A UNITED STATES 
 DEPARTMENT OF 
 COMMERCE 
 PUBLICATION 
 
 *»» T0 'C<v 
 
 '^r ES of ' 
 
 C4(,.4 ■ 
 
 6 A 
 
 
 > J S£ - 
 
 The Economic Development Administration's 
 
 Business 
 oan Program 
 
 AN 
 EVALUATION 
 
 U.S. 
 
 DEPARTMENT 
 
 OF 
 
 COMMERCE 
 
 Economic 
 
 Development 
 
 Administration 
 
Digitized by the Internet Archive 
 
 in 2012 with funding from 
 
 LYRASIS Members and Sloan Foundation 
 
 http://archive.org/details/evaluationofeconOOchil 
 
AN EVALUATION OF THE 
 ECONOMIC DEVELOPMENT ADMINISTRATION'S 
 BUSINESS LOAN PROGRAM 
 
 prepared by 
 
 Chilton Research Services 
 Philadelphia, Pennsylvania 
 
 and 
 
 CONSAD Research Corporation 
 Pittsburgh, Pennsylvania 
 
 for the 
 
 Economic Development Administration 
 
 This economic research study was accomplished by 
 professional consultants under contract with the 
 Economic Development Administration. The statements, 
 findings, conclusions, recommendations, and other 
 data in this report are solely those of the contractors 
 and do not necessarily reflect the views of the 
 Economic Development Administration. 
 
 July 1969 
 Reprinted July 1970 
 
 U.S. DEPARTMENT OF COMMERCE 
 Maurice H. Stans, Secretary 
 
 Rocco C. Siciliano, Under Secretary 
 
 Robert A. Podesta, Assistant Secretary 
 for Economic Development 
 
 
TABLE OF CONTENTS 
 
 Page 
 
 I. INTRODUCTION 1,1 
 
 A. Objective and Method of this Study 1»1 
 
 B. Data Collection 1.2 
 
 C. Analysis 1.10 
 
 II. SUMMARY OF FINDINGS AND CONCLUSIONS AND 
 
 RECOMMENDATIONS FOR FURTHER RESEARCH .... 2.1 
 
 A. Findings 2.2 
 
 B. Conclusions 2.4 
 
 C. Recommendations for Further Research z, 6 
 
 III. THE DIRECT ECONOMIC EFFECTS OF THE EDA 
 
 BUSINESS LOAN PROGRAM 3.1 
 
 A. Employment Impacts 3.2 
 
 B. Income Impacts , 3.4 
 
 C. Costs 3.7 
 
 IV. CHARACTERISTICS OF EMPLOYEES SURVEYED . .... 4.1 
 
 A. Profile of the New Employee 4.1 
 
 B. Characteristics of New Employees 4.2 
 
 C. Special Emphasis Groups 4.6 
 
 V. ANALYSIS OF THE DETERMINANTS OF PROGRAM 
 
 SUCCESS 5.1 
 
 A. Analysis of the Determinants of Success Probability . 5. 3 
 
 B. Analysis of Determinants of the Degree of Success • • 5.11 
 
 C. Conclusion 5.18 
 
 VI. EVALUATION OF THE EDA BUSINESS LOAN PROGRAM . 6.1 
 
 A. The Selection of Evaluation Criteria 6.1 
 
 B. Program Evaluation with Incidence Criteria 6.12 
 
 C. Program Evaluation in Terms of Aggregate Impacts . 6. 25 
 
 D. Conclusion 6. 28 
 
 in 
 
 . 
 
TABLE OF CONTENTS 
 (Continued) 
 
 Page 
 VII. OPTIMAL SIZE OF LOAN 7.1 
 
 APPENDIX A: Calculation of Incidence Multiplier 
 APPENDIX B: Calculation of Indirect Impacts Upon 
 
 Employment and Unemployment 
 APPENDIX C: Community Leadership Survey 
 
List of Exhibits 
 
 Page 
 
 ARA-EDA Business Loan Universe Distribution 1.4 
 
 Completion Rate 
 
 Sample Firms 
 
 Employees in Cooperating Firms 
 
 Employment by Prior Status 
 
 Change in Residence 
 
 Income Status 
 
 Measures of Program Effectiveness 
 
 Changes in Employment Status and Residence 
 
 Heads of Households by Prior Status 
 
 Non-Heads of Households by Prior Status 
 
 Annual Income 
 
 Of New Employees 3. 16 
 
 Increases of New Employees 3.17 
 
 Aggregate Annual Income 
 
 (Low) Income New Employees 3.18 
 
 (Middle) Income New Employees 3.19 
 
 (High) Income New Employees 3. ZO 
 
 Cost Per Employee 3. 21 
 
 1. 
 
 7 
 
 1. 
 
 8 
 
 3 
 
 9 
 
 3 
 
 10 
 
 3 
 
 11 
 
 3 
 
 12 
 
 3 
 
 13 
 
 3 
 
 14 
 
 3 
 
 15 
 
List of Exhibits 
 (Con't) 
 
 Prior and Current Employment Status by: 
 
 Sex 
 
 Age 
 
 Race 
 
 Former Employment Status 
 
 Prior and Current Employment Status by: 
 
 Education 
 
 Heads of Households 
 
 Length of Residence 
 
 Skills 
 
 Holding Another Job 
 
 Nearness to Job 
 
 Questionnaire - Employee Trend Study 
 
 List of Variables in the Analysis of the Determinants of 
 Program Success 
 
 Regression with Success Probability as Dependent 
 
 Variable Number of Observations: 68 
 
 Statistical Estimates of the Discriminant Function . . . 
 
 Estimated Versus Actual Outcome 
 
 Regression with ^ Y as Dependent Variable Number of 
 Observations: 33 
 
 Page 
 
 4.7 
 
 4. 8 
 4.9 
 4.10 
 
 4.11 
 4.12 
 
 4, 
 
 ,13 
 
 4, 
 
 ,14 
 
 4. 
 
 15 
 
 4, 
 
 16 
 
 4.18 
 
 5. 5 
 
 5.7 
 
 5.9 
 
 5.10 
 
 5.13 
 
 vi 
 
List of Exhibits 
 (Con't) 
 
 p *g e 
 
 Regression with ^ E as Dependent Variable Number of 
 
 Observations: 36 5.17 
 
 Direct Income Change Impacts 6.14 
 
 Direct and Indirect Income Impacts 6.16 
 
 Current Value of Future Benefit Stream 6.18 
 
 Benefit-Cost Ratios Based Upon Income Change Impacts ^ 21 
 
 Analysis of Program Costs Per Unit Effectiveness 6. 24 
 
 Total Cost and Benefit Impacts 6. 27 
 
 Dependent Variable - Income Per Dollar Loan 7.2 
 
 Estimated Optimal Loan Size 7.10 
 
 Percent Distribution of Consumer Expenditures by 
 
 Commodity Type and Income Group A-6 
 
 Import Coefficient - Percent Distribution of Response to 
 Questions, "Indicate What Percent of Purchases Are 
 Made in the County in Which You Live, " and Calculation 
 of Average Percent „ # m A-7 
 
 C Matrix - Probability That One Dollar Earned by Group (i) 
 
 is Spent in Commodity (k) and Within County A-8 
 
 Percent Income Distribution by Industry and Value Added as 
 
 a Percent of Purchases A-9 
 
 L Matrix - Probability That One Dollar of Receipts in 
 Commodity Sector (k) is Paid as Earnings to Income 
 Group (j) and Within County A -10 
 
 VII 
 
List of Exhibits 
 (Con't) 
 
 Page 
 
 L Matrix - Probability That a Dollar Spent by Local 
 
 Income Group (i) is Earned by Local Income Group (j) . . . A-ll 
 
 County Growth of Job Opportunities C-8 
 
 Job Opportunities C-9 
 
 Community Services - Expansion C-10 
 
 New Stores C-ll 
 
 Number of Community Leaders by Regions C-12 
 
 Population Change C-13 
 
 Opinion Leaders Questionnaire C-14 
 
 Distribution of Interviews For Opinion Leaders Phase C-19 
 
 Vlll 
 
I. INTRODUCTION 
 
 A. Objective and Method of This Study- 
 It is the legislated goal of the Economic Development Administration 
 to reduce unemployment in areas (counties) where the unemployment 
 rate is in excess of six percent and to reduce poverty in areas where 
 the median family income is less than 40 percent of the national average. 
 In pursuit of those goals, EDA administers what is essentially a four- 
 pronged action program: loans and grants for public works investments, 
 business loans, planning grants, and technical assistance programs. This 
 report presents an evaluation of the economic impacts, the business 
 loan program, as jointly conducted by Chilton Research Services of 
 Philadelphia and CONSAD Research Corporation of Pittsburgh. 
 
 The objective of this study was to evaluate, in terms of its economic 
 impacts on depressed areas, the Economic Development Administration's 
 business loan program. More specifically, this study has sought to 
 determine: 
 
 the extent of program benefits; 
 
 the factors underlying program success; 
 
 the incidence of program benefits. 
 
 1.1 
 
The framework in which the findings of this study are analyzed 
 is that of benefit- cost analysis. This framework has been used in the 
 past by EDA to evaluate its business loan program; but, due to the 
 small number of loan recipients sampled in past studies and to the fact 
 that probability sampling was not employed, the reliability in these past 
 studies has not been known. This study, conducted as it was among a 
 probability sample of loan recipients, corrects these shortcomings by 
 obtaining a number of interviews large enough to permit meaningful 
 statistical analysis. 
 
 Statistical methods, both in the data-collection phase and in the 
 analysis, were used where appropriate. In addition, multiplier analy- 
 sis and Markovian incidence analysis were undertaken. 
 
 B. Data Collection 
 
 In developing a substantial- data base, information was obtained 
 from employees of firms receiving EDA business loans through the 
 medium of a self- administered questionnaire. In addition, opinions and 
 attitudes of leaders of the communities in which the firms located were 
 obtained through the Chilton Research Services TeleCentral system, 
 a battery of Wide Area Telephone Services (WATS) lines at the Chilton 
 Research Services main office in Philadelphia. 
 
 1.2 
 
1. Questionnaire 
 
 The employee questionnaire used in this study was based on. the 
 questionnaire submitted by EDA. The questionnaire was reviewed in 
 meetings with representatives of the Economic Development Adminis- 
 tration, Chilton Research Services, and CONSAD Research Corporation 
 and with Dr. William Miernyk of the University of West Virginia. The 
 questionnaire was formatted by Chilton Research Services. Bureau of 
 Budget approval of this questionnaire was obtained (Bureau of Budget 
 approval number 4I-S-69015), and the questionnaire was reproduced 
 through the facilities of the Chilton Company. 
 
 The questionnaire used for the opinion leaders phase of the study 
 was developed in similar meetings and was formatted and copies pro- 
 duced in the same manner. 
 
 2. Sample Design 
 
 The universe for this study consisted of all ARA-EDA. loans granted 
 prior to June, 1968, and. in the amounts of $100, 000 or more. Of all 
 ARA and EDA loans, this represents approximately 97". 6 percent of the 
 
 total dollars granted and 73. percent of the number of loans. (See Table 1 
 
 th 
 A systematic random sample, based on every n selection, was 
 
 drawn for this study. A random start" was- made and thereafter every 
 
 fourth name was selected from the list of ARA-EDA business loans 
 
 1.3 
 
Table 1 
 
 ARA - EDA Business Loan Universe Distribution 
 
 Distribution by: 
 
 Under $ 50, 000 
 $50,000 - $ 99, 999 
 $ 100,000 - $ 149,999 
 $150,000 - $ 199, 999 
 $200,000 - $499, 999 
 $500,000 - $999,999 
 $ 1, 000, 000 and Over 
 
 Dollar 
 Volume 
 
 Number of 
 Firms 
 
 % 
 
 % 
 
 .6 
 
 13.1 
 
 1. 8 
 
 13. 9 
 
 2-5: 
 
 11. 1 
 
 2.7 
 
 8.7 
 
 16. 3: 
 
 27.2 
 
 14, 9 
 
 11.3 
 
 61.2: 
 
 14.7 
 
 100. 
 
 100.0 
 
 1.4 
 
provided by EDA. From this design, 128 business loans were selected. 
 Of those, 66 loans were eliminated because: 
 
 the loan was granted after June, 1968; 
 
 the loan was less than $100, 000; 
 
 the firm to which the loan was granted was 
 either out of business or in the process of 
 being liquidated; and 
 
 the firm name and location was duplicated 
 in the sample. 
 
 A total of 62 firms was therefore eligible for data collection, 
 
 representing all firms receiving ARA-EDA business loans in the amount 
 
 of $100, 000 or more, prior to June, 1968. 
 
 3. Completion Rate 
 
 In. order to gain as high a questionnaire -completion rate as pos- 
 sible* the following general procedures were employed^ 
 
 a. The Economic Development Administration notified 
 its local offices of the study and identified the con- 
 tractor and the local companies which would be con- 
 tracted. 
 
 b. A letter from EDA's Washington office went to a 
 principal in each of the sampled firms, notifying 
 them of the nature of the study and the importance 
 of their cooperation and identified the contractor. 
 They were also informed that a senior staff member 
 of Chilton Research Services would telephone to set 
 tip an appointment. 
 
 1.5 
 
c » Senior members of Chilton Research Services 
 
 telephoned each sample firm to request coopera- 
 tion and to discuss the procedure to be used in 
 contacting and interviewing the employees. 
 
 d. Interviewers in the field were then told of the 
 procedure that was to be used for each plant 
 they were to visit. 
 
 Most firms requested that the questionnaire be sent to them for 
 distribution to the employees. The Chilton interviewer would then call 
 at a later date, pick up the questionnaires, and review them for com- 
 pleteness while at the plant. Some firms requested that the interviewers 
 distribute the questionnaires one day and call back on a different day. 
 Still others, particularly the firms with a limited number of employees, 
 were willing to bring all the employees together in one room, and to 
 have the interviewer distribute the questionnaires and collect and review 
 them, all in one operation. 
 
 This flexible approach led. to 7"6. 9 percent cooperation among the 
 eligible firms in the sample, ten companies of the 62 having been rated 
 ineligible, as Table 2 shows. 
 
 As to completion rate among eligible employees within the firms, 
 73.3 percent of the employees of cooperating firms filled out the ques- 
 tionnaire. (See Table 3). 
 
 1,6 
 
Table 2 
 Completion Rate of Sample Firms 
 
 Total Sample 
 
 Number Percent 
 
 Firms Selected 62 
 
 Ineligible Firms: 
 
 Seasonal employment, employees not 
 available at time of survey; 4 
 
 Funds not disbursed, cancelled by EDA; 4 
 
 Company on strike- union problems 2 
 
 Eligible Firms: 52: (100.0) 
 
 Firms cooperating in the sample; 40 76. 9 
 
 Waiting for approval from headquarters 
 or firms unable to complete within dead- 
 line; 5 : 9.6 
 
 Refused cooperation; 5: 9. 6 
 
 Other 21 3-9 
 
 1,7" 
 
Table 3 
 
 Completion Rate of Employees in Cooperating Firms 
 
 Number of eligible employees 2*847" 
 
 Number of employees cooperating Z\ 143 
 
 Proportion of employees cooperating 75.3% 
 
 1.8 
 
4. Opinions of Leaders 
 
 Names of community leaders werp obtained by interviewers while 
 calling on the sampled firms. These names were sent to Chilton Re- 
 search Services Philadelphia office and sampled. One hundred com- 
 pleted interviews were obtained in the 40 communities in which the 
 cooperating firms were located. 
 
 5. Timing 
 
 The employee seH-adrninistered questionnaire was mailed to the 
 sample firms starting April 7, 1969. All questionnaires used on this 
 study were received by Chilton Research Services on or before May 14, 
 1969. 
 
 Interviewing for the Opinion Leaders questionnaire began on 
 May 5, 1969, and was completed May 13, 1969. 
 
 1.9 
 
C. Analysis 
 
 Upon completion of the survey by Chilton Research Services, the 
 
 survey data was transmitted to CONSAD Research Corporation for analysis. 
 
 The objectives of the analysis were to determine: 
 
 . the magnitude of direct program impacts in terms of employment 
 and income generation; 
 
 . the incidence of program impacts in terms of goal attainment, 
 including the distribution of jobs to the unemployed and under- 
 employed, and of income to the poor; 
 
 . the characteristics of those who benefitted; 
 
 . the factors underlying the likelihood and magnitude of program 
 success; 
 
 . the indirect benefits generated bv the program through t-he 
 multiplier processes of local economies; 
 
 . the value of the program in meeting social goals in terms of 
 benefits and costs; and 
 
 . the size of loan estimated to generate maximum net program 
 benefits. 
 
 1. The Magnitude and Incidence of Direct Impacts. 
 This analysis required tabulation of the completed questionnaire to 
 :termine the number of jobs 7 and the magnitude of income created by the 
 program. Those impacts were disaggregated by prior status of the employees 
 to determine the extent to which the loan program has reached the unemployed 
 and the poor, as well as in-migrants, in-commuters, and new labor force 
 entrants. They were then further disaggregated to identify the demo- 
 graphic and socio-economic characteristics of each of these incidence 
 
 1. 10 
 
groups. The results are presented in Chapters II and III. 
 
 2. Factors Underlying Program Success. 
 
 The collection of a sizeable data base on program impacts provided 
 the study with an opportunity to apply multivariate statistical techniques 
 for the purpose of identifying the factors most closely associated with 
 program success. However, this necessitated that measures be obtained 
 on the socio-economic characteristics of the counties in which the firms 
 are located, attributes of the firms themselves and of the financing they 
 received. 
 
 Factors most associated with the probability of success, or its 
 complement, the likelihood of default, were analyzed via multiple regres- 
 sion and discriminant analysis. The analysis of the factors most closely 
 associated with the magnitude of success as measured by the number of 
 jobs and by the income generated in each loan-recipient firm made use 
 of multiple regression analysis. The results are described in Chapter V. 
 
 3. Indirect Impacts. 
 
 Income earned by the employees of the EDA loan-recipient firm is 
 either saved, or is spent in the local area or outside the area. To the 
 extent that this income is spent locally, it generates additional income 
 each time it circulates through local hands. The benefits of this effect 
 must be attributed to the EDA loan program. An effort to estimate these 
 indirect impacts required the calculation of an income multiplier. 
 
 1.11 
 
Since it was seen as necessary to the evaluation to identify impacts 
 by incidence group, rather than merely in the aggregate, an incidence 
 multiplier was developed that could provide estimates of the indirect 
 income impacts to low, middle, and high income wage earners. The 
 development of this "incidence multiplier model" is described in 
 Appendix A. 
 
 The "incidence multiplier" was developed out of an a priori formu- 
 lation of the rural income circulation process, in conjunction with know- 
 ledge of the available data. As such, it was necessary to make the sim- 
 plifying assumption that interindustry transactions could safely be ignored. 
 However, for purposes of obtaining an accurate estimate of the aggregate 
 ndirect impacts in the loan-recipient counties, it is clear that the incidence 
 multiplier would understate the actual effect. For this reason, an aggre- 
 gate econometric multiplier model wis developed, and is described in 
 Appendix B. 
 
