Rural Poverty in the United States. Washington: U.S. Government Printing Office, 1968. Schultz, T. W. "Public Approaches to Minimize Poverty," in Fishman, Leo (ed.). Poverty Amid Affluence. New Haven: Yale University Press, 1966. 203 Somers, Gerald G. "Our Experience With Retraining and Relocations," in Gordon, R.A. (ed.). Toward A Manpower Policy. New York: Wiley, 1967. Somers, Gerald G. and Stromsdorfer, Ernst W. "A Benefit Cost Analysis of Retraining," in Seventeenth Annual Proceedings of the Industrial Relations Research Association. Madison: University of Wisconsin Press, 1964. Weber, Arnold R. "Experiments in Retraining: A Comparative Study," in Somers, Gerald G. (ed.). Retraining the Unemployed. Madison: The University of WlSCOnSln PEEEET‘ISEB. ARTICLES, REPORTS AND UNPUBLISHED MATERIALS Bird, Alan R. Poverty in Rural Areas of the United States; Agricultural Economic Report No. 63. Washington: U.S. Department of Agriculture, Economic Research Service, November, 1964. Bird, Alan R. and McCoy, John L. White Americans in Rural Poverty;Agricultural Economic Report No. 124. Washington: U.S. Department of Agriculture, Economic Research SerVice, November, 1967. Bonnen, James T. "Rural Poverty: Programs and Problems," Journal of Farm Economics, Vol. 48, No. 2 (May, 1966), pp. 452-465. Booth, E. J. R. "The Economic Dimensions of Rural Poverty," American Journal of Agricultural Economics, Vol. 51, No. 2 (May, 1969), pp. 428-442. BOrus, Michael P. "A Benefit-Cost Analysis of the Economic Effectiveness of Retraining the Unemployed, Yale Economic Essays, Vol. 4, No. 2 (Fall, 1964), pp. 371—429. Borus, Michael E. "The Cost of Retraining the Hard—Core Unemployed," Labor Law Journal, Vol. 16, No. 9 (September, 1965), pp. 574—583. "Rural Poverty in the United States," Clawson, Marion. 49’ No. 5 (December, 1967), Journal of Farm Economics, Vol. pp. 1227-1233. ' ' ' ll "Economic Criteria for Education and Training, Eckaus, R. S- Vol. 46, No. 2 Review of Economics and Statistics, (May, 1964), pp. 181—190. 205 Fine, Sidney A. "Use of the Dictionary of Occupational Titles to Estimate Educational Investment," The I Journal of Human ResourceS, VQIIXIII, No. 3 _-— (Summer, 1968), pp. 363—375. Gallaway, Lowell E. "The Effect of Geographic Labor Mobility on Income: A Brief Comment," Journal of Human Resources, Vol.‘4, No. 1 (Winter, I969), pp. 103- 09. Gallaway, Lowell E. "Mobility of Hired Agricultural Labor," Journal of Farm Economics, Vol. 49, No. 1 (February, 1967), pp. 32-52. Hathaway, Dale E. "Migration From Agriculture: The Historical Record and Its Meaning," American Economic Review, Vol. 50, No. 2 (May, 1960), pp. 379-391. Jessen, R. J. "The Master Sample Project and its Use in Agricultural Economics," Journal of Farm Economics, Vol. 29, No. 2 (May, 1947), pp. 531—540. Johnson, Harry G. "The Economic Approach to Social Questions," The Public Interest, No. 12 (Summer, 1968), pp. 68—79. Konyha, Marvin E. "Educational Medians in the East North Central States, 1967." East Lansing: Michigan State University, 1969 (mimeo). Lansing, Johan. and Morgan, James N. "The Effects of Geographic Mobility on Income," Journal of Human Resources, Vol. 2, No. 4 (Fall, 1967), PP. 449—460. Lanzilotti, Robert F. "The Poverty Syndrome: A Critical Review," Michigan State UniverSity BuSiness Topics (Spring, 1966), PP. 39—49. Larson, Olaf F. "Discussion: Rural Poverty in the United States," Journal of Farm Economics, Vol. 49, No. 5 (December, 1967), pp. 1234—1236. Lesgold, Alan M. "Analysis of Contingency Tables: Act II," Technical Report No. 14. East LanSing: Michigan State University, Computer Institute for Soc1al Science Research, January, 1968. Federal Manpower Policies and Programs to Kalamazoo: The W. E. Upggin Levitan, Sar A. Combat Unemployment. Institute for Employment Research, February, Main, Earl D. "A Nationwide Evaluation of M.D.T.A. Institut— ional Job Training," Journal of Human Resources, Vol. 3, No. 2 (Spring, 1968), pp. 159-170. 206 Main, Earl D. A Nationwide Evaluation of M.D.T.A. Institut- ional Job Training Programs. Chicago: National opinion Research Center, University of Chicago, October, 1966. Martin, Lee. "Effects of Alternative Federal Policies on Welfare of Rural People," Journal of Farm Economics, Vol. 48, No. 5 (December, 1966), pp. 1267—1276. McCauley, John S. "Manpower Development in Rural Areas," Employment Service Review, Vol. 5, Nos. 3 & 4 (March—April, 1968), pp. 10—15 ff. Nosow, Sigmund. "Retraining Under the Manpower Development and Training Act: A Study of Attributes of Trainees Associated With Successful Retraining." East Lansing: Michigan State University, School of Labor and Industrial Relations, January, 1968 (mimeo). Orshansky, Mollie. "Counting the Poor: Another Look at the Poverty Profile," Social Security Bulletin, Vol. 28, No. 1 (January, 1965), pp. 3—29. Page, David A. "Retraining Under the Manpower Development Act: A Cost—Benefit Analysis," Public Policy, Vol. 13 (1964), pp. 258—267. Perkins, Brian E. and Hathaway, Dale E. The Movement of Labor Between Farm and Nonfarm Jobs; Research Bulletin No. 13. East Lansing: Michigan State University, Agricultural Experiment Station, 1966. Planning Research Corporation. Cost Effectiveness Analysis of On—the—Job and Institutional Training Courses; A Report to the U.S. Department of Labor, Office of Manpower Policy, Evaluation, and Research. Washington: By the Corporation, 1967. Ruttan, Vernon W. "Agricultural Policy in an Affluent Society," Journal of Farm Economics, Vol. 48, No. 5 (December, 1966), pp. 1100—1120. Schultz, Theodore W. "Investment in Human Capital,“ The American Economic Review, Vol. 51, No. 1 (March, 1961), pp. 1—17. Stromsdorfer, Ernst W. "Determinants of Economic Success in Retraining the Unemployed: The West Virginia Experience,“ Journal of Human Resources, Vol. 3, No. 2 (Spring, 1968), pp. 139—168. Thurow, Lester C. "The Causes of Poverty," Quarterly Journal of Economics, Vol. 81, No. 1 (February, 1967), pp. 39—57. 207 Trooboff, Benjamin M. Employment Experience After MDTA Training: A Study of the RelationshipBetween Selected Trainee Characteristics and Post-Trainipg Experience. Atlanta: Georgia State ColIege, School of Business Administration, July, 1968. Tweeten, Luther G. The Role of Education in Alleviating Rural Poverty; Agricultural Economics Report No. 114. Washington: U.S. Department of Agriculture, Economic Research Service, June, 1967. Tweeten, Luther G. Rural Povertyil_Incidence, Causes, and Cures, Processed Series P-590R. Stillwater: Oklahoma State University, Experiment Station, July, 1968. Venkareddy, Chennareddy. "Present Values of Expected Future Income Streams and Their Relevance to Mobility of Farm Workers to the Nonfarm Sector in the United States, 1917-62." Unpublished doctoral dissertation, Michigan State University, 1965. Venkareddy, Chennareddy, and Johnson, Glenn L. "Projections of Age Distributions of Farm Operators in the United States Based upon Estimates of the Present Value of Income," American Journal of Agricultural Economics, Vol. 50, No. 3 (August, 1968), pp. 606-620. PUBLIC DOCUMENTS Miller, Herman P. and Hornseth, Richard A. Present Values of Estimated Lifetime Earnings; Technical Paper No. 16. U.S. Bureau of the Census. Washington: U.S. Government Printing Office, 1967. National Advisory Commission on Rural Poverty. The People LenyBehind. Washington: U.S. Government Printing Office, 1967. National Advisory Commission on Rural Poverty. Rural Poverty in the United States. Washington: U.S. Government Printing Office, 1968. .Sheppard, Harold L. "A Search for New Directions in the War Against Poverty," (Appendix Paper) in Subcommittee on Employment, Manpower, and Poverty of the Committee on Labor and Public Welfare, U.S. Senate (90th Congress, 2nd Session), Toward Economic Securipy for the Poor. Washington: U.S. Government Printing Office, October, 1968. U. S. Bureau of the Census. U.S. Census of Population: 208 1960. Detailed CharacteristiEs. United states Summary. Final Report PC(l)-Ip. Washington: U. S. Government Printing Office, 1963. Department of Agriculture, Economic Research Service. Editing Manual: Rural Life Survey. Washington, 1966 (mimeoil Department of Health, Education and Welfare. Education and Training: A Chance to Advance (Seventh Annual Report of the Department to the Congress on Training Activities Under the MDTA). Washington: U.S. Govern— ment Printing Office, April, 1969. Department of Labor. Dictionary of Occupational Titles, Vol. I, Definitions