 4. Evaluation of the EDA Loan Program. 
 
 A number of criteria could be stipulated for purposes of evaluating 
 the EDA loan program. The more traditional of these is the requirement 
 that the benefit-cost ratios based upon the current discounted values of 
 future income and cost streams exceed unity, where income is taken either 
 as (a) the aggregate income or (b) the increased earnings attributed to the 
 loan program. Two additional bases for evaluation have been suggested, 
 
 1.12 
 
in which the achievement of stipulated goals be achieved in greater 
 magnitude, per dollar cost, than alternative programs. One way of 
 measuring goal achievement is the discounted income stream into an 
 infinite future accruing to the poor, and another is the number of indi- 
 viduals taken off unemployment or out of poverty. All four bases for 
 evaluation were undertaken, and the analyses are presented in Chapter VI. 
 
 5. Optimal Loan Size. 
 
 The question arises as to whether a policy of providing much larger 
 loans would contribute more to goal achievement than smaller loans. 
 Although an initial attempt was made to provide clues to this problem in 
 Chapter V in the analysis of the determinants of loan success, the com- 
 plexity of the analysis required a more extensive treatment than could be 
 undertaken in that chapter. Therefore, a more formal formulation of the 
 problem, together with an effort to identify the correct functional form re- 
 lating size of loan to program benefits, was undertaken. Applying regression 
 analysis and a simple optimization technique, an attempt was made to 
 estimate the optimal loan size that would maximize the net benefit 
 generated by each loan. This analysis is described in Chapter VII. 
 
 1.13 
 
II - SUMMARY OF FINDINGS AND CONCLUSIONS AND RECOMMENDATIONS 
 
 FOR FURTHER RESEARCH 
 
 Before summarizing the results of the study, three preliminary remarks 
 are in order. First, no claim is made that the sample of firms is represen- 
 tative of the present portfolio of EDA business loans. As stated earlier, the 
 universe consisted of ARA and EDA loans in excess of $100, 000 and approved 
 prior to June, 1968. * Over the past several months there has been a shift 
 in the agency's lending policy towards loans larger in absolute amount but 
 smaller as a percent of total project cost than those examined in this study. 
 Hence, to the extent that benefits per dollar of EDA loans increase with 
 the size of loan, as is suggested by the results of this study, the economic 
 impact of the current EDA loan program will be underestimated by these 
 findings. 
 
 Second, as in all investigations involving surveys and interviews, the 
 question of what to do about non-respondents arises. What has been done 
 here, and it is virtually all that can be done, is to assume that the distri- 
 bution of attributes is the same for non-respondents as for respondents. 
 Statistical findings presented in this section and elsewhere in the report 
 reflect that assumption. 
 
 *June '68 was chosen as the cut-off in an effort to capture just perma- 
 nent impacts rather than some that may only be immediate. 
 
 2.1 
 
Finally, despite efforts to the contrary, a gray area inevitably exists 
 between the so-called "hard" findings of a study and interpretations of 
 those findings, i.e. , what are generally thought of as conclusions. For 
 this summary, it was decided that the quantitative results of the survey 
 and inferences drawn from statistical analysis of these data would be 
 labeled "Findings, " and all other interpretations, which relate mainly to 
 causal relationships and policy implications, will be called "Conclusions." 
 
 A. Findings 
 
 A sample of 40 firms receiving $23, 180, 000 in ARA-EDA loans 
 during the period 1962-1967 showed that: 
 
 1. 1817 jobs were created, of which 68% went to individuals 
 who were previously employed full time, 9% to those 
 employed part time, 12% to new labor force entrants, 
 and 11% to people who were formerly unemployed. The 
 "cost per job, " i. e. , amount of EDA loans divided by 
 the number of jobs created, was $12,757. 
 
 2. the average increase in individual earnings was $1, 393 
 per employee. 
 
 3. 425 people, or 23%, went from below to above the poverty 
 level income ($4, 000 per year), but 164, or 9%, went 
 from above to below $4, 000 per year. 
 
 4. the number of people taken off the unemployment or 
 
 poverty rolls (and in some cases, both) as a direct 
 result of the new jobs was 346, or 19%. 
 
 -- The type of individual most likely to obtain a job in the loan recipient 
 firms was a family man, under 35 years of age, white, and a high school 
 graduate. He lives in the same county and within six miles of his job, and 
 
 2.2 
 
has lived in that county for at least nine years. He is a semi-skilled worker 
 who is employed full time, has no additional jobs, and earns under $100 per 
 week. 
 
 Of special interest is that: 
 
 1. Over half of those who were previously categorized as 
 
 poor or unemployed were women. 
 
 2. Non-white employees are much more likely to have 
 previously been unemployed, part time or farm workers. 
 
 3. Almost two-thirds of the previously unemployed were 
 not heads of households. 
 
 Factors associated with the likelihood of a firm's success, i.e., the 
 
 probability that it will remain in existence, include: 
 
 1. its being located in a relatively large town. 
 
 2. its being located in a town with low manufacturing wages 
 and a low growth rate. 
 
 3. the fact that it has received a loan that is large relative 
 to the ARA-EDA average. 
 
 Factors associated with the magnitude of success, measured in terms 
 
 of pre-loan to post-loan change in employment or total wages paid, given 
 
 that the firm remains in existence, include: 
 
 1. larger loans. 
 
 2. better educated labor force. 
 
 3. location in towns having a larger proportion of the labor 
 force employed in manufacturing occupations. 
 
 2.3 
 
4. the fact that the firm is located in a town that is small 
 relative to the others in the sample and characterized by 
 higher levels of out-migration and poverty, and lower 
 growth rates . 
 
 5. the firm is engaged in manufacturing, and the quality 
 of its labor force is not a critical factor. 
 
 An analysis of the most appropriate loan size that should be provided- - 
 
 most appropriate from the standpoint of generating the highest net benefit 
 
 from the program- -showed that even with the most conservative approach, 
 
 loans in excess of $2 million each are indicated. This is four times as 
 
 great as the average loan disbursed during the study period, 1962-1967. 
 
 B. Conclusions 
 
 In terms of performance, it was found that the loan program generated 
 aggregate benefits far in excess of its cost. When benefits are considered 
 applicable to all recipients regardless of prior income or employment 
 status, benefit-cost ratios were found without exception to be greater 
 than unity. Even when benefits are viewed as applicable only to econo- 
 mically distressed individuals, the program directly or indirectly (as a 
 result of the multiplier process) caused 520 people to be removed 
 either from the poverty or unemployment rolls, at a cost ranging from 
 $17, 500 to $25, 000 per person. This cost is undoubtedly less than the 
 cost to society of letting these individuals (and in many cases, their 
 dependents also) remain unemployed or in poverty over their lifetimes. 
 
 2.4 
 
On the basis of the finding that 69 percent of the direct increase in 
 income was incident to individuals who are not poor, and from the 
 analysis of the magnitude and incidence of multiplier effects, it is 
 concluded that a program designed to generate economic impacts by 
 stimulating economic development must generate an impact far in 
 excess of that needed to attain goals relating strictly to poverty and 
 unemployment. 
 
 Through statistical analysis of the factors underlying the success 
 of firms and the program itself, it was found that greater impacts 
 were obtained in regions having better quality manufacturing 
 labor, but undergoing economic decline as indicated by out- 
 migration and negative growth rates. Evidently, the reduction 
 in the competition for a good labor force results in lower wage 
 rates, which gives the EDA assisted industry an additional, and 
 effective, competitive advantage. 
 
 Finally, the study suggests that the recent trend in EDA toward 
 larger loans is a step in the right direction. In fact, the 
 evidence indicated that loans in excess of $2. 5 million each would 
 maximize net program benefits. However, the results relating 
 to EDA's share of total project cost were neither as favorable 
 nor as conclusive with respect to current policy. The variable 
 
 2.5 
 
representing EDA's share consistently bore a positive relation- 
 ship with measures of success of the loans. This finding is in- 
 consistent with the hypothesis underlying the current policy of 
 reducing the relative size of the agency's participation, but it 
 should be noted that the variable failed to pass tests of statistical 
 significance. Two --and obviously conflicting- -explanations can 
 be given to this. The first is that the result is a valid one, and 
 it is due to the fact that the effects of the higher cost of capital 
 obtained from sources other than EDA more than offset the 
 "leverage" which the agency realizes by reducing its degree of 
 participation. Second, is the argument that projects being financed 
 under the new policy are qualitatively different from those included 
 in the sample, and the results of the statistical analysis are there- 
 fore not applicable to this issue. The authors of this report take 
 a neutral position with respect to these two points of view. 
 
 C. Recommendations for Further Research 
 
 An analysis of the type described in this study, if properly designed, 
 can make use of a relatively limited budget to provide answers 
 to a number of relevant questions. But budgets do create limitations, 
 and study designs always seem to have left some stones unturned, 
 especially when they are viewed in retrospect. Therefore, several 
 recommendations for future work will be made. 
 
 2.6 
 
The analysis of income impacts would be made more informative 
 by broadening the survey to organize income data around the family 
 rather than the individual. Although such a study would be more 
 costly and difficult to administer, it would be valuable to know whether 
 the added earnings of an individual, who remains poor, nevertheless 
 contributes enough to the family of which he is a member to raise it 
 out of poverty. 
 
 The musical-chairs effect, discussed briefly below and in more 
 detail in Chapter VI, remains uncovered. According to the optimistic 
 view of this effect, although 70 percent of the jobs created in loan- 
 recipient firms went to persons who were previously employed full 
 time, a certain portion of the jobs released by these individuals will 
 be obtained by the poor and unemployed. The remainder being obtained 
 by those previously employed full time, who release jobs, a certain 
 portion of which . . . and so on. If this view is correct, the program 
 will yield several times more benefits than if the effect does not operate 
 at all. If, however, a larger portion of the jobs released by the 
 previously employed are not filled due to the upward pressure on local 
 wages caused by the reduced labor supply (See Chapter VI), then the 
 musical-chairs effect will be relatively small. 
 
 2.7 
 
What is needed is an in-depth study, perhaps a case study, tracing 
 whether the jobs released by "job- switchers" were filled either by the 
 unemployed, in-migrants, new entrants, by other job-switchers, or 
 went unfilled, resulting in a contraction of the local economy that did 
 not receive EDA aid. 
 
 A larger base of data on disbursed loans and their correlates, 
 extending to 75-100 observations, is needed to improve the statistical 
 analyses attempted in Chapter V and VII. The data would describe the 
 employment level of each firm and the total wage bill, size of loan and 
 a more detailed characterization of the firm, as well as locational 
 characteristics. This information could be gathered largely by area 
 representatives in local field offices, and without requiring survey 
 techniques. 
 
 The analysis of optimal loan size in Chapter VII is highly dependent 
 upon the correct identification of the nature of the mathematical 
 relationship between the size of loan and the magnitude of benefits 
 generated by the loan-recipient firm. Further research is needed 
 on the exact form of this relationship. In addition, other variables 
 relating to the magnitude of impact should be included. To meet these 
 analytic requirements, it will be necessary to perform a more intensive 
 analysis, employing more advanced statistical methods, and making use 
 of a considerably larger sample. 
 
 2.8 
 
III. THE DIRECT ECONOMIC EFFECTS OF THE EDA BUSINESS LOAN 
 PROGRAM 
 
 This chapter is to describe and analyze the direct economic effects 
 generated by 40 firms as a result of EDA and ARA loans to these firms. The 
 analysis included changes in employment status and income status. Only 
 changes in the status of new employees will be considered in this chapter. 
 
 In order to define changes which have the most relevance for EDA, special 
 emphasis is placed onthree groups; the formerly unemployed, the poor, and 
 heads of households. These categories are not mutually exclusive and an ef- 
 fort was made to determine the extent of overlapping. For this study, an annual 
 income of less than $4000 for an individual employee was considered a 
 poverty-level income.* 
 
 The following summarized the direct impacts of $23, 180, 000 in 
 loans disbursed by ARA and EDA to a sample of 40 firms: 
 
 1. 1,817 people gained jobs from EDA loan-recipient firms after 
 the loans were received. 68% of these employees had previously 
 been employed full time. 23%had been unemployed or under- 
 employed (Table 1). 
 
 2. 6% of the new employees had migrated into the county in which the 
 EDA firm was located. 14% of the new employees were currently 
 in-commuters and lived in nearby counties (Table 2). 
 
 * The decision to focus upon individuals earning under $4000 per year raises 
 a question as to whether the low income wage earners should be considered as 
 poor. Although poverty income is properly discussed on a family or household 
 basis rather than on an individual ba^is, the employee survey reported in this 
 study was taken at place of work rather than at place of residence, making 
 income estimation on a family basis considerably more complicated. In 
 addition an annual income of $4,000 rather than $3, 500 was chosen as the 
 cutoff point since the latter would have surpassed the large number of effects 
 reported in employees in the low income range between $4i000 and $4,000. 
 
 3.1 
 
3. Among those hired by EDA loan-recipient firms, the annual 
 increase in income has averaged $1, 393 per employee. Because 
 of this additional income, there was a net increase of 315 em- 
 ployees with incomes above poverty level (Table 3). 
 
 4. The number of people taken off the poverty or unemployment rolls 
 as a direct result of the new jobs were 346 (19% of the total) 
 (Table 4). 
 
 A. Employment Impacts 
 
 Data was analyzed for 40 loan-recipient firms with a total of 2, 848 
 employees. Of these, 1, 817 were considered to be new employees (hired 
 after the date of approval of an EDA or ARA loan).* 
 
 Table 5 shows changes in employment and residence status among 
 these new employees. 58 of the 1, 031 employees hired before the date of the 
 loan worked for companies who received EDA loans to prevent them from go 
 ing bankrupt and, on this, could have been, but were not, included in the 
 measure of success of the program. 
 
 It is to be noted that 68% of the new employees had previously been 
 employed full time in another job. This high proportion is over twice that 
 expected on the basis of the Miernyk study (which reported that 34% of its 
 sample had been fully or partially employed). ** However, like the 
 
 * No information was available on how many of these employees 
 were hired to fill positions in existence before the EDA loan. 
 
 ** Evaluation of ARA-EDA Loan Program , Office of Economic Research, 
 Economic Development Administration, November, 1968, page 2. It should 
 be noted, however, that the low proportion reported in the Miernyk study 
 occurred during a period when the national unemployment rate was considerably 
 higher --5. 6% versus the current 4%. 
 
 3.2 
 
Miernyk study, the survey results showed that many of the previously 
 unemployed had not been members of the labor force and that most of 
 these new labor force entrants were women (See Chapter III). 
 
 Former part-time employees (171) make up about 10% of the total 
 of new employees, and only 14 of these were still employed part time. 
 Almost all these employees responded that, while they had been working 
 part time before, they did not prefer to work part time. They were, 
 therefore, considered underemployed. 
 
 Data on new labor force entrants reinforces the belief that new 
 employment opportunities must greatly exceed employment targets for 
 the unemployed and underemployed. The number of new labor force 
 entrants (216) is larger than the number of formerly unemployed (196). 
 
 However, about one-fourth of the new labor force entrants classified 
 themselves as heads of households who probably (a) would have entered 
 the labor force in the near future or (b) were labor force dropouts who 
 had been unemployed for a long period of time and were no longer actively 
 looking for work. 
 
 Almost 80% of the heads of households are now employed full time 
 and had been employed full time. Former unemployed and underemployed 
 persons made up 14% of the heads of households .while almost 2 9% of the non- 
 heads of households hired had been unemployed or underemployed (See 
 Tables 6 and 7). 
 
 3.3 
 
Most (86%) of the new employees were county residents. About 
 7% of these were in-migrants who had not been living in this county 
 before starting their present job. About 1% of the new employees were 
 in-migrants who had moved to nearby counties and were currently in- 
 commuters (See Table 2). 
 
 B. Income Impacts 
 
 The following discussion deals only with direct income benefits of 
 EDA loans. Indirect benefits generated by the multiplier process, both 
 in terms of income and employment, will be described in Chapter V . 
 
 Income data was derived on an annual basis from the weekly earn- 
 ings of EDA income and other jobs reported in the survey and annual 
 net earnings from farm employment. 
 
 For the total group of new employees, income from EDA employ- 
 ment over the period of a year was $7, 748, 000. The average annual 
 income from EDA employment was $4,264, only slightly above the 
 poverty level (See Table 8). 
 
 Since most employees had no other job or farm income, those who do 
 have additional income, make a great deal more than the $260 annual difference 
 between current income and EDA income indicated by the means. Average EDA 
 incomes of heads of households are abour 18% higher than incomes for non- 
 heads of households. Additional income from other sources of employment 
 is also higher for heads of households. In fact, heads of households accounted 
 for most of the income from other jobs and farm employment 
 
 3. A 
 
The increase in income from prior employment to current employ- 
 ment was computed on an annual increase basis (See Table 9). Since being 
 hired by their current employers, the average annual increase for heads of 
 households was about $381 less per year than for non-heads of households. 
 Since their current average incomes were more than those of non-heads 
 of households, the prior income of heads of households exceeded the prior 
 incomes of the non-head of households by more than the 18% current 
 difference. 
 
 The annual average increase was not uniform among low, middle 
 and high income groups. New employees were divided into three groups 
 according to their total current annual income. The low-income group 
 was made up of those individuals earning less than $4,000 a year. The 
 middle income group included those reporting incomes between $4, 000 
 and $10,000 a year and the high income group included those making 
 above $10,000 per year. 
 
 Over 50% of the new employees were still earning less than $4, 000 
 a year (See Table 10). Mean current income ($3, 150) was only slightly 
 more than mean EDA income ($3, 133) for the poor. This implies that 
 few secondary jobs were available to supplement poverty-level incomes. 
 
 Of the 913 new employees with poverty-level incomes, less than half 
 were heads of households. Their mean EDA income and mean total income 
 were slightly more than the means for non-head of households. However, 
 the average annual increase in income for heads of households ($429) 
 
 3.5 
 
was about one -third of the average annual increase for those that were 
 not heads of households ($1, 209) in the low income group. 
 
 Almost 71% of the formerly unemployed were in the low income group. 
 However, only 22% of the unemployed were heads of households with poverty 
 level incomes. The major gains in raising income above poverty levels were 
 for heads of households (See Table 4). 
 
 The middle income group included 479 new employees who were formerly 
 in the poverty group but now earned over $4, 000 a year (See Table 3). However, 
 164 employees who reported former incomes above $4, 000 now had incomes 
 below the poverty level. Most of these indicated a decrease of $1, 000 - $2, 000 
 a year, which may not be unusual in a rural situation where other sources of 
 jobs were closed or moved to different locations. A few employees reported 
 decreases of up to $7, 000 a year, and no apparent explanation is available in 
 these cases. * 
 
 This movement between income groups results in a net increase for the 
 middle income group of 315 employees. For this income group the difference 
 between EDA income average ($5, 350) and the total current income ($5, 619) was 
 much higher than that of the lower income group ($1, 800)*. 
 
 Heads of households made up about 69% of this group and mean incomes were 
 about $350 higher for them than for non-heads of households. Annual average 
 increase was less than two-thirds of that for middle income employees who 
 are not heads of households (Table 11). 
 
 * This data is available from responses to this initial survey questionnaire 
 provided by Chilton Research Services. (See Exhibit IV-1) 
 
 3.6 
 
For the over-$10,000 income group, the average income from other 
 sources ($5,719) almost equaled average income from EDA-financed 
 jobs ($6,511). From the high figure for secondary income and the large 
 average annual increase in income ($6, 906), it was evident that EDA had 
 greatly expanded the earnings total for this group (See Table 12). 
 
 Heads of households in this category had a much smaller propor- 
 tion of EDA income to total income and, as in the poverty and middle 
 income groups, a smaller average annual increase in income. 
 
 C. Costs 
 
 Over the five-year period 1962-67, the 40 firms whose employees 
 were surveyed received at least one business loan from EDA-APA, 
 Four firms received two loans in this time period. 
 
 Total amount of the loans was $23, 180, 000. Individual loan amounts 
 ranged from $56, 000 to $2, 620, 000, the average being about $527, 000. 
 For the proposed projects, EDA and ARA loan money was supplemented 
 by $15,435,000 from other sources -- private industry and state and 
 local governments. The major portion (60%) of the project money, how- 
 ever, was from EDA and ARA. 
 
 Seven companies having a total of 514 employees had not hired any 
 new employees since the loan approval. * Therefore, these firms gen- 
 
 * Loan approval dates for these companies averaged over five 
 years ago. 
 
 3.7 
 
erated no direct employment or income impacts. Thirty- three companies 
 had hired a total of L, 874 new employees and retained 460 employees who 
 were hired before theloan was received. 
 
 The size of the companies varied greatly -- from one with only four 
 employees to one with 650 employees. The average company had approxi- 
 mately 70 employees of which 57 were new employees. 
 
 On the basis of the data for 40 firms, the loan cost per new job 
 created was $IZ, 40Q, or said another way, 8. 1 jobs were created for every 
 $100, 000 in EDA loans. This estimate is based upon the number of employees 
 working in EDA loan-receipient firms hired after the data of loan. In- 
 cluded are the 58 jobs that would have been. destroyed by bankruptcy. This 
 estimate is considerably higher than most of the cost figures reported by 
 the Evaluation of the ARA and EDA Loan Program. * If, instead of counting 
 only those employees- hired after the data of loan, we base our estimate 
 upon the total employment in EDA loan-assisted firms, then the loan cost 
 per job becomes $8, 139, which is equivalent to 12.3 jobs per $100, 000 in 
 EDA loan. These figures are considerably closer to those reported in the 
 earlier EDA evaluations (See Table 13). 
 
 * Miller, S.,D. Gaskins, C. Liner, Evaluation of ARA and EDA Loan 
 Program , Office of Economic Research, Economic Development Ad- 
 ministration, November 10, 1968. 
 
 3.8 
 
Table 1 
 
 Employment by Prior Status 
 
 " — — _^____ Present 
 Prior " — . -_^_Status 
 
 Status " — — 
 
 Full Time 
 
 % Full Time 
 
 Part Time 
 
 % Part Time 
 
 Employed Full Time 
 
 1217 
 
 67.0 
 
 17 
 
 1.0 
 
 Employed Part Time 
 
 157 
 
 8.6 
 
 14 
 
 .7 
 
 New Labor Force Entrants 
 
 202 
 
 11. 1 
 
 14 
 
 .7 
 
 Unemployed 
 
 188 
 
 10.4 
 
 8 
 
 .5 
 
 Total 
 
 1764 
 
 97. 1 
 
 53 
 
 2.9 
 
 3.9 
 
Table 2 
 Change in Residence 
 
 ^-^^ TO 
 
 
 
 FROM ^^-\^^ 
 
 County Resident 
 
 In-Commuter 
 
 County Resident 
 
 1448 
 
 14 
 
 In- Commuter 
 
 
 223 
 
 In-Migrant 
 
 111 
 
 21 
 
 3.10 
 
Table 3 
 
 Income Status 
 
 * "s^^Current Status 
 Prior Status V-N * ,, *««^^ 
 
 Poor 
 
 Not Poor 
 
 Poor 
 
 Not Poor 
 
 749 
 164 
 
 479 
 425 
 
 TOTAL 
 
 913 
 
 904 
 
 3.11 
 
Table 4 
 Measures of Program Effectiveness 
 
 "~~ --— -^_^ Prior 
 Above ■ — -^______^ Status 
 
 Poverty &i EmplT^ — ^___ 
 
 Poor -Employed 
 
 Poor & Unemployed 
 
 Mot Poor-Unemployec 
 
 Not Heads of Households 
 Heads of Households 
 Total 
 
 -40 
 198 
 
 90 
 41 
 
 33 
 34- 
 
 158 
 
 131 
 
 , 
 
 57 
 
 * Net effect. Number raised from poverty minus the number whose income 
 dropped below $4000. 
 
 3.12 
 

 4) 
 O 
 
 C 
 
 o 
 
 a' 
 
 m 
 
 ui 
 .2 
 
 6 
 >- 
 O 
 
 i— ■ 
 
 t 
 
 W 
 
 c 
 
 en 
 
 1) 
 
 c 
 
 rtS 
 
 U 
 
 1—1 
 rt 
 +-> 
 o 
 
 H 
 
 3 
 o 
 H 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 cm 
 
 »— t 
 
 r- 
 
 i — i 
 
 ■ 
 
 <M 
 
 NO 
 
 CO 
 
 In- 
 Commuter 
 
 
 
 
 
 
 
 
 
 
 
 
 in 
 
 CM 
 
 CM 
 
 CO 
 
 m 
 CM 
 
 County 
 Resident 
 
 
 
 CO 
 
 
 CO 
 
 
 r- 
 
 O 
 
 "t 
 
 
 o 
 o 
 
 vO 
 
 in 
 
 O 
 
 
 1, 559 
 
 6 
 
 • r-l 
 
 H 
 i 
 
 -t-> 
 
 i— 1 
 
 O 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 •«t 
 
 "* 
 
 co 
 
 CO 
 
 in 
 
 In- 
 
 Commuter 
 
 
 
 
 
 
 
 
 
 
 
 PO 
 
 o 
 
 t 
 
 f— * 
 
 CO 
 
 County 
 Resident 
 
 1 — 1 
 
 CM 
 
 f— ( 
 
 oo 
 
 r~ 
 
 co 
 CO 
 
 CO 
 
 (M 
 
 - 
 
 o 
 
 nO 
 
 >tf 
 
 <tf 
 
 o 
 
 r- 
 
 ^ 
 •^ 
 
 
 r-l 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r- 
 
 r- 
 
 1 
 
 in 
 
 •4* 
 
 
 -t-> 
 
 
 
 
 
 
 
 
 
 
 
 #— * 
 
 ul 
 
 o 
 
 CO 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 (M 
 
 —J 
 
 r j 
 
 f—< 
 
 c- 
 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 - 
 
 
 
 
 - 
 
 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 1 ' 
 
 
 
 
 —* 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 £ 
 
 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 • iH 
 
 -4-1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 i 
 
 In- 
 Commu 
 
 
 
 
 
 
 
 
 
 
 
 » — i 
 
 in 
 
 m 
 
 rj 
 
 o 
 
 |3 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 
 (Nl 
 
 CM 
 
 in 
 
 CM 
 
 -*j 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 >s C 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 t! «J 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O cn 
 
 u » 
 
 CO 
 
 r- 
 
 CO 
 
 (M 
 
 o 
 
 oo 
 
 in 
 
 CO 
 
 ■* 
 
 in 
 
 vO 
 
 fM 
 
 —< 
 
 vO 
 
 in 
 
 
 CO 
 
 c"> 
 
 r> 
 
 vO 
 
 _l 
 
 co 
 
 
 
 
 o 
 
 (M 
 
 ■* 
 
 CO 
 
 vXJ 
 
 •~" 
 
 
 O 
 
 H 
 
 
 
 ■tf 
 
 
 
 
 
 f— 1 
 
 o 
 
 1 ' 
 
 ' ' 
 
 *~ * 
 
 m 
 
 
 « 
 
 
 
 
 
 p— * 
 
 
 
 
 
 
 -" 
 
 
 
 
 ^ H 
 
 go 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 B 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ■*-> 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 x; 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4-> 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 T) 
 
 
 
 
 
 -a 
 
 
 
 
 
 •a 
 
 
 Pres 
 rior Statu 
 
 1) 
 
 s 
 
 •r-t 
 
 H 
 i 
 
 r- 1 
 
 
 a) 
 
 e 
 
 H 
 i 
 
 W 
 
 O 
 i— < 
 
 1 
 
 . — 1 
 
 -*-t 
 O 
 
 0) 
 
 e 
 
 •H 
 
 H 
 i 
 
 3 
 
 a) 
 
 e 
 
 H 
 i 
 
 IT} 
 
 w 
 
 >- 
 
 o 
 
 t— < 
 
 s 
 
 rd 
 
 
 0) 
 H 
 3 
 
 0) 
 
 a 
 
 •r-S 
 
 H 
 i 
 
 w 
 
 0) 
 
 o 
 
 I— 1 
 
 s 
 
 i— < 
 
 i S 
 
 cm 
 
 fa 
 
 Ph 
 
 £ 
 
 p 
 
 H 
 
 h 
 
 i n, 
 
 2 
 
 P 
 
 H 
 
 fa 
 
 Mh 
 
 ^, 
 
 H 
 
 I H 
 
 s^uopisa^i A^unoQ 
 
 3.13 
 
 qUT?.I^TJA[-UI 
 
 l^iox 
 
Table 6 
 Heads of Households by Prior Status 
 
 — — _________ Present 
 
 Prior __Status 
 
 Full Time 
 
 % Full Time 
 
 Part Time 
 
 % Part Time 
 
 Status ■ — 
 
 
 
 
 
 Employed Full Time 
 
 844 
 
 79.7 
 
 13 
 
 1. 2 
 
 Employed Part Time 
 
 76 
 
 7.2 
 
 1 
 
 . 1 
 
 New Labor Force Entrants 
 
 51 
 
 4.8 
 
 1 
 
 . 1 
 
 Unemployed 
 Total 
 
 72 
 
 6.8 
 
 1 
 
 . 1 
 
 1043 
 
 98. b% 
 
 i. 
 
 1. 5% 
 
 3.14 
 
Table 7 
 Non-Heads of Household by Prior Status 
 
 " — - — _______^ Present 
 
 Prior "-" , — ■ __ ______Status 
 
 Full Time 
 
 % Full Time 
 
 Part Time 
 
 % Part Time 
 
 Status ■ 
 
 
 
 
 
 Employed Full Time 
 
 373 
 
 49.2 
 
 4 
 
 0.6 
 
 Employed Part Time 
 
 81 
 
 10.7 
 
 13 
 
 1.7 
 
 New Labor Force Entrants 
 
 151 
 
 19.9 
 
 13 
 
 1.7 
 
 Unemployed 
 Total 
 
 116 
 
 15.3 
 
 7 
 
 . 9 
 
 721 
 
 95. 1 
 
 1 37 
 1 
 
 4.9 
 
 3.15 
 
Table 8 
 Annual Income of New Employees 
 
 ^\^ Income 
 
 Employee ^"\^^ 
 Group ^^^^^ 
 
 Number 
 of Cases 
 
 EDA Annual 
 Income 
 
 Mean 
 Annual 
 
 EDA 
 Income 
 
 Annual 
 
 Current 
 
 Income 
 
 Mean 
 Annual 
 Current 
 Income 
 
 All New Employees 
 
 1,817 
 
 7, 748,000 
 
 4,264 
 
 8.247,000 
 
 4,538 
 
 Heads of Households 
 
 1,059 
 
 4,824,000 
 
 4,555 
 
 5, 265,057 
 
 4, 972 
 
 Non Head of Households 
 
 758 
 
 2, 924,000 
 
 3, 858 
 
 2, 981,606 
 
 3, 934 
 
 3.16 
 
Table 9 
 Annual Income Increases of New Employees 
 
 """"----^^^ Income 
 Employee '"^-^^^ 
 Group — ^ 
 
 Numbe r 
 of Cases 
 
 Total Annual 
 Increase 
 
 Average Annual 
 Increase 
 
 All New Employees 
 
 1.817 
 
 $ 2, 531,000 
 
 $ 1,393 
 
 Heads of Households 
 
 1.059 
 
 1,306.667 
 
 1. 234 
 
 Non Heads of Households 
 
 758 
 
 1,224, 118 
 
 1,615 
 
 3.17 
 
Table 10 
 
 Aggregate Annual Income (Low) 
 Income New Employees 
 
 """""•^^^Employee Group 
 Income ^^^^-^^^ 
 
 Total Group 
 
 Heads of 
 Households 
 
 Non Heads of 
 Households 
 
 EDA Income 
 
 2,860, 196 
 
 1,335,850 
 
 1,524,346 
 
 Total Current Income 
 
 2,876,411 
 
 1.34S, 122 
 
 1,531,289 
 
 Annual Increase in Income 
 
 773,308 
 
 181,411 
 
 591,897 
 
 Number of Employees 
 
 913 
 
 423 
 
 490 
 
 3.18 
 
Table 11 
 
 Aggregate Annual Income (Middle) 
 Income New Employees 
 
 *"*■ .^Employee Group 
 
 Income """^■^^^ 
 
 Total Group 
 
 Heads of 
 Households 
 
 Non Heads of 
 Households 
 
 EDA Income 
 
 4,600,948 
 
 3,258,337 
 
 1,342,611 
 
 Total Current Income 
 
 4,832,127 
 
 3,441,430 
 
 1,390,697 
 
 Annual Increase in Income 
 
 1,453,612 
 
 864,431 
 
 589, 181 
 
 Number of Employees 
 
 860 
 
 597 
 
 263 
 
 3.19 
 
Table 12 
 
 Aggregate Annual Income (High) 
 Income New Employees 
 
 mployee Group 
 
 Total Group 
 
 Heads of 
 Households 
 
 Non Heads of 
 Households 
 
 286,499 
 
 230,041 
 
 56,458 
 
 Total Current Income 
 
 538,125 
 
 478,505 
 
 59,620 
 
 Annual Increase in Income 
 
 303,875 
 
 260,835 
 
 43,040 
 
 Number of Employees 
 
 44 
 
 39 
 
 5 
 
 3.20 
 
Table 13 
 
 Cast Per Employee 
 
 Study 
 
 Loan Amount 
 per Employee 
 
 Employees per 
 $100,000 of loan 
 
 Special Impact Survey 
 
 Survey of Eight ARA-EDA 
 
 Business Loan- 
 Recipient Firms 
 
 Employment and Financial 
 Data from EDA National 
 and Field Office Files 
 
 Office of Audits Review 
 
 Chilton/CONSAD 
 
 $7,705 
 5,291 
 
 5,492 
 
 5,434 
 12,369 
 
 13.0 
 18.9 
 
 18.2 
 
 18.4 
 8.1 
 
 3.21 
 
IV. CHARACTERISTICS OF EMPLOYEES SURVEYED 
 
 In Chapter III, the employment and income generated by the firms that 
 received EDA loans was described. This chapter will describe the socio- 
 economic characteristics of these employees in order to provide a view 
 of the factors that affect the incidence of program benefits. 
 
 Exhibit IV -1 is a copy of the questionnaire that was furnished to 2, 847 
 employees of 40 firms. Of these, 2,127 questionnaires were filled out and 
 returned, and 2, 020 were properly completed. 1, 344 of these were "new" 
 employees . * 
 
 A profile of the average new employee is provided in Section A of this 
 chapter. Section B provides details on characteristics of new employees, 
 in general, and on characteristics of subgroups such as the poor, 
 unemployed and new labor force entrants. 
 
 A. Profile of the New Employee 
 
 The average new employee of the EDA loan recipient firm is a young 
 man, probably under 35 (62%), married, two children. He is most likely 
 to be white (77%) and has probably attended or graduated from high school 
 (68%). 
 
 * Hired after the date of approval of the EDA loan. 
 
 4.1 
 
This employee lives in the same county in which he works (86%), 
 probably within 6 miles of his employer's location (53%), and is a long- 
 term (9 years or more) resident of the community (69%). 
 
 He is employed full time as an semi-skilled worker (64%) and 
 holds only one job in spite of the fact that he earns less than $ 100 a 
 week (77%). He has probably been employed by this company for less 
 than two years (80%). 
 
 Before the employee was hired by the EDA loan recipient firm, 
 he -was employed full time in the same county (57%). He worked for his 
 former employer for three years or less (62%*) and made under $80 
 a week as a semi-skilled worker (6l%*). When he left his former job, 
 his employer probably hired someone to replace him (62%*). 
 
 If the new employee is a woman, she is much more likely to have 
 previously been unemployed or to be a new labor force entrant (42%). 
 She probably is not the head of a household and makes less than $ 4, 000 
 a year (74%). 
 
 B. Characteristics of New Employees** 
 
 Questionnaires with complete information were returned by 1, 344 
 employees hired after the date of approval of an EDA loan. As described 
 
 * % of former full time employees. 
 ** Hired after the date of approval of the EDA loan. 
 
 4.2 
 
in Chapter III, of these 1,344 employees, 97% are currently employed 
 full-time and only 3% are employed part-time. Over two-thirds wero 
 employed full-time before they were hired by their current employer. 
 About 10% had been employed part-time and about 11% had been unem- 
 ployed. The remaining 12% were new entrants to the labor force. * 
 More than half of the total group currently have poverty level incomes. 
 
 Responses were analyzed for 9 characteristics by current and 
 former employment status. A brief summary of the distribution of 
 characteristics is given here. Contained in this chapter are tables 
 showing details employee characteristics. 
 
 T» Sex- 'See Figure I) 
 
 Of the replies received, 62% were from men and 38% were from 
 women. Most of the men. (80%) had' formerly been employed full-time, 
 while less than, half (48%) of the women had been full-time employees. 
 The previous unemployment rate of the women (17. 9%) was almost 
 three times as high as that of the men (6.4%). 
 
 Although almost all of the new employees (97% of both men and 
 women) were currently working full time, only 38% of the men had 
 poverty level incomes, while 74% of the women did. 
 
 * Had not been employed and were not actively looking for work. 
 
 4.3 
 
*.. .ttge. (bee Figure 2) 
 
 The new employees were slightly younger than those who had 
 been hired before the EDA loan. 
 
 The under-21 group had the lowest rate of former full-time em- 
 ployment ( 54% ) and the highest rates of former part-time (20% ) 
 .nd unemployed (12 % ,) workers. The unemployment rate decreas .& 
 as the age of the workers increases. 
 
 3. Race. (See Figure 3) 
 
 There were more Negroes among the new employees (20%) than 
 among employees hired before the date of the loan (10%). The formei 
 employment status of minority groups differed greatly from that of th 
 white employees. (See Table 1) 
 
 All of the minority groups had: a significantly larger proportic 
 of poverty-level incomes (71%) than did the whites (46%). 
 
 4. Education. (See Figure 4) 
 
 Most of the new employees had attended or graduated from high 
 school. The average number of years of school attendance was about 
 10. 
 
 5. Heads of Households. (See Figure 5) 
 
 About 55% of the new employees considered themselves heads 
 of households. Most of these had been previously employed full-time 
 (55%). About half of the non-heads- of households had been employed 
 
 4.4 
 
full time. However, about 75% of those who were not heads of house- 
 holds indicated that they contributed to the support of a family unit. 
 
 6. Length of Residence. (See Figure 6) 
 
 Most of the new employees were long-term residents of the county; 
 68 % had lived in the same town for more than nine years. Less than 
 40% of the new employees who had lived in the area less than one year 
 had poverty-level incomes, while 53% of the long term residents had 
 incomes of less than $4,000 a year. The long-term residents also 
 had the highest previous unemployment rate (12 % ), as compared with 
 6 % for the less -than- one -year group. 
 
 7. Skills. (See Figure 7) 
 
 Most of the new employees were operatives (63 % ) . An addi- 
 tional 33% were in four skill groups: craftsmen and foremen ( 12 %) 
 of the total), clerical workers ( 8% ), laborers ( 6 % ), and service 
 workers ( 5%). 
 