of Titles; Vol. II, Occupational Classification, 3rd. ed. Washington: U.S. Government Printing Office, 1965. Department of Labor. Manpower Report of the President and Report on Manpower Requirements, Resources, Utilization, and Training. Washington: U.S. Government Printing Office, April, 1967; April, 1968; January, 1969. Department of Labor. Report of the Secretary of Labor on Manpower Research and Training_Under the MDTA.i Washington: U.S. Government Printifig Office, 1966. Department of Labor. Selected Characteristics of ”Occupations (Physical Demands, Working Conditions, Training Time) 1966 - A Supplement to the Dictionary of Occupational Titles, 3rd. ed. Washington: U.S. Government Printing Office, 1966. Department of Labor. Selected Characteristics of Occupations by Worker Traits and Physical Strength, 1968 (§ppplement 2 to the Dictionary of Occupational Titles, 3rd. ed.). Washington: U. S. Government Printing Office: 1968. Department of Labor, Manpower Administration. Formal ‘Occupational Training of Adult Workers: Manpower Automation Research Monograph No. 2. Washington: By the Department, December, 1964. Department of Labor, Manpower Administration. The Influence of MDTA Training on Earnings, Manpowei__ Evaluation Report No. 8. Washington: By the Department, December, 1968. U.S. 209 Department of Labor, Manpower Administration. Register of ProjectspApproved Under the Manpower Development and Training Act Through June, 1965. Washington: by the Department, 1965. Senate. Employment and Training Legislation — 1968 Background Information Supplement. Committee Print, 90th Congress, 2nd Session. Washington: U.S. Government Printing Office, June, 1968. APPENDICES APPENDIX A DEVELOPMENT OF BASIC TRAINING REQUIREMENTS AND RESPONDENTS" SVP AND GED LEVELS Basic training requirements for low income individuals were based upon a selected list of occupational titles and the SVP levels of those titles as presented in the Dictionary of Occupational Titles. A. Occupational Titles Used The job-titles which were used to determine the basic training requirements for poverty—level workers were those jobs for which unemployed and underemployed workers have been retrained under Federal programs authorized by the Manpower Development and Training Act (MDTA) of 1962 (as ammended). A listing of all MDTA projects approved through June, 1965, was the source used to obtain the list of job titles, a list which totaled over 900 titles.130 Specifically, there were 917 titles, 393 of which were institutional courses and 524 of which were OJT programs (157 occupations appeared in both type programs). Seventy—five of the occupations were identified as predominantly female occupations. For each occupation or job-title listed, the DOT code number was found in Volume II of the Dictionary of Occupational Titles, then the corresponding GED and SVP . 130U.S. Department of Labor, Manpower Administration, Register of Projects Approved Under the MDTA Through June, 1965 (Washington: by the Department, 1965). 211 were obtained from the 1966 DOT Supplement. In order to evaluate GED and SVP "requirements" on the basis of total trainees in each occupation, the following sampling and weighting procedures were applied. It was noted that trainee enrollments in approved institutional courses were considerably larger than those in OJT courses, so it was decided to weight institutional courses by the ratio of institutional to OJT average enrollments. A random sample of ten states was selected, using a random number table,131 and average enrollments in these states were computed (Table A—l). It should be i i noted that these averages, 43.47 for institutional programs ‘ and 19.4 for OJT programs, differ considerably from averages which would result if the totals for enrollments and projects presented on pages 3-4 of the Register of Projects Approved were used to compute averages. This results because, in computing these averages in the ten-state sample, all multi—occupation programs, those programs whose file was incomplete and did not give the occupational title, those providing only pre—vocational training, and those whose total enrollment was not given were all eliminated.' The ratio of institutional to OJT average enrollments in occupational courses which could be identified, then,was 131Robert G. D. Steel and James H. Torrie, Principles . and Procedures of Statistics (New York: McGraw Hill, 1960), ‘ Appendix Table A.l, pp. 428—431. 