 Laborers and clerical workers had the lowest rates for former 
 full-time employment (56% and 61%) and the highest rates of part-time 
 employment (23% and 13%). The highest rates of previous unemploy- 
 ment were for operatives ( 13% ) and service workers (14%^. 
 
 4.5 
 
8. Holds Another Job. (See Figure 8) 
 
 Only 6 % of the new employees held another job. This group 
 included 16% of all former farm workers but only 2% of the formerly 
 unemployed. 
 
 9. Nearness to Job. (See Figure 9) 
 
 Over half of the new employees lived within six miles of their 
 job. Almost 3 6% of the full-time employees and 24% of the part-time 
 employees lived within three miles. Slightly more than 10% lived 
 more than 20 miles from their work. 
 
 C. Special Emphasis Groups 
 
 Some distinguishing characteristics of groups given special 
 emphasis in Chapter 5 were noted. 
 
 Over half (55%) of the 690 employees categorized as poor were 
 women. Women also make up a majority of the previously unemployed 
 (63%) and of the new labor force entrants (76%). Because of the large 
 number of women in the previously unemployed and new labor force 
 entrants groups, most of them also appear as non- heads of households. 
 
 More Negro employees are poor (28%) than are non-poor (12%). 
 
 Current part-time employees are more likely to be under 21 (34% 
 and to be employed as service workers (31%) or clerical employees 
 (26%). They are also more likely to be non-heads of households (77%). 
 
 A. 6 
 

 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 tK 
 
 E> 
 
 in 
 
 in 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 4-J 
 
 
 •«* 
 
 in 
 
 
 
 
 
 
 
 
 
 _^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 O 
 Oh 
 
 d 
 
 2 
 
 CI 
 
 CO 
 
 I s - 
 
 m 
 
 
 
 
 
 
 
 
 
 o 
 
 CO 
 
 
 
 (M 
 
 CO 
 
 
 
 
 
 
 
 
 
 o 
 
 
 CO 
 S3 
 
 t— " 
 
 «P 
 
 sO 
 
 ro 
 
 
 
 
 
 
 
 
 
 o 
 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 
 
 o 
 
 tP 
 
 
 
 
 
 
 
 
 <* 
 
 
 «; 
 
 
 d 
 
 ro 
 
 ■— < 
 
 
 
 
 
 
 
 
 
 ro 
 
 
 H 
 
 
 2 
 
 00 
 
 m 
 
 
 
 
 
 
 
 
 
 -< 
 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 a) 
 H 
 
 4a) 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 2 
 
 W 
 05 
 
 t£ 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 05 
 
 Pi 
 
 B 
 
 
 
 
 
 
 
 
 
 
 00 
 
 in 
 
 D 
 U 
 
 Oh 
 
 o 
 2 
 
 
 ■* 
 
 
 
 
 
 
 
 
 
 CO 
 
 cm 
 
 4) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 £ 
 
 ^ 
 
 CO 
 
 00 
 
 
 
 
 
 
 
 
 
 o 
 o 
 
 
 
 — t 
 
 
 ^o 
 
 ro 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 •— ■ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 vD 
 
 po 
 
 
 fc 
 
 
 
 Z 
 
 a 
 
 o 
 o 
 
 
 
 
 
 
 
 
 o 
 
 ro 
 
 
 
 CO 
 
 LTl | 
 
 
 
 — I I ■ ■ ,„■.,■—.,, , 
 
 
 
 
 
 
 ' — — ■■■ 
 
 s 
 
 &■ 
 
 
 ci 
 eg 
 
 
 
 
 ' 
 
 
 
 
 
 o 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 U, 
 
 
 
 2 
 
 
 (VJ 
 
 
 
 
 
 
 
 
 
 CO 
 
 o^ 
 
 cr- 
 
 t 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 r- 
 
 rn 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 ro 
 
 v£) 
 
 
 
 
 
 
 
 
 
 c 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <D 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 C 
 
 o 
 2 
 
 CO 
 
 r\] 
 
 
 
 
 
 
 
 
 
 in 
 
 , , 
 
 
 2 
 
 U"l 
 
 O^ 
 
 
 
 
 
 
 
 
 
 ■<r 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 to 
 
 W 
 
 tP 
 
 
 
 
 
 
 " 
 
 
 
 
 
 o 
 o 
 
 
 
 z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 d 
 
 o> 
 
 _ 
 
 
 
 
 
 
 
 
 
 o 
 
 rsl 
 
 H 
 
 
 z 
 
 m 
 
 m 
 
 
 
 
 
 
 
 
 
 — • 
 
 
 CO 
 
 05 
 
 
 
 
 "-' 
 
 
 
 
 
 
 
 
 
 
 
 <u 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 _s 
 
 wO 
 
 0^ 
 
 , 
 
 
 
 
 
 
 
 
 
 o 
 
 
 »-i 
 
 
 &- 
 
 in 
 
 "Sf 
 
 
 
 
 
 
 
 
 
 o 
 
 
 05 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 A 
 
 4-1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 (4 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 rt 
 
 d 
 Z 
 
 <*• 
 
 (M 
 
 
 
 
 
 
 
 
 
 vO 
 
 ro 
 
 
 PL, 
 
 r- 
 
 in 
 
 
 
 
 
 
 
 
 
 fM 
 
 — ' 
 
 (0 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 E 
 
 wO 
 
 ro 
 
 t- 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 &- 
 
 r^ 
 
 <\J 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 — 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 •—1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 • 
 
 * 
 
 s 
 
 
 
 
 
 
 
 
 
 rn 
 
 r- 
 
 
 Cn 
 
 o 
 
 •£> 
 
 ^r 
 
 
 
 
 
 
 
 
 
 ~-4 
 
 <M 
 
 
 2 
 
 O 
 
 (M 
 
 
 
 
 
 
 
 
 
 O^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 C J3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O fd 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ W 
 
 X 
 
 
 C 
 
 
 
 
 
 
 
 
 
 J 
 
 2 £ 
 
 
 
 
 
 «J 
 
 
 
 
 
 
 
 
 
 <q 
 
 
 
 
 c 
 
 £ 
 
 o 
 
 
 
 
 
 
 
 
 
 H 
 
 O 
 
 4J 
 
 
 
 
 s 
 
 ^____ 
 
 > 
 
 
 
 
 
 
 
 
 
 H 
 
 
 4.7 
 

 
 
 r^ 
 
 NO 
 
 in 
 
 ro 
 
 l-H 
 
 
 
 
 
 
 
 o 
 
 — _. 
 
 
 >, 
 
 6> 
 
 l-H 
 
 «— 1 
 
 i— i 
 
 
 — 
 
 03 
 
 T- 
 
 vO 
 
 ■* 
 
 —* 
 
 
 o 
 
 
 
 M 
 > 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ~- 
 
 
 
 o 
 Z 
 
 LTl 
 
 | 
 
 
 oo 
 
 vO 
 
 0^ 
 
 
 00 
 
 r> 
 
 T- 
 
 CO 
 
 o 
 
 CM 
 
 
 o 
 
 i — i 
 
 ~* 
 
 o 
 
 00 
 
 r- 
 
 m 
 
 vO 
 
 CO 
 
 M 
 
 
 
 c> 
 
 
 
 k 
 
 1 ' 
 
 "" , 
 
 ~~ 
 
 
 
 
 
 
 
 
 
 vD 
 
 
 
 
 LD 
 
 00 
 
 vO 
 
 ro 
 
 o 
 
 0^ 
 
 CT^ 
 
 vD 
 
 ■* 
 
 ^H 
 
 
 O 
 
 
 
 •—I 
 
 t£ 
 
 1—1 
 
 "■* 
 
 •"• 
 
 — ' 
 
 — i 
 
 
 
 
 
 
 
 o 
 
 
 co 
 
 rt 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 p 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 r» 
 
 H 
 
 H 
 
 
 r- 
 
 O- 
 
 LTl 
 
 -0 
 
 ro 
 
 CO 
 
 CO 
 
 oo 
 
 r~ 
 
 co 
 
 "* 
 
 <* 
 
 ^O 
 
 < 
 
 
 
 Z 
 
 o 
 
 co 
 
 «— < 
 
 r- 
 
 CO 
 
 —4 
 
 ^H 
 
 r- 
 
 in 
 
 ~h 
 
 
 co 
 
 
 
 *- H 
 
 CM 
 
 M 
 
 — < 
 
 "~ l 
 
 "* 
 
 ""' 
 
 
 
 
 
 — ' 
 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 —w 
 
 
 
 
 
 
 H 
 Z 
 
 <u 
 
 
 ^ 
 
 O 
 
 CO 
 
 m 
 
 m 
 
 ro 
 
 in 
 
 00 
 
 in 
 
 CO 
 
 CO 
 
 o 
 
 
 £ 
 
 l£ 
 
 co 
 
 — ' 
 
 rt 
 
 
 
 
 
 
 
 
 
 o 
 
 
 w 
 
 PS 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 LTl 
 
 K 
 
 ^ 
 
 . 
 
 CO 
 
 vO 
 
 in 
 
 (VJ 
 
 rg 
 
 — ' 
 
 M 
 
 CO 
 
 rg 
 
 1-1 
 
 rt 
 
 00 
 CO 
 
 ro 
 
 D 
 
 rt 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 U 
 
 P. 
 
 Z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <v 
 
 
 -* 
 
 00 
 
 vO 
 
 co 
 
 o 
 
 O 
 
 o> 
 
 sD 
 
 ■^ 
 
 ,_, 
 
 
 o 
 
 
 
 § 
 
 s£ 
 
 "" ' 
 
 - " H 
 
 ■" ' 
 
 ■-* 
 
 -" 
 
 
 
 
 
 
 
 o 
 
 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 in 
 
 t 
 
 
 d 
 
 o 
 Z 
 
 "# 
 
 co 
 
 o 
 
 't 
 
 ,-H 
 
 M 
 
 ^h 
 
 in 
 
 in 
 
 r~ 
 
 CO 
 
 o 
 
 (V} 
 
 
 fa 
 
 00 
 
 ro 
 
 ~-i 
 
 r- 
 
 co 
 
 , — i 
 
 ^H 
 
 r~ 
 
 m 
 
 p-i 
 
 
 
 
 
 
 — 
 
 N 
 
 CJ 
 
 — 
 
 — 
 
 — — i 
 
 — « 
 
 
 
 
 
 mA 
 
 
 
 
 
 LTl 
 
 m 
 
 in 
 
 •<? 
 
 vD 
 
 co 
 
 00 
 
 r- 
 
 co 
 
 CO 
 
 
 o 
 
 
 
 s 
 
 Pi 
 
 t> 
 
 ~~* 
 
 " 
 
 —' 
 
 ~* 
 
 — ' 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CT~ 
 
 
 h 
 
 o 
 2 
 
 ■* 
 
 -* 
 
 ^ 
 
 CO 
 
 m 
 
 CO 
 
 r~ 
 
 vO 
 
 CO 
 
 co 
 
 
 CO 
 
 
 
 i 
 
 O 
 
 ^ 
 
 r~ 
 
 co 
 
 rg 
 
 T-t 
 
 -o 
 
 00 
 
 oo 
 
 « 
 
 in 
 
 
 
 o 
 
 
 
 C XI 
 4) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 — < 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 . 
 
 
 • 
 
 ■* 
 
 co 
 
 00 
 
 _ 
 
 oo 
 
 •— i 
 
 (M 
 
 —i 
 
 r~ 
 
 O 
 
 O 
 
 m 
 
 
 
 D 
 
 o 
 Z 
 
 rg 
 
 co 
 
 ^~* 
 
 (M 
 
 
 *~* 
 
 •■^ 
 
 ~~* 
 
 
 
 
 rf 
 
 
 
 
 r- 
 
 r- 
 
 rg 
 
 rg 
 
 a- 
 
 r- 
 
 r~ 
 
 m 
 
 PO 
 
 
 
 o 
 
 
 co 
 
 W 
 
 6- 
 
 --< 
 
 —i 
 
 p-i 
 
 —i 
 
 
 ^< 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ~"* 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <; 
 
 Z 
 
 
 
 r^- 
 
 00 
 
 O 
 
 o» 
 
 in 
 
 r^ 
 
 _ 
 
 oo 
 
 m 
 
 ~-i 
 
 O 
 
 o 
 
 h 
 
 CO 
 K 
 
 
 Z 
 
 ro 
 
 (M 
 
 1—1 
 
 '"- 
 
 "~* 
 
 rsl 
 
 ~~* 
 
 
 
 
 
 •x> 
 
 
 <b 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 E 
 
 
 (V) 
 
 r- 
 
 "* 
 
 O 
 
 _ 
 
 in 
 
 m 
 
 co 
 
 rg 
 
 (M 
 
 — 
 
 o 
 
 
 •— i 
 
 
 fc£ 
 
 CO 
 
 _< 
 
 ^-« 
 
 
 fm4 
 
 
 
 
 
 
 
 o 
 
 
 « 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 — 1 
 
 
 Oh 
 
 .,_. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 • 
 
 O 
 
 _ 
 
 co 
 
 _^ 
 
 ■* 
 
 v£l 
 
 vD 
 
 "J< 
 
 CM 
 
 (M 
 
 — i 
 
 m 
 
 "T 
 
 
 Oh 
 
 o 
 Z 
 
 <* 
 
 <N 
 
 ■""* 
 
 ~~ ' 
 
 "^ 
 
 
 
 
 
 
 
 r%] 
 
 
 01 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6 
 
 s£ 
 
 rg 
 
 r- 
 
 r- 
 
 •*}* 
 
 o 
 
 00 
 
 o 
 
 -JD 
 
 in 
 
 CM 
 
 
 o 
 o 
 
 
 
 h 
 
 •— * 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 
 >o 
 
 r- 
 
 o 
 
 in 
 
 O 
 
 r> 
 
 Tf 
 
 in 
 
 co 
 
 in 
 
 CO 
 
 CO 
 
 rg 
 
 I 
 
 o 
 2 
 
 o 
 
 LTl 
 
 vO 
 
 (VJ 
 
 0> 
 
 ^0 
 
 00 
 
 m 
 
 Tf 
 
 
 
 cr 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 U 
 
 
 C -C 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 0) 
 
 
 rS 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <; o 
 
 W 
 
 rg 
 
 Pi 
 
 ■* 
 
 cr^ 
 
 ■<t 
 
 o~- 
 
 •tf 
 
 o 
 
 •tf 
 
 a- 
 
 Tj< 
 
 o 
 
 X 
 
 c: 
 
 < 
 
 a rt 
 
 P > 
 
 
 
 
 4; 
 
 rg 
 
 00 
 
 ro 
 
 ro 
 
 -* 1 
 
 ■* 
 
 in 
 
 in 
 
 sD 
 
 rt 
 
 H 
 
 O 4^ 
 
 
 
 
 X) 
 
 c 
 
 i 
 
 i 
 
 in 
 
 i 
 
 o 
 
 1 
 
 m 
 
 i 
 
 o 
 
 1 
 
 in 
 
 i 
 
 o 
 
 i 
 m 
 
 1 
 
 O 
 
 in 
 
 O 
 H 
 
 ^ 
 
 — 
 
 
 
 JD 
 
 CO 
 
 (NJ 
 
 to 
 
 (O 
 
 ~*$' 
 
 N 1 
 
 in 
 
 m 
 
 v£> 
 
 -X) 
 
 K-H .-i 
 
 4.8 
 
CURRENT STATUS 
 
 >s 
 
 M 
 
 4) 
 
 > 
 
 o 
 
 p. 
 
 ^ 
 
 
 CO 
 
 en 
 
 - 
 
 
 
 
 
 
 
 o 
 o 
 
 
 d 
 2 
 
 
 (M 
 
 CO 
 
 vO 
 
 o 
 
 
 
 
 
 
 co 
 
 
 d 
 
 o 
 
 ^ 
 
 
 o 
 
 IN] 
 
 CM 
 
 ~ 
 
 
 
 
 
 
 
 o 
 o 
 
 
 d 
 
 Z 
 
 o 
 
 o 
 
 00 
 P0 
 
 o 
 
 o 
 
 
 
 
 
 
 en 
 en 
 
 
 0) 
 
 g 
 
 H 
 
 Ph 
 
 t£ 
 
 LTl 
 
 
 oo 
 
 
 
 
 
 
 
 
 o 
 o 
 
 
 d 
 Z 
 
 
 o 
 
 en 
 
 o 
 
 o 
 
 
 
 
 
 
 
 (M 
 
 6 
 
 ^ 
 
 
 00 
 
 M 
 
 - 
 
 
 
 
 
 
 
 o 
 o 
 
 
 d 
 Z 
 
 m 
 
 o 
 o 
 
 O 
 INI 
 
 m 
 
 o 
 
 o 
 
 
 
 
 
 
 o 
 o 
 
 en 
 
 CT- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 H 
 
 CO 
 
 « 
 
 o 
 
 « 
 
 E 
 
 h 
 
 & 
 
 en 
 
 in 
 
 m 
 
 o 
 
 - 
 
 
 
 
 
 
 
 o 
 o 
 
 
 d 
 2 
 
 o^ 
 ^f 
 
 in 
 
 CO 
 
 co 
 
 - 
 
 o 
 
 
 
 
 
 
 en 
 
 a 
 
 i 
 
 >. 
 
 
 t— i 
 
 1) 
 
 c 
 
 ^ 
 
 o 
 
 00 
 
 in 
 
 m 
 
 
 
 
 
 
 
 
 O 
 
 o 
 
 
 d 
 2 
 
 en 
 
 (Ni 
 
 (N) 
 
 r^ 
 
 o 
 
 o 
 
 
 
 
 
 
 (VJ 
 
 *f 
 
 W 
 2 
 
 j£ 
 
 00 
 O0 
 
 o 
 
 - 
 
 - 
 
 
 
 
 
 
 
 o 
 o 
 
 
 d 
 Z 
 
 o 
 
 vD 
 
 (VJ 
 
 (Nl 
 
 o 
 
 
 
 
 
 
 o 
 
 vO 
 
 uf. 
 
 s 
 
 Ph 
 
 t£ 
 
 T 
 -o 
 
 en 
 
 OJ 
 
 
 
 
 
 
 
 
 c 
 o 
 
 
 d 
 2 
 
 00 
 
 
 m 
 
 o 
 
 o 
 
 
 
 
 
 
 <M 
 
 ro 
 
 £ 
 
 •— * 
 
 "3 
 
 ^ 
 
 r- 
 
 (M 
 
 (VI 
 
 - 
 
 
 
 
 
 
 
 o 
 o 
 
 
 d 
 2 
 
 
 O 
 
 vO 
 
 co 
 
 o 
 
 
 
 
 
 
 o 
 o^ 
 
 en 
 
 « <; u W 
 
 
 O 
 
 u 
 
 bo 
 
 z 
 
 c 
 
 T) 
 C 
 t— 1 
 
 c 
 x; o 
 
 C 0) 
 
 a, c 
 co *^ 
 
 O 
 
 
 
 
 
 
 <, 
 
 H 
 O 
 H 
 
 01 
 
 c 3 
 
 O tf 
 
 4-» .-J 
 
 £ > 
 t. *< 
 O *_. 
 
 c •>' 
 1 — t H- 
 
 
 
 
 
 
 
 I 
 
 i.9 
 
 
 
 
 
 
 
 
Table 1 
 
 Former Employment Status 
 
 
 Unemployed 
 
 Part-Time 
 
 Fa: 
 
 rm Worker 
 
 White 
 
 11% 
 
 8% 
 
 
 5% 
 
 Negro 
 
 8% 
 
 16% 
 
 
 13% 
 
 Indians 
 
 25% 
 
 11 % 
 
 
 29% 
 
 * Percentage given is percent of total racial 
 groups (i.e. , white) of new employees. 
 
 4.10 
 

 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 
 
 >N 
 
 6^ 
 
 ,H 
 
 ro 
 
 00 
 
 <M 
 
 oo 
 
 r- 
 
 (-H 
 
 
 
 
 o 
 
 
 
 4-> 
 
 
 
 ■— ' 
 
 
 ro 
 
 ro 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 
 
 Z 
 
 ro 
 
 o^ 
 
 <* 
 
 CO 
 
 s 
 
 0^ 
 
 r- 
 
 
 
 
 o 
 
 M 
 
 
 
 
 
 co 
 
 in 
 
 ,—4 
 
 m 
 
 ■<}< 
 
 
 
 
 
 <J^ 
 
 t 
 
 
 fc 
 
 
 
 
 (VJ 
 
 (M 
 
 
 
 
 
 
 vT) 
 
 
 
 
 „ 
 
 - 
 
 r~ 
 
 ,_ 
 
 t* 
 
 o 
 
 (M 
 
 ( , 
 
 
 
 o 
 
 
 
 •— < 
 
 t£ 
 
 
 —' 
 
 
 ro 
 
 ro 
 
 •-* 
 
 
 
 
 
 o 
 
 
 to 
 
 rt 
 ■*-» 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 
 H 
 
 
 r- 
 
 t 
 
 o 
 
 O^ 
 
 n< 
 
 *-+ 
 
 ro 
 
 o 
 
 
 
 rf 
 
 -O 
 
 < 
 H 
 CO 
 
 H 
 Z 
 
 o 
 Z 
 
 
 in 
 
 o 
 
 ~-< 
 
 r> 
 
 ro 
 
 oJ 
 
 r* 
 
 
 
 ■* 
 
 vO 
 
 
 
 ^* 
 
 1—1 
 
 <tf" 
 
 Tt< 
 
 ■" ' 
 
 
 
 
 
 ro 
 
 
 0) 
 
 
 
 LTl 
 
 CO 
 
 (M 
 
 a- 
 
 ro 
 
 CO 
 
 
 
 
 O 
 
 
 g 
 
 *P 
 
 
 
 
 ro 
 
 M 
 
 — ' 
 
 — ' 
 
 
 
 
 o 
 
 
 W 
 
 H 
 *-> 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 (M 
 
 ro 
 
 (M 
 
 _. 
 
 m 
 
 m 
 
 o 
 
 
 
 00 
 
 m 
 
 tn 
 
 • 
 
 
 
 
 
 -- 
 
 
 
 
 
 
 ro 
 
 (VJ 
 
 ID 
 
 nJ 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 u 
 
 Oh 
 
 Z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 
 . 
 
 rg 
 
 r^ 
 
 , , 
 
 r- 
 
 o 
 
 , 
 
 ^H 
 
 
 
 O 
 
 
 
 § 
 
 ^ 
 
 
 ■"• 
 
 
 ro 
 
 ro 
 
 "•* 
 
 
 
 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 •— * 
 
 
 
 
 
 
 
 
 
 
 
 
 xD 
 
 
 
 3 
 
 o 
 Z 
 
 r- 
 
 N 
 
 r- 
 
 r~ 
 
 ro 
 
 vO 
 
 oo 
 
 v£> 
 
 
 
 O 
 
 ro 
 
 
 fa 
 
 
 LTl 
 
 a- 
 
 o 
 
 CO 
 
 rM 
 
 — * 
 
 1 — 1 
 
 
 
 ro 
 
 (VJ 
 
 
 
 
 
 
 ^ 
 
 T 
 
 — < 
 
 
 
 
 
 — 
 
 
 
 
 
 •s< 
 
 m 
 
 
 r^ 
 
 0^ 
 
 ^ 
 
 
 
 
 
 O 
 
 
 
 E 
 (« 
 
 ^ 
 
 
 i— i 
 
 ■" H 
 
 ro 
 
 rg 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 fa 
 
 o 
 Z 
 
 i< 
 
 ■^ 
 
 o 
 
 ro 
 
 
 ^ 
 
 o 
 
 o 
 
 
 
 ro 
 
 9\ 
 
 
 ■ 
 
 o 
 
 & 
 
 
 ro 
 
 O 
 
 a* 
 
 O 
 
 CO 
 
 
 
 
 
 O 
 
 
 
 
 •M4 
 
 
 (SJ 
 
 -* 
 
 
 
 
 
 
 o 
 
 
 
 Q. 
 
 
 
 
 
 
 
 
 
 
 — 
 
 
 
 (-. w 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ~ o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 c 
 D 
 
 • 
 
 o 
 
 0^ 
 
 ro 
 
 (VJ 
 
 CO 
 
 pg 
 
 ^ 
 
 o 
 
 
 
 in 
 
 • — i 
 
 
 o 
 Z 
 
 
 *~* 
 
 •"■* 
 
 >tf 
 
 in 
 
 ~^ 
 
 
 
 
 
 ^ 
 
 
 
 
 
 o- 
 
 -O 
 
 
 CO 
 
 s 
 
 vO 
 
 
 
 
 o 
 
 
 CO 
 
 W 
 
 £? 
 
 
 
 
 ro 
 
 ro 
 
 
 
 
 
 
 o 
 
 
 
 fa 
 
 
 
 
 
 
 
 
 
 
 
 
 1 ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <; 
 
 Z 
 
 o 
 
 o 
 
 in 
 
 O 
 
 0^ 
 
 ' (M 
 
 m 
 
 o^ 
 
 _ 
 
 
 
 o 
 
 m 
 
 H 
 
 
 2 
 
 
 »-« 
 
 
 "tf 
 
 O 
 
 — i 
 
 
 
 
 
 ^j0 
 
 ro 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 — ' 
 
 
 o> 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 £ 
 
 
 
 m 
 
 r- 
 
 O 
 
 sO 
 
 o 
 
 (M 
 
 Tf 
 
 
 O 
 
 
 n 
 
 
 6° 
 
 
 
 
 ro 
 
 ro 
 
 .—I 
 
 
 
 
 
 o 
 
 
 Ph 
 
 
 
 
 
 
 
 
 
 
 
 
 
 f--4 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 fa 
 
 . 
 
 o 
 
 sO 
 
 cr- 
 
 CO 
 
 O 
 
 o 
 
 fM 
 
 m 
 
 
 
 ^o 
 
 ro 
 
 
 
 
 2 
 
 
 
 
 ro 
 
 "* 
 
 (NJ 
 
 
 
 
 
 (NJ 
 
 *""* 
 
 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 g 
 
 
 ~H 
 
 rj 
 
 CO 
 
 M 
 
 O 
 
 a~ 
 
 ^H 
 
 ^H 
 
 
 
 o 
 
 
 
 s£ 
 
 
 —i 
 
 
 ro 
 
 ro 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 "" 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 fa 
 
 
 r- 
 
 «* 
 
 0^ 
 
 O 
 
 00 
 
 ^ 
 
 ._ 
 
 o 
 
 
 
 ro 
 
 r~ 
 
 
 o 
 
 
 *"4 
 
 gs 
 
 o^ 
 
 (M 
 
 co 
 
 |MH 
 
 *H 
 
 
 
 r^t 
 
 (M 
 
 
 
 
 ' — 1 
 
 
 n 
 
 ro 
 
 
 
 
 
 
 CT^ 
 
 
 
 
 (— 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 TJ 
 
 
 
 j: 
 
 JT 
 
 
 
 ^ 
 
 
 
 
 
 
 
 z 
 
 C 
 
 
 x: 
 u 
 CO 
 
 oo 
 X 
 
 QO 
 
 
 
 o 
 
 
 
 
 0) 
 C £> 
 
 
 
 o 
 
 0) 
 
 4»J 
 
 u 
 
 
 T> 
 
 
 T3 
 
 
 
 
 O rt 
 
 D 
 Q 
 
 <; 
 u 
 
 
 4-1 
 
 <; 
 
 ^ o 
 u o 
 
 •* .c 
 
 0) (j 
 
 T 5 
 
 o £ 
 
 2 ^ 
 
 5 "J 
 
 rt £ 
 -a £ 
 
 rj id 
 
 U U 
 
 T3 
 
 « 2 
 
 4-» 1 
 
 (U 
 
 ^ ° 
 
 c aj 
 0) — J 
 
 ti o 
 
 3 on 
 -a « 
 
 rj ^4 
 
 ^ a 
 
 4-) 
 
 (t) 
 
 3 
 
 
 
 H 
 
 O 
 
 rj rt 
 *-. 
 
 W 
 
 
 
 Z co 
 
 <; o 
 
 O O 
 
 <; to 
 
 O co 
 
 <; o 
 
 O <J 
 
 
 
 
 
 H 
 
 
 U — 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4.11 
 

 
 
 
 vO 
 
 "* 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 >> 
 
 ^ 
 
 iO 
 
 ■<* 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ""• 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 o 
 
 PH 
 
 
 
 Z 
 
 00 
 
 oo 
 
 (M 
 
 
 
 
 
 
 
 
 
 in 
 
 00 
 
 
 
 
 r- 
 
 CO 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 ,_, 
 
 ^ 
 
 in 
 
 f 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 to 
 
 rt 
 
 
 
 
 
 
 
 
 
 
 
 
 ^"* 
 
 
 
 D 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 ■ 
 
 
 
 H 
 
 H 
 
 # 
 
 o-^ 
 
 PJ 
 
 
 
 
 
 
 
 
 
 o 
 
 0> 
 
 
 < 
 
 
 o 
 
 co 
 
 NO 
 
 
 
 
 
 
 
 
 
 CO 
 
 o 
 
 
 H 
 to 
 
 
 Z 
 
 r- 
 
 LTl 
 
 
 
 
 
 
 
 
 
 — ' 
 
 — ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 Z 
 
 E 
 
 
 —i 
 
 r- 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 ^ 
 
 CI 
 
 r- 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CS 
 
 Ih 
 
 
 — 
 
 •>* 
 
 
 
 
 
 
 
 
 
 m 
 
 00 
 
 
 
 rd 
 
 d 
 
 * H 
 
 pj 
 
 
 
 
 
 
 
 
 
 p-> 
 
 PJ 
 
 
 Ph 
 
 Z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4) 
 
 
 00 
 
 ro 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 ^ 
 
 LTl 
 
 t*< 
 
 
 
 
 
 
 
 • 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 d 
 
 00 
 
 00 
 
 
 
 
 
 
 
 
 
 PJ 
 
 ro 
 
 
 
 Z 
 
 
 r<1 
 
 in 
 
 . ■ — 
 
 
 
 
 
 
 
 
 
 
 ^° 
 
 
 
 
 ^ 
 
 
 CT- 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 S 
 
 r- 
 
 fM 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 ft) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 k 
 
 d 
 Z 
 
 
 PJ 
 
 
 
 
 
 
 
 
 
 o 
 
 PJ 
 
 If, 
 
 i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 sS 
 
 «o 
 
 rr 
 
 
 
 
 
 
 
 
 
 o 
 
 
 f 
 
 
 w—i 
 
 (■^ 
 
 nO 
 
 
 
 
 
 
 
 
 
 o 
 
 i 
 
 y. 
 
 
 0) 
 
 
 
 
 
 
 
 
 
 ] 
 
 
 1 — 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 t,ii 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 "* 
 
 
 c 
 
 d 
 Z 
 
 ,— I 
 
 <?• 
 
 
 
 
 
 
 
 
 
 o 
 
 -0 
 
 Ua 
 
 co 
 D 
 H 
 
 D 
 
 in 
 
 oo 
 
 
 
 
 
 
 
 
 
 Tf 
 
 
 - 
 
 W 
 
 & 
 
 (V) 
 
 
 
 
 
 
 
 
 
 
 o 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 — ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <: 
 
 z 
 
 d 
 
 vO 
 
 00 
 
 
 
 
 
 
 
 
 
 >* 
 
 _ 
 
 
 H 
 
 
 z 
 
 ro 
 
 —i 
 
 
 
 
 
 
 
 
 
 in 
 
 CO 
 
 
 CO 
 
 
 
 
 "*■' 
 
 
 
 
 
 
 
 
 
 —* 
 
 
 
 a 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 £ 
 
 
 ■* 
 
 -XI 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 1— t 
 
 
 &- 
 
 ■V 
 
 in 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 K 
 
 h 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ou 
 
 •i-i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 u 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 R) 
 
 d 
 Z 
 
 ■<J" 
 
 00 
 
 
 
 
 
 
 
 
 
 CM 
 
 f» 
 
 
 
 Ph 
 
 LTl 
 
 vO 
 
 
 
 
 
 
 
 
 
 (M 
 
 ^ 
 
 
 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 E 
 
 
 oo 
 
 pj 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 6> 
 
 sO 
 
 C"> 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ~* 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 "a 
 
 . 
 
 00 
 
 r~ 
 
 
 
 
 
 
 
 
 
 m 
 
 in 
 
 
 
 o 
 
 o 
 
 oo 
 
 
 
 
 
 
 
 
 
 oo 
 
 m 
 
 
 
 2 
 
 LTl 
 
 PJ 
 
 
 
 
 
 
 
 
 
 00 
 
 
 
 
 to 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Q 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 tn 
 
 J 
 
 
 
 
 
 
 
 
 
 
 
 
 <u 
 
 
 o 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 C .O 
 
 
 to 
 
 s 
 
 tn 
 
 
 
 
 
 
 
 
 
 
 
 _d 
 
 
 Q 
 
 w 
 
 OJ 
 
 o 
 
 
 
 
 
 
 
 
 
 
 *j •-* 
 
 
 w 
 
 X 
 
 CO 
 
 D 
 
 O 
 
 E 
 
 >* 
 
 Z 
 
 
 
 
 
 
 
 
 
 < 
 
 H 
 O 
 
 E > 
 
 u * 
 
 O 4-. 
 
 
 ■- — %- 
 
 
 
 
 
 
 
 
 
 
 1 
 
 H 
 
 c ^^ 
 
 A. 12 
 
CURRENT STATUS 
 
 4-> 
 
 u 
 
 > 
 
 o 
 
 t£ 
 
 vD 
 
 CO 
 
 ^ 
 
 CO 
 
 CO 
 
 CO 
 
 ~ 
 
 eg 
 
 eg 
 
 eg 
 f- 
 
 o 
 : ° 
 
 
 d 
 
 2 
 
 
 eg 
 
 eg 
 
 rg 
 
 CO 
 
 CO 
 
 eg 
 
 O 
 
 eg 
 
 eg 
 
 co 
 
 cr- 
 
 CO 
 
 co 
 
 t— < 
 rt 
 
 o 
 H 
 
 ^ 
 
 00 
 
 co 
 
 in 
 
 •* 
 
 CO 
 
 CO 
 
 - 
 
 eg 
 
 eg 
 
 0^ 
 
 O 
 
 o 
 
 
 
 z 
 
 o 
 
 eg 
 
 in 
 
 rt" 
 
 in 
 
 00 
 co 
 
 rf 
 
 00 
 
 eg 
 
 eg 
 
 o~- 
 
 -JO 
 
 CO 
 CO 
 
 r- 
 
 B 
 
 ^ 
 
 m 
 
 v£> 
 
 
 ^O 
 
 - 
 
 CO 
 
 CO 
 
 CO 
 
 vO 
 
 vO 
 
 O 
 O 
 
 
 o 
 Z 
 
 - 1 
 
 (NJ 
 
 o 
 
 eg 
 
 ■* 
 
 - 
 
 - 1 
 
 — 
 
 eg 
 
 eg 
 eg 
 
 CO 
 
 r- 
 eg 
 
 B 
 
 H 
 
 f — i 
 a 
 
 ^ 
 
 CO 
 
 CO 
 
 in 
 
 ■* 
 
 CO 
 
 co 
 
 ~ 
 
 eg 
 
 eg 
 
 a-- 
 
 O 
 
 o 
 
 
 d 
 Z 
 
 CO 
 
 o 
 
 O 
 
 m 
 
 rg 
 
 in 
 
 CO 
 
 CO 
 
 r- 
 
 o 
 eg 
 
 eg 
 
 CO 
 
 o 
 o 
 
 CO 
 
 rg 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 PRIOR STATUS 
 
 S 
 u 
 
 s£ 
 
 iTl 
 
 - 
 
 eg 
 
 eg 
 
 CO 
 
 ■* 
 
 
 
 - 
 
 eg 
 
 GO 
 
 o 
 o 
 
 --. i 
 
 eg| 
 
 i 
 
 d 
 Z 
 
 iTl 
 
 - 
 
 eg 
 
 eg 
 
 CO 
 
 ■* 
 
 o 
 
 o 
 
 - 
 
 co 
 
 o-- 
 
 Uner iploy- 
 
 e :l 
 
 s£ 
 
 ■* 
 
 ■ 
 
 CO 
 
 eg 
 
 ~ 
 
 N 
 
 CO 
 
 - 
 
 eg 
 
 
 o 
 o 
 
 in 
 
 d 
 Z 
 
 •~o 
 
 r- 
 
 ■* 
 
 CO 
 
 eg 
 
 CO 
 
 ^ 
 
 - 
 
 CO 
 
 eg 
 
 W 
 
 h 
 
 Z 
 
 ^ 
 
 CO 
 
 ^ 
 
 CO 
 
 in 
 
 t 
 
 CO 
 
 eg 
 
 CO 
 
 ** 
 
 
 o 
 o 
 
 d 
 Z 
 
 CO 
 
 r- 
 
 in 
 
 oo 
 
 ^o 
 
 in 
 
 CO 
 
 m 
 
 r~ 
 
 o 
 
 o 
 
 V 
 
 .§ 
 H 
 
 u 
 
 a 
 Ph 
 
 ^ 
 
 00 
 
 tM 
 
 CO 
 
 in 
 
 eg 
 
 eg 
 
 eg 
 
 eg 
 
 CO 
 
 r- 
 
 o 
 
 d 
 Z 
 
 o 
 
 CO 
 
 ^ 
 
 v£) 
 
 eg 
 
 CO 
 
 eg 
 
 co 
 
 ^r 
 
 co 
 co 
 
 in 
 eg 
 
 6 
 
 H 
 
 ■— i 
 
 P— 4 
 
 3 
 
 t£ 
 
 O"^ 
 
 CO 
 
 -£> 
 
 *f 
 
 CO 
 
 rf 
 
 -* 
 
 -« 
 
 - 
 
 CO 
 ~J3 
 
 o 
 
 d 
 2 
 
 o 
 oo 
 
 vO 
 
 eg 
 
 rg 
 in 
 
 r- 
 co 
 
 00 
 
 eg 
 
 CO 
 
 co 
 
 0> 
 
 eo 
 
 eg 
 
 CO 
 
 r- 
 o 
 
 CO 
 
 LENGTH OF 
 RESIDENCE 
 
 Less than one 
 year 
 
 U 
 
 a 
 
 V 
 
 <u 
 c 
 
 O 
 
 en 
 u 
 
 rt 
 
 aj 
 >> 
 o 
 
 tn 
 U 
 
 rt 
 
 4) 
 1) 
 U 
 
 s: 
 H 
 
 01 
 
 u 
 
 rt 
 
 0) 
 
 u 
 
 3 
 O 
 
 to 
 u 
 
 rt 
 
 <u 
 
 > 
 
 in 
 U 
 
 rt 
 
 X 
 CO 
 
 tn 
 M 
 
 rt 
 
 <L> 
 fH 
 
 c 
 
 0) 
 