212 Table A—1. Number of MDTA Projects Approved and Number of Trainees Enrolled in Selected States, by Program Type, Through June, 1965 No. Inst No. Inst No. OJT No. OJT State Projects Trainees Projggts Trainees 3 Arizona 48 2,739 19 97 4 Arkansas 69 1,724 43 218 5 California 463 22,360 42 2,263 15 Illinois 237 13,893 22 1,824 16 Indiana 102 4,659 19 62 17 Iowa 76 2,652 5 212 32 New Jersey 191 7,405 97 1,130 38 Oklahoma. 49 1,635 4 162 40 Pennsylvania 309 10,547 95 1,044 44 South Dakota 20 379 28 252 Total 1,564 67,993 374 7,264 Average 43.47 19.4 'Source: U.S. Department of Labor, Manpower Adminis— tration, Rggister of Projects Approved Under the MDTA Through June, 1965 (Washington: by the Department, 1965) pp. 3-4. 43.47/19.4 or 2.24. The number of institutional courses approved should be multiplied by this factor to reflect the greater importance of institutional as compared to OJT programs in terms of total enrollments. However, to compensate partially for the fact that OJT programs have been given more emphasis in recent years than data through 1965 would indicate, the weight factor was rounded to 2.0. A second weight factor was utilized to account for the greater frequency of some programs. Using the ten randomly sampled states, a frequency count of all occupational programs was made, with a weight factor of one given each 2133 time the occupational category was listed. For those programs which were not given in any one of the ten sample states, it was assumed that they were listed no more than five times in the non-sampled states and they were also given a weight of one. B. SVP Median Levels The weighted number of job titles with various SVP scales are given, by sex, in Table A—2. The job titles are also classified in Table A—2 according to their relationship to the median SVP, this being defined as the SVP level at which the median of the total (weighted) number of job titles is found. For male job titles the median falls at SVP = 6 and for female categories the median falls at SVP = 4. A "median range" was established for each sex category, with the range for males including one SVP level above and one below the median SVP, and for females this range included only the median SVP level and the one above it. This "median range" for males comprised 70.5 percent of all job titles; for females it contained 83 percent of all titles. Upon examination of the over 900 job—titles given in the Register of Projects Approved, it was found that the occupations could be roughly broken down into four broad categories according to SVP scales as follows: SVP l and 2 = unskilled occupations SVP 3 and 4 = semi—skilled SVP 5 and 6 = skilled SVP 7 and 8 = highly skilled , 214 Table A-2. Weighted Number of MDTA Approved Retraining Courses by SVP Scales, by Sex Males Females SVP No. of Job' Relation to SVP No. of Job Relation to titles median titles median 1 4 l 0 2 158 2 34 < median (median 3 258 3 47 4 265 4* 612 i = "median 5 292 5 255 range" (867=82.9%) 6* 613 ="median range " 6 66 (2032=70.5%) } 7 1127 , 7 31 { -. >median 8 66 8 1 >median 9 l “ 9 0 *Median SVP Level Source: U.S. Department of Labor, Manpower Administration, Register of Projects Approved Under the MDTA Through June, 1965 (Washington: By the Department, 1965). Thus, the "median range" for male occupations consisted of either skilled or highly skilled occupations, and for females it included semi—skilled and skilled occupations. APPENDIX B DEVELOPMENT OF RETRAINING POTENTIALS For determining retraining potentials, the data from Appendix A has been tabulated as shown in Tabde B-l, with the weighted number of occupations listed by GED and corresponding SVP scales and broken down by male and female occupations. In addition, median SVP levels corresponding to each GED level, by sex, Were calculated as shown in Table B-2. Inasmuch as vocational abilities are focused upon, and presumably improved, when occupational retraining courses are given, respondents' GED levels were considered to be the independent variables and their SVP levels were treated as dependent variables. Thus, for a given GED, a respondent was compared to the median SVP corresponding to that GED level of MDTA retraining programs to determine if he had a potential for retraining based on his general level of educational development. For example, since the median SVP for male occupations which "require" a GED of 4 is an SVP of 7, a respondent whose GED was 4 and whose SVP also equalled 4 could have a potential for retraining to raise his SVP level from 4 to as high as 7. 