 > 
 
 L0 
 
 in 
 
 u 
 
 rt 
 <v 
 
 >H 
 
 -C 
 GO 
 
 W 
 
 1 
 Nine Years or 
 M ore 
 
 <, 
 H 
 O 
 H 
 
 c _o 
 
 O rf 
 
 rt rt 
 
 E < 
 
 O *j 
 o 
 
 c • • 
 
 4.13 
 

 ^ 
 
 s£ 
 
 
 
 
 00 
 
 
 
 ** 
 r^ 
 
 m 
 
 
 o 
 
 o 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 —4 
 
 
 
 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 
 
 2 
 
 — i 
 
 o 
 
 —1 
 
 ON 
 
 — . 
 
 s0 
 
 — . 
 
 m 
 
 — , 
 
 m 
 
 o 
 
 (Nl 
 
 
 
 
 
 
 
 Tf 
 
 
 <vl 
 
 r- 
 
 CO 
 
 
 in 
 
 t 
 
 s 
 
 
 p< 
 
 
 
 
 
 
 
 •* 
 
 
 
 
 vO 
 
 
 
 
 _ 
 
 
 _, 
 
 00 
 
 
 (Nl 
 
 ^r 
 
 v£> 
 
 
 o 
 
 O 
 
 
 
 f— i 
 
 ^ 
 
 
 
 
 
 
 '- 
 
 vO 
 
 
 
 
 O 
 
 
 CO 
 
 D 
 
 nJ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 
 H 
 
 
 vO 
 
 o 
 
 o 
 
 0> 
 
 in 
 
 *o 
 
 o 
 
 m 
 
 m 
 
 m 
 
 s£> 
 
 "* 
 
 
 
 o 
 2 
 
 
 
 
 o 
 
 
 in 
 
 00 
 
 I s - 
 
 
 00 
 
 O 
 CO 
 
 o 
 
 to 
 
 H 
 2 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 0) 
 
 
 cl 
 
 
 ro 
 
 vD 
 
 
 en 
 
 vO 
 
 , ( 
 
 
 s 
 
 O 
 
 
 £ 
 
 ^ 
 
 
 
 
 (NJ 
 
 
 
 (NJ 
 
 CO 
 
 
 
 o 
 
 
 W 
 
 H 
 
 4-1 
 
 In 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o* 
 
 -H 
 
 o 
 
 ~ 
 
 CT> 
 
 o 
 
 ~ 
 
 o^ 
 
 _ 
 
 o 
 
 CO 
 
 in 
 
 CO 
 
 00 
 (Nl 
 
 u 
 
 (X 
 
 2 
 
 
 
 
 
 
 
 
 
 
 
 
 
 0) 
 
 
 _^ 
 
 
 _^ 
 
 00 
 
 
 (Nl 
 
 vO 
 
 in 
 
 
 sO 
 
 o 
 
 
 
 E 
 
 ^ 
 
 
 
 
 
 
 "- 1 
 
 sO 
 
 
 
 
 o 
 
 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .— < 
 
 
 
 
 
 
 
 
 
 
 
 
 (■H 
 
 
 
 3 
 
 o 
 2 
 
 in 
 
 o 
 
 o^ 
 
 O 
 
 in 
 
 in 
 
 O 
 
 (Nl 
 
 m 
 
 o 
 
 r- 
 
 oo 
 
 
 tn 
 
 . — i 
 
 
 
 o 
 
 
 m 
 
 ~f 
 
 vO 
 
 
 00 
 
 (Nl 
 
 m 
 
 
 
 
 
 
 —1 
 
 
 — 
 
 CO 
 
 
 
 
 — 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 , 
 
 
 
 
 
 
 sD 
 
 
 
 
 
 
 
 £ 
 
 t£ 
 
 - 
 
 
 - 
 
 ~ 
 
 
 vb 
 
 r- 
 
 xO 
 
 ■* 
 
 o 
 
 o 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 k 
 
 o 
 2 
 
 ~"* 
 
 o 
 
 '— , 
 
 ^ 
 
 o 
 
 in 
 
 
 in 
 
 f 
 
 J-l 
 
 0^ 
 
 - 
 
 1 
 
 
 
 
 
 
 
 
 
 r- 
 
 
 n£> 
 
 
 
 
 _o 
 
 *£ 
 
 
 
 '"' 
 
 in 
 
 
 00 
 
 
 
 
 o 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 — 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 c 
 
 • 
 
 o 
 
 o 
 
 ^1 
 
 r~ 
 
 o 
 
 (Nl 
 
 «o 
 
 o 
 
 o 
 
 00 
 
 ^< 
 
 (Nl 
 
 CO 
 
 D 
 
 o 
 2 
 
 
 
 
 
 
 ^* 
 
 o 
 
 
 
 
 t? 
 
 
 W 
 
 t£ 
 
 
 
 - 
 
 r» 
 
 
 r- 
 
 r- 
 
 "t 
 
 
 m 
 
 O 
 O 
 
 
 D 
 H 
 
 h 
 J 
 
 
 
 
 
 *■* 
 
 
 
 
 
 
 
 ~~* 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 «, 
 
 2 
 
 
 
 CO 
 
 o 
 
 _, 
 
 r- 
 
 o 
 
 o 
 
 ■* 
 
 vO 
 
 o 
 
 in 
 
 in 
 
 o 
 
 H 
 
 
 2 
 
 
 
 
 (Nl 
 
 
 —i 
 
 o 
 
 
 
 
 in 
 
 rn 
 
 CO 
 
 
 
 
 
 
 
 
 "* 
 
 
 
 
 ~ 
 
 
 <u 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 s 
 
 *£ 
 
 (N) 
 
 
 
 (Nl 
 
 
 sT> 
 
 r- 
 
 r^ 
 
 - 
 
 ^ 
 
 o 
 
 o 
 
 
 « 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 "^ 
 
 
 Oh 
 
 4-> 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 (T) 
 
 • 
 
 (Nl 
 
 o 
 
 o 
 
 "* 
 
 o 
 
 r- 
 
 00 
 
 (T- 
 
 _ 
 
 0^ 
 
 o 
 
 o^ 
 
 
 
 
 2 
 
 
 
 
 ~* 
 
 
 
 vO 
 
 
 
 
 (NJ 
 
 
 V 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 £ 
 
 E£ 
 
 4 
 
 
 _ 4 
 
 f- 
 
 _H 
 
 «♦ 
 
 ■<* 
 
 in 
 
 ~-, 
 
 vO 
 
 o 
 o 
 
 
 
 .— 1 
 
 
 
 
 
 
 
 *"* 
 
 vO 
 
 
 
 
 ~~* 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 
 (NJ 
 
 o 
 
 00 
 
 _ 
 
 in 
 
 r>- 
 
 — . 
 
 00 
 
 ■^ 
 
 —i 
 
 r~ 
 
 ro 
 
 
 o 
 
 
 
 
 vO 
 
 
 (NI 
 
 r- 
 
 *tf 
 
 
 in 
 
 oo 
 
 in 
 
 
 2 
 
 
 
 
 
 
 — ' 
 
 m 
 
 
 
 
 00 
 
 
 
 
 
 
 
 
 
 
 
 
 in 
 
 
 
 
 
 
 
 
 
 IQ 
 
 
 
 
 
 
 u 
 
 in 
 
 
 
 
 
 
 
 
 u 
 
 
 
 
 TJ 
 
 
 0) 
 
 u 
 
 
 
 
 
 
 
 
 ~- v 
 
 
 
 
 C 
 
 
 ^ 
 
 <u 
 
 
 
 V 
 
 
 
 
 r-l 
 
 y, M 
 
 
 
 
 rt 
 
 
 
 u 
 
 
 
 1—4 
 
 
 
 
 nj 
 
 C rtj 
 
 
 
 
 
 o 
 
 o 
 
 
 
 C Si 
 
 
 
 
 e 
 
 rt c 
 
 
 
 
 C 
 
 en 
 
 J3 
 
 
 
 o a 
 
 CO & 1-1 
 
 J J 
 
 
 
 o 
 
 10 
 
 in "> 
 
 en 
 
 OJ 
 
 (D 
 
 
 <u c 
 £ i 
 w £ 
 
 M O 
 
 U k 
 
 CD 
 
 > 
 
 rf 
 
 J 
 
 in 
 
 J 
 
 £ > 
 
 O 4-1 
 
 kh r-4 
 
 
 
 
 en 
 0) 
 
 o 
 
 M 
 
 CD 
 
 £ £ 
 
 u u 
 
 to 
 c 
 
 V 
 
 'J 
 
 U 
 
 in 
 
 o 
 
 ■ — • 
 
 to 
 
 M 
 
 D- 
 
 o 
 
 U 
 
 > 
 
 J-l 
 
 <U 
 
 £ 
 
 |H 
 
 n 
 
 u 
 o 
 
 <; 
 
 H 
 O 
 H 
 
 4. 14 
 

 
 
 CM 
 
 oo 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 >* 
 
 6> 
 
 
 0> 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 """ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 d 
 2 
 
 (NJ 
 
 r~ 
 
 
 
 
 
 
 
 
 
 o 
 
 CO 
 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 00 
 
 <*< 
 
 
 
 
 o 
 
 •* 
 
 
 
 
 
 
 
 • 
 
 
 o 
 
 
 
 ^ 
 
 t£ 
 
 
 CJN 
 
 
 
 
 
 
 
 
 
 o 
 
 
 to 
 
 D 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 ** H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 CO 
 
 
 < 
 
 H 
 
 d 
 
 (M 
 
 CO 
 
 CM 
 
 
 
 
 
 
 
 
 
 CO 
 
 r- 
 
 H 
 10 
 
 
 2 
 
 
 
 
 
 
 
 
 
 
 " H 
 
 
 ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 2 
 
 W 
 
 0) 
 
 E 
 
 s£ 
 
 o- 
 
 00 
 
 f-l 
 
 (N) 
 
 co 
 
 
 
 
 
 J 
 
 
 
 
 o 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 « 
 
 (-1 
 
 w 
 
 r*- 
 
 CO 
 
 
 
 
 
 
 
 
 
 oo 
 
 in 
 
 
 nJ 
 
 o 
 
 
 
 
 
 
 
 
 
 CO 
 
 CM 
 
 Ch 
 
 2 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r\i 
 
 
 
 
 
 
 
 
 
 
 
 
 V 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 £ 
 
 & 
 
 vO 
 
 •<* 
 
 
 
 
 
 
 
 
 
 o 
 o 
 
 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 d 
 2 
 
 m 
 
 o 
 
 co 
 (M 
 
 
 
 
 
 
 
 
 
 
 m 
 o 
 
 CO 
 
 — ta 
 
 CM 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ' 
 
 
 
 
 - 
 
 ■* 
 
 
 
 
 
 
 
 
 
 
 
 
 S 
 
 t£ 
 
 o 
 
 00 
 
 
 
 
 
 
 
 
 
 o 
 o 
 
 
 
 nJ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 h 
 
 
 
 2 
 
 in 
 
 00 
 
 
 
 
 
 
 
 
 
 CO 
 
 0^ 
 
 ON 
 
 i 
 
 
 ~* 
 
 00 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 *>4 
 
 ^ 
 
 (NJ 
 
 o 
 
 
 
 
 
 
 
 
 
 o 
 
 o 
 
 I 
 
 
 it o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 C 
 
 d 
 2 
 
 CO 
 
 (NJ 
 
 
 
 
 
 
 
 
 
 m 
 
 _ 
 
 
 O 
 
 
 Tf 
 
 
 
 
 
 
 
 
 
 "* 
 
 
 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <y- 
 
 
 
 
 
 
 
 
 
 
 
 w 
 
 w 
 
 J# 
 
 _* 
 
 
 
 
 
 
 
 
 
 
 
 o 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 — i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 2 
 
 d 
 
 (M 
 
 CO 
 
 
 
 
 
 
 
 
 
 o 
 
 in 
 
 
 2 
 
 
 in 
 
 
 
 
 
 
 
 
 
 vO 
 
 (NJ 
 
 
 
 
 
 •""• 
 
 
 
 
 
 
 
 
 
 — 1 
 
 
 a> 
 
 
 _ 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 E 
 
 wO 
 
 . 
 
 CO 
 
 
 
 
 
 
 
 
 
 o 
 
 
 t-t 
 
 
 &- 
 
 t>» 
 
 O^ 
 
 
 
 
 
 
 
 
 
 o 
 
 
 « 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Oh 
 
 44 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 14 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 AS 
 
 o 
 
 O 
 
 o 
 
 
 
 
 
 
 
 
 
 •43 
 
 CO 
 
 
 b 
 
 2 
 
 
 - 
 
 
 
 
 
 
 
 
 
 cm 
 
 H 
 
 V 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 £ 
 
 ^ 
 
 00 
 
 ro 
 
 
 
 
 
 
 
 
 
 o 
 o 
 
 
 
 H 
 
 
 
 On 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^-« 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 • 
 
 00 
 
 •<* 
 
 
 
 
 
 
 
 
 
 iM 
 
 CO 
 
 
 Ch 
 
 o 
 
 sO 
 
 ■<* 
 
 
 
 
 
 
 
 
 
 
 (NJ 
 
 
 2 
 
 
 CO 
 
 
 
 
 
 
 
 
 
 ON 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4> 
 
 « 
 
 
 
 
 
 
 
 
 
 
 
 
 
 c 23 
 
 w 
 
 
 CO 
 
 
 
 
 
 
 
 
 
 
 
 O rt 
 
 co £ 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 £ 7Z 
 
 P Eh 
 
 
 >H 
 
 2 
 
 
 
 
 
 
 
 
 
 _) 
 
 nJ rt 
 
 J o 
 o 2 
 
 X <; 
 
 O 
 
 »-3 
 
 
 
 
 
 
 
 
 
 
 
 H 
 O 
 
 H 
 
 u < 
 
 4.15 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 - 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 
 
 
 >N 
 
 e£ 
 
 o 
 
 co 
 
 00 
 
 IN) 
 
 Tf 
 
 co 
 
 oo 
 
 ro 
 
 
 o 
 
 
 
 
 4-1 
 
 
 *" h 
 
 (M 
 
 1— 1 
 
 ~-H 
 
 f— 1 
 
 
 
 
 
 PPH 
 
 
 
 
 U 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 d 
 Z 
 
 O"- 
 
 O 
 
 _ 
 
 in 
 
 in 
 
 00 
 
 ro 
 
 ro 
 
 
 ><* 
 
 00 
 
 
 
 o 
 
 ^ 
 
 a- 
 
 (M 
 
 00 
 
 a^ 
 
 m 
 
 in 
 
 *"* 
 
 
 00 
 vO 
 
 t 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 O 
 
 in 
 
 CO 
 
 <* 
 
 ■* 
 
 co 
 
 0^ 
 
 ro 
 
 
 o 
 
 
 
 co 
 
 *— • 
 n) 
 
 t£ 
 
 
 eg 
 
 ~ 
 
 " 
 
 ~* 
 
 
 
 
 
 o 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 ro 
 
 
 
 H 
 
 H 
 
 
 r» 
 
 r- 
 
 vD 
 
 ro 
 
 ro 
 
 vO 
 
 m 
 
 r- 
 
 
 ro 
 
 r- 
 
 
 <, 
 
 
 d 
 
 ~* 
 
 m 
 
 ro 
 
 00 
 
 00 
 
 o 
 
 -^ 
 
 ro 
 
 
 ro 
 
 r~ 
 
 
 H 
 
 co 
 
 
 Z 
 
 CO 
 
 ro 
 
 ro 
 
 " H 
 
 ■"■ 
 
 — ' 
 
 *■* 
 
 
 
 - - 
 
 
 
 0) 
 
 £ 
 
 
 _ 
 
 Tf 
 
 tJ< 
 
 I s - 
 
 t . 
 
 o- 
 
 rf< 
 
 CT- 
 
 
 O 
 
 
 
 z 
 
 & 
 
 
 
 
 
 
 
 "-" 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 w 
 
 05 
 
 H 
 
 4-» 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 <tf 
 
 in 
 
 m 
 
 v£> 
 
 ■tf 
 
 ro 
 
 m 
 
 ro 
 
 
 m 
 
 oo 
 ro 
 
 
 £ 
 
 IX) 
 
 o" 
 
 
 
 
 
 
 
 
 
 
 ro 
 
 
 
 U 
 
 & 
 
 Z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4) 
 § 
 
 
 o 
 
 o 
 
 00 
 
 >* 
 
 rf 
 
 00 
 
 00 
 
 ro 
 
 
 o 
 
 
 
 
 sS 
 
 '- 1 
 
 (M 
 
 ** 
 
 — ' 
 
 — ' 
 
 
 
 
 
 o 
 
 
 
 
 EH 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Cn 
 
 d 
 
 ro 
 
 CM 
 
 rt 
 
 r- 
 
 co 
 
 ro 
 
 o 
 
 •<* 
 
 
 oo 
 
 
 
 
 Z 
 
 ro 
 
 ro 
 
 ro 
 
 r- 
 
 r- 
 
 O 
 
 — ■ 
 
 ro 
 
 
 ro 
 
 ro 
 
 
 
 
 —* 
 
 m 
 
 rXJ 
 
 — i 
 
 — i 
 
 i — i 
 
 — 1 
 
 
 
 i — . 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 £ 
 u 
 it 
 
 t£ 
 
 -£> 
 
 a- 
 
 cr 
 
 co 
 
 •o 
 
 "* 
 
 (M 
 
 •* 
 
 
 o 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 k 
 
 6 
 z 
 
 vD 
 
 oo 
 
 oo 
 
 r» 
 
 m 
 
 x* 
 
 " H 
 
 f 
 
 
 ro 
 
 cr- 
 
 
 i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 ^ 
 
 vO 
 
 Tf 
 
 ro 
 
 m 
 
 ^t< 
 
 O 
 
 r- 
 
 ^h 
 
 
 o 
 
 
 
 
 
 
 (M 
 
 1-0 
 
 
 
 
 
 
 
 o 
 
 
 
 
 a 
 
 p -a 
 
 M (1) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 0) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 c 
 
 d 
 Z 
 
 o-> 
 
 ■t 
 
 ro 
 
 (M 
 
 o 
 
 ^C 
 
 o 
 
 ro 
 
 
 •* 
 
 ro 
 
 
 
 D 
 
 
 ro 
 
 ro 
 
 <M 
 
 <M 
 
 '~* 
 
 ■ ' 
 
 
 
 ■* 
 
 
 
 
 
 s£ 
 
 
 O 
 
 in 
 
 o 
 
 CO 
 
 o 
 
 r- 
 
 ro 
 
 
 o 
 
 
 
 CO 
 
 W 
 
 X 
 
 ro 
 
 H 
 
 _ 
 
 _« 
 
 _c 
 
 
 
 
 o 
 
 
 
 E- 
 
 
 
 
 
 
 
 
 
 
 
 
 — ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 z 
 
 d 
 
 <M 
 
 m 
 
 t 
 
 ro 
 
 _ 
 
 ^o 
 
 ^, 
 
 ro 
 
 
 vO 
 
 o 
 
 
 
 z 
 
 — . 
 
 T(< 
 
 <M 
 
 rH 
 
 M 
 
 
 — . 
 
 
 
 m 
 
 ro 
 
 
 10 
 
 
 
 
 
 
 
 
 
 
 
 
 — ' 
 
 
 
 <u 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 £ 
 
 ^ 
 
 LO 
 
 >* 
 
 r- 
 
 in 
 
 co 
 
 vO 
 
 00 
 
 ro 
 
 
 o 
 
 
 
 t-i 
 
 .,- 
 
 6^ 
 
 ^-4 
 
 tvj 
 
 (VI 
 
 ^_, 
 
 
 
 
 
 
 o 
 
 
 
 « 
 
 H 
 
 
 
 
 
 
 
 
 
 
 
 i*H 
 
 
 
 cm 
 
 -*-> 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 tH 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 nJ 
 
 d 
 2 
 
 O- 
 
 o 
 
 ro 
 
 "* 
 
 o 
 
 r- 
 
 o 
 
 ro 
 
 
 m 
 
 Tf 
 
 
 
 (^ 
 
 ~* 
 
 ro 
 
 ro 
 
 ro 
 
 ~~* 
 
 
 ^ H 
 
 
 
 ro 
 
 " 
 
 
 V 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6 
 
 wO 
 
 _-. 
 
 m 
 
 ^O 
 
 T 
 
 t 
 
 co 
 
 cr- 
 
 ro 
 
 
 o 
 
 
 
 
 6^ 
 
 »H 
 
 (M 
 
 p>h 
 
 •-* 
 
 *H 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 "■* 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 *3 
 
 . 
 
 r- 
 
 00 
 
 ^O 
 
 <* 
 
 _ 
 
 o 
 
 ■^ 
 
 CT~ 
 
 
 00 
 
 ro 
 
 
 
 Cn 
 
 o 
 
 o 
 
 CO 
 
 r* 
 
 r0 
 
 ro 
 
 vO 
 
 00 
 
 ro 
 
 
 o 
 
 ro 
 
 
 
 Z 
 
 
 ro 
 
 — 
 
 - 1 
 
 - 1 
 
 
 
 
 
 cr- 
 
 
 
 
 
 
 0) 
 
 
 
 
 
 
 
 0) 
 
 
 
 a> 
 
 
 
 co 
 
 CO 
 
 
 c 
 
 o 
 
 
 
 
 O^ 
 
 CT^ 
 
 O 
 
 u 
 
 o 
 
 
 
 C XI 
 
 O IT) 
 
 
 
 W 
 
 u 
 
 c 
 
 o 
 
 o 
 
 O^ 
 
 ■tf 
 
 c> 
 
 0^ 
 
 £ 
 
 
 
 
 
 
 Z 1 
 
 « 
 
 <; 
 w 
 
 Z 
 
 3 
 
 
 
 — < 
 
 -C 
 
 «0 0) 
 
 <u -p 
 
 I 
 O 
 
 in 
 o 
 
 O^ 
 
 1 
 
 o 
 
 o 
 o 
 
 1 
 
 o 
 iri 
 
 ro 
 i 
 
 o 
 
 o 
 
 u 
 
 o 
 
 o 
 o" 
 
 
 H 
 O 
 
 O *j 
 
 
 - 
 
 
 
 -* 
 
 rO 
 
 -o 
 
 - 1 
 
 "- 1 
 
 ro 
 
 ro 
 
 
 -i f-+ 
 
 
 A. 16 
 
EXHIBIT IV -1 
 
 4.17 
 
Chilton Research Services 
 Philadelphia, Pennsylvania 
 
 Study it 9138 
 Bureau of Budget 
 #41 - S - 69015 
 Exp. December, 1969 
 
 EMPLOYEE TREND STUDY 
 
 We are conducting a worker study to determine their work patterns during the last three or 
 four years. You can h^lp us greatly by providing the information requested in this 
 questionnaire. You are never identified with your answers which are combined with all 
 other answers to form a picture of the area as a whole. Most questions can be answered by 
 checking a box. Please disregard the small numbers before the boxes — they are for tabu- 
 lation purposes only. Thank you. 
 
 Your Name: 
 
 Home Telephone Number 
 
 First, we would like to get some general characteristics of your background. 
 PLEASE CHECK ONLY ONE BOX FOR THE FOLLOWING QUESTIONS: 
 
 1) Please check whether you are male or female? 
 
 9-1 | 1 Male 2 | | Female 
 
 2) Are you married, never marrieH, widowed or divorced? 
 
 10-1 _J Married 2 Never married 3f~ Widowed 4 | ^Divorced 
 
 3) Please check the last grade or class you completed in your schooling. 
 
 11-1 ! Never Attended School 5 JT Graduated High School 
 
 2 Attended grammar school 
 did not graduate 
 
 3 Graduated Grammar School 
 
 4 Attended High School 
 did not graduate 
 
 6 Attended College 
 did not graduate 
 
 7 Q Graduated College 
 
 8 Q Graduate Work 
 
 4) Please check the one box which tells us your age. 
 
 12-1 \^\ Under 21 
 
 2 □ 21 - 24 
 
 3 □ 25 - 29 
 
 4 Q 30 - 34 
 
 5 □ 35 - 39 
 
 6 □ 40 - 44 
 
 7 □ 45 - 49 
 
 8 □ 50 - 54 
 
 9 | | 55 - 59 
 
 CD 60 " 64 
 X £] 65 & Over 
 
 4.18 
 

 - '■"■ 
 
 
 
 
 5) 
 
 please check ONLY one box. 
 
 
 
 
 * 
 
 13 _1 [J \miTE 2 
 
 l>f j SPANISH AMERICAN 
 
 □ 
 
 NEGRO 3 
 5 ORIENTAL 
 
 f~] INDIAN 
 
 6) 
 
 Including yourself, how many members c 
 
 >f your family are there living with you? 
 
 
 14-1 □ One 5 
 
 D 
 
 Five 9 
 
 |~~1 Nine 
 
 
 2 Qj Two 6 
 
 D 
 
 Six o 
 
 i| Ten or More 
 
 
 3 Q Three 7 
 
 D 
 
 Seven 
 
 
 
 A Q Four 8 
 
 D 
 
 Eight 
 
 
 7) 
 
 How many of them are children under 18 
 
 years of age? 
 
 
 
 15-0 ~2 None 4 
 
 D 
 
 Four 8 
 
 D Ei 8 hc 
 
 
 1 Q One 5 
 
 D 
 
 Five 9 
 
 I | Nine or More 
 
 
 2 Q Two 6 
 
 □ 
 
 Six 
 
 
 
 3 □ Three 7 
 
 □ 
 
 Seven 
 
 
 8) Are you the principal bread winner of the family? 
 
 16-X ~] YES SKIP TO Q. 9 1 [j NO , please answer: 
 
 a) Do you contribute to the support 
 of the family? 
 
 17-1 Q YES X □ NO 
 
 9) Please list the City or Town, County and State in which you live. 
 
 (City or Town) (County) 
 
 (State) ~ ~~ ~~ ' 4<19 
 
' in) About how long have you been living in this town? PLEASE CHECK ONLY ONE BOX. 
 
 20-0 Less than 1 year A |_J Four years 8 [_J Eight years 
 
 1 ~] One year 5 Q Five years 9 Qj Nine years or More 
 
 2 "~] Two years 6 Q Six years 
 
 3 Three years 7. j_[ Seven years 
 
 llj. a. Were you living in this county before you started your present job or not? 
 21-X [J YES 1 [J NO 
 
 b. If not, please specify the county in which you lived? 
 
 22- 
 
 (County) 
 
 (State) 
 
 12) Please indicate about how much of your purchases, none, some, half, most or all, for 
 the following items are spent in the county in which you live. (PLEASE CHECK ONE BOX 
 FOR EACH TYPE OF PURCHASE) 
 
 SOME- HALF- MOST- 
 NONE APPX. APPX. APPX. ALL 
 Type Of Purchase 25% 50% 75% 
 
 Food 
 
 Family Clothing 
 
 Large purchases such as 
 
 Appliances, furniture, 
 
 23-0 i I 111 211 3 I | 4 | 
 
 cars, etc. 
 Other purchases 
 
 24-0 I 
 25-0 [ 
 
 26-0 r 
 
 1 I 
 
 1 I j 
 
 i n~i 
 
 2 j i 
 
 2 CZj 
 
 2 i . 
 
 3 I i 
 
 3 I i 
 
 Now, let us talk about your current employment 
 
 13) About how long have you been working for this company at this location? PLEASE CHECK 
 ONLY ONE BOX. 
 
 27-0 Less than 1 year 
 
 1 One year but less than 18 months 
 
 2 18 months but less than 2 years 
 
 3 Q 2 years but less than 2 1/2 years 
 
 ^ ~Z\ 2 1/2 years but less than 3 years 
 
 5 3 years but loss than 3 1/2 years 
 
 4.20 
 
 6 [H 3 1/2 yrs. but less than 4 yrs. 
 
 7 Q 4 years but less than 4 1/2 yrs. 
 
 8 Q 4 1/2 years but less than 5 yrs. 
 
 9 Five years or more 
 
.,» ' \ D id you work for this company somewhere else? 
 
 2 8-l D YES 2 □ N ° 
 
 h) If YES, where was that, in what City or Town, County or State? 
 
 (City or Town) 
 
 (County) 
 
 (State) 
 
 29- 
 
 151 Do you generally work full-Lime (35 hours each week or more) or part-time (less than 35 
 hours each week)? 
 
 30-1 □ Full-T 
 
 lme 
 
 2 Q Part-Ti 
 
 lme 
 
 16) Durine the past 12 months, about how.manv months have you worked for this company 
 it this location? 
 
 31-1 | | One Month 
 
 Five Months 
 
 32-1 I | Nine Months 
 
 2 j | Two Months 
 
 6 ( | Six Months 
 
 2 ( [ Ten Months 
 
 3 I ! Three Months 
 
 7 | Seven Months 
 
 3 I | Eleven Months 
 
 4 i I Four Months 
 
 8 I | Eight Months 
 
 4 I | Twelve Months 
 
 17) What is your present occupation, that is, what do you do? 
 
 33- 
 
 (Type of Work Done) 
 
 (Title) 
 
 18) About how far do you travel to work each day? 
 
 (MILES) 
 
 (One Way) 
 34- 
 
 19) Please check the ONE box which shows how much you normally earn each week from this 
 Job, before deductions. 
 
 35-0 Q Less than $40 
 
 1 □ $40 to $59 
 
 2 Q $60 to $79 
 
 3 □ $80 to $99 
 
 4 Q $100 to $119 
 
 5 □ $120 to $139 
 
 6 [J $140 to $159 
 
 7 [J $150 to $179 
 
 36-1 □ $180 to $199 
 
 2 □ $200 to $219 
 
 3 Q $220 and Over 
 
 4.21 
 
Do you have another job besides this one? 
 
 37- 1 J YES X O NO- If NO, SKIP TO QUESTION 24 
 
 21) 
 
 What do you do at your other job? 
 
 (Type of Work Done) 
 
 38- 
 
 (Title) 
 
 22) please check the ONE box which shows how much you normally made each week from this 
 other job in 1968. 
 
 39-1 Less than $40 
 
 2 j~~j $40 to >59 
 
 3 □ $6Q to $79 
 
 4 Q $80 to $99 
 
 5 □ $100 to $119 
 
 6 Q $120 to $139 
 
 7 □ $140 to $159 
 
 8 □ $160 to $179 
 
 40-lD $1 80 to $199 
 l{3 $200 to $219 
 3lJ $220 and Over 
 
 23) Where is this job located — in what City or Town, County and State? 
 
 (City or Town) 
 
 (County) 
 
 41- 
 
 (State) 
 
 24) Now, please check the ONE box which best describes your previous employment status, 
 that is, before you took the job described in questions 13 thru 19. 
 
 42- I. [ j Worked on a f ai 
 
 (Now please answer Questions 
 25 to 29 on blue form only) 
 
 3 f | Worked Part-Time-Less than 
 35 hours a week 
 
 (Answer Questions 35 thru 38 
 on gold form only) 
 
 2 { ] Worked full-time somewhere else 
 ...over 35 hours a week and for 
 more than 45 weeks a year 
 
 (Answer Questions 30 thru 34 on 
 greer. form only) 
 
 4 I I Unemployed 
 
 (Answer Questions 39 thru 44 
 on pink form only) 
 
 4.22 
 
I For Office Use Only | Col. 43 
 
 Those who previously worked on a farm 
 please answer Questions 25 thru 29. 
 
 25) Did you own the farm? 
 
 44-1 LJ YES 
 
 2 j I NO 
 
 26) Did you work full-time(at least 35 hours a week) or part-time (less than 3_. hours 
 a week) on the farm? 
 
 45-1 I I Full-Time 
 
 2 j ( Part-Time 
 
 a. If worked only part-time, please indicate whether you preferred 
 to work part-time or not. 
 
 46 -i[ j Preferred Part-Time 
 
 2 1 j Did Not 
 
 27) About how many months a year did you work? 
 
 l, 7 - ]| J One Month 5 | | Five Months 
 
 3 ~| Two Months 6 | | Six Months 
 
 .1 j Three Months 7 | j Seven Months 
 
 ^LJ Four Months 8 \_ j Eight Months 
 
 48-1 I J Nine Months 
 
 2 I I Ten Months 
 
 3 I J Eleven Months 
 
 4 I I Twelve Months 
 
 22) Hov much did you earn a year (After deducting farm expenses such as feed, fertilizer 
 aiid equipment)? 
 
 49-I ~^\ Less than $1,000 
 
 2 □ $1,000 to $1,999 
 
 3 Q $2,000 - $2,999 
 
 4 Q $3,000 - $3,999 
 
 5 P $4,000 - $4,999 5Q-1 □ $7,500 - $9,999 
 
 6 □ $5,000 - $5,999 2 □ $10,000 - $14,999 
 
 7 Q $6,000 - $6,999 3 P $15,000 & Over 
 
 8 □ $7,000 - $7,499 
 
 29) a) Do you still work on the farm in your spare time? 
 51 -1 p YES 2 □ NO 
 
 b) If you still work on the farm, about how much do you earn from the farm in a year? 
 52-1 Q Less than $1,000 5 Q $4,000 - $4,999 53-1 □ $7,500 - $9,999 
 
 6 □ $5,000 - $5,999 2 Q $10,000 - $14,999 
 
 7 [I] $ 6 . 000 ~ 56,999 3 □ $15,000 & Over 
 
 8 □ $7,000 - $7,499 
 
 4.23 
 
 2 P $1,000 to $1,999' 
 
 3 P $2,000 - $2,999 
 
 4 □ $3,000 - $3,999 
 
• 
 
 - 
 
 For Office Use Only 
 
 Col. 5/, 1 
 
 30) 
 
 
 Those who previously worked somewhere else 
 full-time, please answer Questions 30 to 34 
 
 , 
 
 About how much did you earn in an average work week at your previous job before 
 deductions? 
 
 
 Less than $40 4 Q $100 to $119 56-1 □ $180 to $199 
 
 
 $40 to $59 5 □ $120 to $139 2 □ $200 to $219 
 
 
 2 □ $60 to $79 6 Q] $140 to $159 3 □ $220 and Over 
 
 
 3 Q $80 to $99 7 □ $160 to $179 
 
 
 31) About how long did you work there? 
 
 
 57- 
 
 
 
 32) 
 
 Where was this company located ... in what City or Town, County and State? 
 
 
 (City or Town) 
 
 (County) 
 
 (State) 
 
 58- 
 
 33) 
 
 What was your occupation, that is, what did you do? 
 
 
 - 
 
 59- 
 
 (Type of Work Done) 
 
 (Title) 
 
 34) 
 
 When you left, did your employer (Company) hire someone to replace you? 
 
 
 60 -1 Q YES 2 □ NO 3 □ DON'T KNOW 
 
 
 
 4.24 
 
 
 
 
For Office Use Only 
 
 Col. 01 
 
 If worked less than 35 hours each week previous to beginning 
 your present job, please answer questions 35 thru 38- 
 
 35) Did you prefer to work less than 35 hours a week? 
 
 62-1 [7 YES 
 
 2 Q NO 
 
 36) About how many hours did you work in a normal week? 
 
 (,3-1 ;23 Less than 10 hours 4 Qj 20 to 24 hours 
 
 2 Q] 10 to 14- hours 5 Q 25 to 29 hours 
 
 3 j^] 15 to 19 hours 6 ] ~\ 30 to 34 hours 
 
 37) Please check the ONE box which shows how much you earned in an average week before 
 deductions from this job. 
 
 t 4-l [7] Less than $20 
 
 2 □ $20 to $29 
 
 3 □ $30 to 30 
 
 4 [J $40 to $49 
 
 5 P $50 to $59 65-1 Q $120 to $149 
 
 6 □ $60 to $69 2 □ $150 & Over 
 
 7 □ $70 tc 489 
 
 8 □ $90 to $119 
 
 38) What did you do at that job? 
 
 (Type of Work Done) 
 
 66- 
 
 (Title) 
 
 4.25 
 
For Office Use Only j Col 67 
 
 Those who were unemployed, 
 please answer Questions 39 to 44 
 
 39) Were you actively looking for work? 
 6 8 1 □ YES 
 
 2 □ NO 
 
 40) Please check all the reasons why you were not working, 
 PLEASE CHECK AS MANY AS APPLY . . . 
 
 69_ 1 Long Illness 
 
 2 ^3 Care for children 
 
 3 Enrolled in school 
 
 4 ^] Military 
 
 5 Didn't want work 
 
 6 [_ Awaiting recall to a Job 
 
 7 No work available 
 
 8 □ Laid off 
 
 9 Q Moved 
 
 SP None of the above 
 
 41) Were there any other reasons? Please Explain:_ 
 
 70- 
 
 42) About how long were you out cf work? PLEASE CHECK ONLY ONE BOX. 
 
 71 -1 Less than 6 weeks 2 6 to 15 weeks 3 Q Over 15 weeks 
 
 43) Were you receiving any welfare payments or unemployment in3urence irhen you cero 
 out of work? 
 
 72-1 □ YES 
 
 2 □ NO 
 
 44) If YES i about how much were you receiving ea;h month? 
 
 $ 
 
 73- 
 
 4.26 
 
V. ANALYSIS OF THE DETERMINANTS OF PROGRAM SUCCESS 
 
 The factors underlying the success of a business loan program 
 can be analyzed simply in terms of their effects upon the likelihood that 
 loans will be in default, or equivalently, that firms will fail. Indeed, 
 such an analysis has much to contribute to program planning by clarifying 
 the risk elements characterizing each loan so that loan approval decisions 
 can be rationally based. 
 
 "Where high risk loans are in conformity with program objectives, 
 at the very least the analysis of factors underlying default will provide 
 planners with a means for estimating future program costs. But where 
 other objectives underlie the program, such as the maximization of 
 poverty reduction, then the default rate, or in more positive tone, the success 
 likelihood, becomes an instrumental variable that affects the extent to 
 which program objectives will be realized. In this case, knowledge of 
 the factors underlying the probability of success is a necessity for purposes 
 of program control. 
 