215 ~ H H - - - - I - - - - - - - a Na.ns N km s - - - an N - n oN N - w neonanxm , snamnm ., Na map NNN - - NH mmn an - Ne arm men - N . m wHN NaN - - N GN on - H can HNH - G Ne.nm "wasnnxm N as NHN s - - NH as - H In 0N N m fl - ON aeN In - - OH am an - om ma SH 4 so.HN . ,o newnnnmeEmm A/ - s amn man - - N mm as - N mm om m l .2 NG.m - - an as Ms - - m cm - - HN an N u wannnxmas . - - - - a - - - - - - - - H m a m N H m N: m N m a m N mnw>an Anunso .Nu.nmqu canoe nos-wnu-ao nmaonnaunuman m>m Hanna an conusnnnumna new unmouom mwamz moma .wcsh fiwsousH pw>oumm< mwmusoo m>m tum Qmw twuswfimz Hum OHQMH .An - H - - - - H - - - - w NH.N . uuwHHme NHHmHm - NN - - - N H - NH - - N J - om 4H - - - N - 4N N - o NN.om , 15:me A - HH EN - - - H a m ONH - m. : . - N was m - - - mm H mwN H a No.mo namHHme-Haom - - we H - - - a - HN - m .mm L - - 4H 0N - - - N i o m N 2 NN.m ustHmeap - - - - - - - - - - - H m a m N H m a m a m N mH.053 AHuHso «Nu.um:Hv Hmnop Hoe-msu-ao HmmmHusanmaH ism HHme NH coauonHuumHQ emu uncouom moamfimm moma.o::h fiwsoufifi to>ouom< mongooo m>m tom emu touswfloz twscpucoo Hum GHQmH 218 Table B-2. Required GED Levels and the Corresponding Median SVP Levels of MDTA Training Courses Approved Through June, 1965, by Sex GED Median SVP Male l Female 1 2 - 2 3 2 3 5 4 4 7 6 5 7 7 6 _ _ Source: Data in Tables A-1 and A-2. APPENDIX C PRESENT LEVELS OF EDUCATION AND TRAINING Measures of the present levels of education and training possessed by poverty-level respondents were derived from the survey data, and these levels were then converted to equi- valent GED and SVP measures. A. General Educational Development Measures Several methodological problems in developing these measures were readily apparent. To determine accurately the present level of skills and abilities of respondents would require the administration of a complex battery of general aptitude and vocational aptitude tests. Even then, it is recognized that such tests have numerous weaknesses and that such verbal instruments cannot be a perfect measure of an individual's skill level and cannot, therefore, be used to make definite predictions concerning labor market performance. Numerous attitudinal and motivational factors enter into an individual's actual performance in the labor market. While social-psychologists have recently made much progress in developing measuring devices for these factors, there appears to be no instrument available as yet to greatly improve upon the predictive capacities of the general and vocational aptitude tests which, as noted, have only limited predictive ability. At any rate, the sample size in the RLS study precluded the administration of any such battery of tests. As an 219 220 alternative, the number of years of schooling completed by the respondents was used as a proxy for a measure of general aptitude. This is admittedly not a perfect substitute for three basic reasons. One is response bias. It is known that persons with lesser amounts of formal education tend to overstate their educational level when responding to survey questionnaires (RLS poverty-level respondents tend to have below average educational levels) and some report last grade attended d.132 The second reason even though it was not complete is that the quality of schooling differs widely between some school systems, and even many times within the same school system for different individuals. Thirdly, many individuals have received self-education well beyond that received from formal schooling either through individual efforts or as a part of specific job training. However, no more appropriate measure was available from the survey data so years of schooling were used as the best substitute available for a precise measure of respondents' general educational abilities. The years of schooling completed were converted to an "equivalent" value on the GED scale. This was done in 132U.S. Bureau of the Census, U.S. Census of Population: 1960. Detailed Characteristics, United States Summary. Final Report, PC (1) — 1D (Washington: U.S. Government Prifiting Office, 1963), pp. XVIII—XIX. 221 133 and accordance with the translation made by Eckaus must be interpreted with the precautions which he specified. The conversion categories are shown in Table C-l, which duplicates the Eckaus categories except that a range of grades, such as 3-5, is used for each GED category instead of just one grade, in this case grade 4 only. Table C-l. General Educational Development Categories GED Category School Year Equivalent in Years 1 0-2 2 3—5 3 6-8 4 9—11 5 12—14 6 15-17 7 >17 Source: R.S. Eckaus, "Economic Criteria for Education and Training," Review of Economics and Statistics, Vol. 46, No. 2 (May, 1964), p. 184, Table 2. Eckaus noted, in particular, that he had conflicting advice in making the translation and that he chose that which represents "higher standards" for the general school system. The translation, he knew, was controvercial and he did not intend it to be definitive.134 133R.S. Eckaus, "Economic Criteria for Education and Training," Review of Economics and Statistics, Vol. 46, No. 2 (May, 1964), pp. 181—190. 134Ibid., p. 185 (note). 222 Fine emphasized, in addition, that the U.S. Department of Labor "resolutely avoided" attempts to translate their GED estimates into a time scale related to years of schooling completed, because a job's GED requirement could have been rated at a different level for each scale — reasoning, mathematical, and language development.135 With these precautions in mind, the Eckaus translation was, nevertheless, utilized. It was believed that the upward bias of the Eckaus scale which resulted from using "higher standards" for school systems was offset here by the use of a range of grade levels for each GED category. B. Specific Vocational Preparation Measures High School and College Training If the respondent indicated that he was attending regular school or college at the time of the survey, no SVP score was calculated for him. To determine the extent of high school vocational training received, the high school program was subdivided as follows: . College preparatory program 3 Business program Vocational program General program Only one program given Did not attend high school O‘iU‘I-bWNH . n . In general, few respondents over 25 years of age had attended high school when such specialized programs were a 135Fine, "Use of the Dictionary of Occupational Titles...," p. 367. 223 part of the curriculum, so most took courses only in the general program and no SVP equivalent was credited for this program. Vocational agriculture courses would have been the major exception to this general finding. For female respondents, home economics courses were not considered to be vocational training. For those respondents having taken courses in the high school business or vocational programs, the following computational procedures were utilized: 1. It was assumed that high school vocational courses given only in grades 11 and 12 were advanced enough to qualify as actual job training (below these grades most vocational courses were considered to be introductory in nature). 2. It was assumed that students attended vocational courses two hours each school day. The school year was about 180 days long resulting in 360 hours of training which were equivalent to nine weeks of 40 hours each - the normal work-week. 3. Following Fine,136 high school vocational training was considered equivalent to one—half the number of hours of on—the-job training. Thus, each year of high school vocational training, equaling nine 40 hour weeks, was equivalent to 4.5 weeks of SVP equivalents. Allowing for absences from school, and to avoid the use of fractions of weeks, the SVP equivalent of one year of high school vocational training was set at 4 weeks. For respondents having taken the college preparatory program in high school, no SVP equivalents were given. For each respondent who attended college, the following computat- ional procedures were utilized: 1. It was assumed that the first year of college consisted of basic or introductory courses and 136Ibid., p. 368. did not contain any vocational instruction. 224 2. Beyond the first year of college, each year was regarded as one following Fine. igglf specific vocational preparation, 3. Since each regular college year usually consists of thirty weeks of class attendance, each year of college completed beyond the first year was set equivalent to 15 weeks of specific vocational preparation. Thus, for respondents having taken vocational programs in high school and for those having completed two or more years of college, the SVP weeks equivalent given in Table C-2 were applied to determine the amount of SVP obtained from school or college type courses. Table C-2. SVP Weeks Equivalent of High School Vocational and College Training Years of schooling SVP Years of Schooling SVP completed including weeks completed, no high weeks high school vocation- equiva- school vocational equiva- al lent lent <11 0 -— -— ll 4 -— -- 12 8 -- —- 13 8 (14 0 14 23 14 15 15 38 15 30 >15 53 >15 45 Vocational or Technical Training When evaluating the SVP equivalent of training that was exclusively of a vocational or technical nature, only the 137Ibid. 225 latest training course taken within the last ten years was considered, and only training courses lasting more than six weeks were included. It was assumed that training taken ; more than ten years earlier would presently do little to improve the income position of these families as evidenced by the fact that they were now in poverty, and a training program of less than six weeks was assumed to make less than a marginal contribution to an individual's earning ability. Training which was included was vocational or technical training courses, apprenticeship, and formal, government sponsored on-the—job training programs. Each week of attendance in any of these training programs was considered as a full week of vocational preparation, with two exceptions. If the training course was described as "home economics," it was not considered to be vocational training, and if it was described as a "school or college type course," each week of attendance was considered to be oneehalf of a week of vocational preparation as instructed by Fine. Furthermore, adjustments i were made (arbitrarily) for training which had not been used for some time or which had never been used. The SVP ratios of total attendance time credited for the several categories of when the training was taken and how much it was used are shown in Table C—3. Briefly, for unused training, that taken more than 5 years previously was not counted at all; that taken 3—5 years previously was given one—half credit; that taken 1-2 years previously was given full credit. 2267 For training that had been used but which may have been recently unused, that taken before 1962 and last used before 1965 was given one-half credit; that taken 1962 through 1964 and last used before 1965 was given three-quarters credit; last 10 years and last given full credit. and any training taken within the used in the past two years was Table C—3. SVP Ratio of Unused Training Year SVP ratio Year Latest Year SVP ratio Training if never Training Training was if recently Taken used Taken used unused 1956—1961 0.0 1956—1961 Before 1965 0.5 1962-1964 0.5 1962—1964 Before 1965 0.75 1965—1967 1.0 1956-1967 1965-1967 1.0 By adding the SVP weeks equivalent of high school vocational courses, years of college attendance, and other vocational and technical training courses taken by each respondent, a total measure of their vocational preparation in terms of weeks of training was obtained. This measure was then converted to the specific vocational preparation (SVP) scale developed by the U.S. Department of Labor which is given in Table C-4. 227 Table C—4. SVP Equivalents of Weeks of Vocational Trainingp Training Time SVP in Weeks Equivalent 0 1-4 5-13 14-26 27-52 53-104 105-208 209-520 >520 \OmQGWubWNI-J Source: U.S. Department of Labor, Selected Characteristics of Occupations, A Supplement to the Dictionary of Occppational Titles, 3rd ed. (Washington: U.S. Government Printing Office, —196'6T,' p. A-5 . luau 0 6 3 5. 2. 4 1 3_ O 3 9 2 1 3 .