 But since program objectives center around the achievement of economic 
 and welfare magnitudes, success will be measured by more than the "no 
 default" criterion. The extent of success as measured by unemployment 
 reduction, poverty amelioration, or contribution to total area product will 
 be essential elements of the evaluation indicators. Here, the factors that 
 condition the degree of success are of interest to the program planning 
 process. Knowledge of such factors will aid in the loan approval decision 
 
 5.1 
 
through knowledge of conditions that should be avoided, and will provide 
 information on the factors controllable by the program itself. 
 
 In this chapter, accordingly, tvo types of analyses will be made: 
 
 1. Factors affecting the probability of success, and 
 
 2. Factors affecting the magnitude of success. 
 
 The factors to be analyzed in each case will be grouped into three 
 sets: characteristics of the firm; characteristics of the loan; and socio- 
 economic characteristics of the environment in which the firm is to be 
 located. Of these three sets of factors, the second, characteristic of 
 the loan, is directly controllable in the planning process. The size of the 
 loan, for example, of critical interest in the selection of loan applicants 
 and in program budgeting, is subject to control wifhin the program. If 
 the magnitude of success increases more than proportionately with the mag- 
 nitude of the loan, the larger loans will be indicated, up to the point of 
 diminishing returns (O L. in Figure 1). But this is not meant to imply that 
 the characteristics of the loan are of dominant importance among the fac - 
 
 tors that underlie program success 
 
 Payroll (P) of 
 the Firm 
 
 cost 
 
 L Size of loan ($) 
 
 Figure !• Factors Affecting Size of Loan Decision 
 
 5.2 
 
It is one of many potentially important factors, among which are those 
 that, while not directly subject to program control, are subject to con- 
 sideration in the loan approval process. Thus, if a firm has special 
 labor requirements, and these factors are known to have a depressing, 
 if not pre-emptive effect upon the "payroll curvets in curve P~ in Figure 2), 
 then labor applicants possessing such characteristics can be avoided in 
 favor of those with characteristics indicating higher probability of success. 
 
 Payroll (P) 
 
 \^/ 
 
 Size Of Loan 
 
 Figure 2: Factors Affecting Loan Approval Decision 
 
 A. Analysis of the Determinants of Success Probability 
 
 In this section, multivariate analytic techniques are applied for 
 assessing the determinants of success probability. The first such 
 application is the use of multiple regression analysis (ordinary least 
 squares) to the following relation:* 
 
 * See Table 1 for list of variables in relation to equation 1 
 
 5.3 
 
(1) « = a + Zb. SIC. + Zb. ACT. + g EDATO + g, EDAPR + g,TIME 
 e=l l j=l J J *^_ 3 
 
 Char. Char, 
 
 of Firm of Loan 
 
 + hj POP + h 2 APOP + h 3 MTG + h M/ED+ h %FEM 
 
 + h 6 M/AG + h ? M/WG + h g %MG + h %INC + h 1Q %PUT 
 
 Characteristic of Community 
 
 Tlie analysis was made using various logarithmic transformations of 
 the: data, as well as of the data in its original form. The analysis was made 
 upon, data for 68 counties (observations) of which 27 contained default-loan 
 firms,, and 41 contained successful firms, all of which obtained ARA loans 
 between 1962 and 1965. * Given that the dependent variable is limited only 
 
 to zero and one values, the predictive power of an Initial regression in which 
 
 2 
 
 all variables were included was encouraging (R = .567). From this run, 
 
 nine variables were found to be most appropriate for inclusion for further 
 analysis. The results of a linear regression of these variables are shown 
 rn cxdiirrrns 1 and 2 of Table 2. ## 
 
 Following a series of tests using logarithmic transformations on the 
 variables, eight variables were found most appropriate for inclusion, the 
 
 results? are shown in columns 3 and 4 of Table 2. This regression provides 
 
 22 
 
 aaH of. . 58, with all eight variables significantly associated with the depen- 
 dent: variable. Of considerable interest are the following: 
 
 U.. A non-linear (log) transformation for "size of EDA loan" provides 
 a better fit than a linear transform. 
 
 # The successful firm sample included most of the^ surveyed firms, 
 bxit crontained many not surveyed. 
 
 ** The analysis was that of step-wise multiple regression using BIO- 
 
Table 1 
 List of Variables * 
 
 6 - is if firm defaulted; 1 if firm succeeded 
 
 SIC. = is industry classification: i= 1 if industry is resource 
 
 oriented; i=Z if labor oriented; i=3 if recreau'on oriented 
 
 ACT. = is firm classification: j=l if firm is new; j = 2 if branch 
 
 plant; j— 3 if plant expansion; j=4 if loan save firm 
 
 b., b. = dummy coefficients (0, 1) 
 
 EDATO = is size of EDA loan 
 
 EDA PR = is size of EDA loan as a percent of total project cost 
 
 TIME = is number of months, since the date of loan approval 
 
 POP = is I960 population of the county in which firm is located 
 
 APOP = is the change in. county population, 1950-1960 
 
 MIG = is tire county net migration rate, 1955-1960 
 
 M/ED = is the median county education. (years school completed) 
 
 %FEM = is the percent of females in county labor force 
 
 M/AG = is the median age of the county population 
 
 M/WG = ia the median, county manufacturing wage, I960 
 
 %MG = is the percent county labor force in manufacturing industry, 
 
 I960 
 
 * Sources: Indicators of characteristics of firm and of loans were 
 taken from EDA loan records provided by the Office Of Business Loans, 
 EDA. Socio-economic characteristics of communities were taken from 
 City and County Data Book, I960 , with the exception of the income growth 
 rate data, which was taken from OBE estimates, Department of Commerce 
 
 5.5 
 
Table 1 (continued) 
 
 %INC = is the aggregate county income growth rate, 1962-1966 
 
 %PVT ~ is the % of families earning under $3000 (poverty) 
 
 5.6 
 
TABLE 2 
 
 Regression with Success Probability as Dependent 
 Variable Number of Observations: 68 
 
 ■ 
 
 Linear Variables 
 
 Mixed Transform 
 
 Logarithmic 
 
 coefficient 
 
 t- value 
 
 coefficient 
 
 t-value 
 
 Transform 
 
 SIC (1) 
 
 -0.278 
 
 -2. 1 
 
 -0. 139 
 
 -2.6 
 
 Yes 
 
 SIC (2) 
 
 -0.305 
 
 -2.3 
 
 -0. 123 
 
 -2. 2 
 
 Yes 
 
 ACT (1) 
 
 -0. 181 
 
 -2. 1 
 
 -0. 192 
 
 -2.3 
 
 
 TIME 
 
 -2.801 
 
 -6.6 
 
 -2.773 
 
 -6.8 
 
 
 EDATO 
 
 0. 171 
 
 +2. 2 
 
 0. 1.14 
 
 +2. 8 
 
 Yes 
 
 POP 
 
 0.386 4-2.7 
 
 0.342 
 
 4-2.4 
 
 
 M/ED 
 
 0.336 +0.9 
 
 deleted 
 
 
 
 M/WG 
 
 -1.690 
 
 -2. 1 
 
 -1.34S 
 
 -1. 9 
 
 
 %INC 
 
 -1. 155 
 
 -2. 
 
 -0. 33 T 
 
 -2.3 
 
 Yes 
 
 CONSTANT 
 
 2.886 
 
 
 2-833 
 
 
 
 56 
 
 .58 
 
 5.7 
 
2. Although the size of county as measured by population size 
 has a positive effect upon the likelihood of success, variables indicating 
 economic growth (median wage and inccme growth rate) have a negative 
 affect. 
 
 3. Time since date of the loan is the most significant of all varia- 
 bles, and is negatively related. This is undoubtedly reflective of the fact 
 that firms have increased likelihood of failure the longer they are in exis- 
 tence (80% of all new firms in the United States fail in the first five years). 
 
 The next type of analysis of factors underlying the success proba- 
 bilities of EDA loan-recipient firms was discriminant analysis. * Because 
 the firm status grouping that was to be predicted by the discriminant 
 function was the same as that used as the dependent variable in the 
 regression analysis (success or failure) the results of this analysis 
 essentially mirrors that of the preceeding. Nevertheless, the form of 
 the estimated result is clearly more useful- for policy purposes than 
 that of the regression analysis:, since the discriminant function unam- 
 biguously states whether a given firm is expected to succeed or fail, 
 while certain difficulties becloud the interpretation of the regression 
 result.** The results of the analysis are shown in Tables 3 and 4. 
 
 * Given a set of observations belonging to one of two groups (the set 
 of firms belonging to the success or the failure group), discriminant analysis 
 estimates a function, as a linear combination of the attributes of the observa- 
 tions that best serves to predict which of the two groups an observation is ex- 
 pected to belong. 
 
 ** The predicted value of the (0, 1) regression may easily exceed unity or 
 take on a negative value, a result inconsistent with the probability interpretation 
 of the dependent variable. See David Durand's application of discriminat 
 analysis to discriminate between good and bad consumers installment loans, in 
 G. Fintner, Econometrics , New York: John Bailey, 1952. Original reference 
 is D. Durand, "Risk Elements in Consumer Installment Financing", Studies 
 in Consumer Installment Financing 8, NBER, 1941. 
 
Table 3 
 
 Statistical Estimates of the Discriminant Function 
 
 2 
 
 Mahalanobis D 
 
 5. 315" 
 
 
 . 
 
 Dividing Point 
 
 -0.2935- 
 
 
 
 Variable 
 
 Co efficient 
 
 Variable 
 
 Coefficient 
 
 SIC(l) 
 
 -0. 042. 
 
 APOP 
 
 -0. 003 
 
 SiC(2) 
 
 -u. 049 
 
 MIG 
 
 0.030 
 
 AGT(l) 
 
 -0. 0Z5f 
 
 M/ED 
 
 0.046 
 
 ACT(2) 
 
 a.. ODD 
 
 %FEM 
 
 0.086 
 
 ACT(3) 
 
 0.005: | 
 
 M/AG 
 
 -0.015 
 
 EDATO 
 
 -a..2fi55 ! 
 
 M/¥G 
 
 -0.242 
 
 EDAPR 
 
 0.023: 
 
 %MG 
 
 -0.019 
 
 TIME 
 
 0.040' 
 
 %INC 
 
 -0.169 
 
 POP 
 
 0.035 
 
 %PUT 
 
 -0.004 
 
 5.9 
 
Table 4 
 
 Estimated Versus Actual Outcome 
 
 ~**"^^-A ctual 
 Est i ma!Hv^ 
 
 Succeed 
 
 Fail 
 
 Succeed 
 
 Fail 
 
 Total 
 
 3.6 
 
 5 
 
 41 
 
 1 
 26 
 27 
 
 5.10 
 
Table 3 gives the coefficients of the variables in the discriminant function 
 
 as well as a measure of its ability to utilize these variables as the basis 
 
 2 
 
 for discriminating between success and failure (the Mahalanobis D ) 
 
 failure! The dividing point between success and failure is also provided. 
 In Table 4, the estimation of whether each firm would have been expected 
 to succeed or fail, based on the underlying factors, is shown, grouped 
 according to the actual success or failure outcome for the firm. Five 
 of the 41 firms that succeeded were "predicted" to fail, while only one of 
 the 27 firms that failed was expected to succeed. 
 
 B. Analysis of the Determinants of the Degree of Success 
 
 In this section, ordinary least squares regression will be applied 
 for identifying the variables that, are significantly related to the degree 
 of success, given of course, that the firm did not fail- Two analyses 
 are made: the first analyzing, the determinants of total wages paid * 
 to employees of KJJ.A loan-recipient firms (AY), and the second analyzing 
 the number of employees* hired in each such firm (AE) since the date 
 of loan approval.. The independent variables included in both analyses are 
 exactly the same as those used in the preceeding section (See Table 1). 
 
 The income change (AY) regression for all variables in their original 
 
 Z 
 
 (linear) form provided an R of .74**, with ten significant variables, as 
 
 * These measures are obtained from the survey of Chilton Research 
 Services, Inc. , and are "blown-up" to reflect the total employment from 
 each firm. 
 
 ** Not corrected for degrees of freedom. 
 
 5.11 
 
shown in columiio 1 <x«.d 7. of Table 5. Based upon this analysis, as well 
 as a number of analyses made with varying logarithmic trans formations 
 of the variables, it was decided to use a semi-log transform of the 
 regression equation (taking the logarithm of the dependent variable only), 
 implying that the true form of the function is exponential (i. e. , AY = 
 
 L ). The results of this analysis, where all variables but 
 
 : Vl e b 2 x 2 . . . e bnX »). 
 
 three (TIME, APOP, and M/AG) entered, indicate high predictive power 
 
 2 
 
 far the relation, with R - . 85*, as shown in columns 3 and 4 of Table 5. 
 
 Ten variables were found to be. significant (t -Z). Because only 33 ob- 
 servations could be included in the analysis, it was necessary to reduce 
 the number of variables in the regression. Removing four additional 
 
 variables from rh<* r-pgrefrsion (ACT(2) r EDAPR, %FEM, and M/WG) 
 
 2 
 
 provides an R of. 8.0, * with nearly all the included variables being 
 
 significantly. The findings of this : semi-log regression were as follows: 
 
 1. Neither the firm classification (ACT.) northe degree of poverty 
 in the county were significantly associated with the degree of success. ** 
 
 2. SIC type made a difference, with the resource-oriented industries 
 having a small advantage over the labor-oriented industries, and almost 
 certainly over the recreation-oriented- industries, which, while necessarily 
 
 * Not corrected for degrees of freedom. 
 
 ** This is in addition to the other- deleted variables, including time 
 since loan, EDA percent participation, the three environmental character- 
 istics of average age, wage and percent female, and the population change 
 measure which was collinear with the migration measure. 
 
 5.12 
 
TABLE 5 
 
 Regression with A-Y as Dependent Variable 
 Number of Observations: 33 
 
 Variable 
 
 No Transform 
 
 Semi- Log T 
 
 ransform 
 
 Semi- Log T 
 
 ransform 
 
 coefficient 
 
 t- value 
 
 coefficient 
 
 t- value 
 
 coefficient 
 
 t-value 
 
 SIC (1) 
 
 0.050 
 
 +0.2 
 
 1. 118 
 
 +2. 1 
 
 1.495 
 
 +3.2 
 
 SIC (2) 
 
 0. 091 
 
 +0.3 
 
 0.837 
 
 +1. 6 
 
 1.285 
 
 +3.0 
 
 ACT (1) 
 
 -1.037 
 
 -3.2. 
 
 -0.989 
 
 -2.0 
 
 -0. 142 
 
 -0.4 
 
 ACT (2) 
 
 -1.238 
 
 -3.8 
 
 -1. 074 
 
 -2.0 
 
 deleted 
 
 
 ACT (3) 
 
 -1. 230 
 
 -3.7 
 
 -1.330 
 
 -2.5 
 
 -0.468 
 
 -1.3 
 
 TIME 
 
 -0.421 
 
 -0.7 
 
 deleted 
 
 
 deleted 
 
 
 EDA TO 
 
 0. IL2 
 
 +0.8 
 
 0.700 
 
 +3.5 
 
 0.654 
 
 +3.3 
 
 EDAPR 
 
 2.358 
 
 +L. 6 
 
 3^01.7 
 
 +1.3 
 
 deleted 
 
 
 POP 
 
 -0.295 
 
 -1. 3 
 
 -1.375 
 
 -4, 2 
 
 -1.438 
 
 -4.7 
 
 A POP 
 
 Q.QI6 
 
 0.0 
 
 deleted 
 
 
 deleted 
 
 
 MEG 
 
 -0. 134 
 
 -0. L 
 
 -1.806 
 
 -2.2 
 
 -1.453 
 
 -1.8 
 
 M/ED 
 
 2.508 
 
 +2.5 
 
 6.448 
 
 +3. 9 
 
 5.795 
 
 +3. 9 
 
 %FEM 
 
 -I. 357 
 
 -0.6 
 
 3 . 645. 
 
 +0. 9 
 
 deleted 
 
 
 M/AG 
 
 0.803 
 
 +0.4 
 
 deleted 
 
 
 deleted 
 
 
 M/WG 
 
 0. 985 
 
 +0. 5 
 
 -Z. 909 
 
 - .9 
 
 deleted 
 
 
 %MG 
 
 1.9Q3 
 
 +1. 8 
 
 5'. 668 
 
 +4.3 
 
 5.899 
 
 +4. 5 
 
 %INC 
 
 -0. 767 
 
 -Q.6: 
 
 -5.7T6 
 
 -3.0 
 
 -3.821 
 
 -2.3 
 
 %PUT 
 
 1.413 
 
 +LZ 
 
 1. 781 
 
 +1.0 
 
 1.836 
 
 + 1.2 
 
 CONSTANT 
 
 -3.210 
 
 
 -11.775 
 
 
 -10. 117 
 
 
 R 2 * 
 
 .74 
 
 
 .85- 
 
 
 .80 
 
 
 ♦ Uncorrected for degrees of freedom 
 
 5.13 
 
deleted from the regression, almost certainly would be negatively- 
 associated with the success measure. 
 
 3. Size of loan is significantly associated with the success measure 
 in a positive exponential relation as shown on curve segment AB in 
 Figure 3. This finding, derived as "part of the package" of the overall 
 multi-variate regression, indicates that the benefit-cost ratio increased 
 as size of loan increases but this pattern must reach diminishing re- 
 turns at same scale of loans. 
 
 AY 
 
 Figure 3: Relation of Size a€ Loan, tor Size of Loan Payoff 
 However, the evidence indicates that diminishing returns do not set in 
 until some point beyond the largest loans of those included in this study. 
 Although most of the observed EDA Loans undoubtedly lie on curve segment 
 AB, the statistical technique used in this analysis is not sufficiently re- 
 fined to identify whether the curve begins to tail off at its upper point, 
 continuing onto segment BD, or continues its increasing. path onto segment 
 BC. An attempt to clarify this far purposes of identifying the optimal 
 size of loan implied in the data is presented in Chapter 6 of this study. 
 
 5.14 
 
4. As might be expected, the size and quality of the county labor 
 force has an important impact upon the magnitude of success. * Both 
 the educational attainment of the labor force, and the proportion of t : ie 
 labor force in manufacturing occupations are very significantly and 
 positively associated with degree of success. 
 
 5. Quite surprisingly, the degree of success is inhibited rather 
 than enhanced by community size and growth. Larger populations, and 
 higher income growth rates along with net immigration, are associated 
 with a smaller degree of success as measured by the aggregate payroll 
 of the loan-recipient firm. This finding, which has important implica- 
 tions for the "equity versus efficiency" controversy, ** is shown in all 
 three regressions of Table 5. is consistent with the negative impact of 
 community economic growth upon success of loans. It is further supported 
 by noting that Table 2 shows, that lower wages, together with higher educa- 
 tion (indicating lahor force quality), enhances loan success. However, 
 Table 2. indicates that the si2re of population enhances the probability of 
 success, in contrast to its dampening effect on the magnitude of success 
 shown, in Table 5. These findings indicate that industries thrive in pro- 
 
 * This result was also indicated in a recent report by CONSAD Researcr. 
 Corporation, A Study of the Effects of Public Investment , for the Economic 
 Development Administration, 19&9, Chapter III. 
 
 ** A number of economists have argued that loans in higher growth cen- 
 ters will be more efficient, but that since the declining areas are thought to 
 be in greater need, are counter-equity. But the evidence here indicates that 
 loans to declining areas are both equitable and efficient. In discussion with a 
 consultant to the pulp and paper industry, Mr. Meyer Hyman, a locational 
 rationale for industrialists considering location in relatively rural regions is 
 the existence of a sizeable and experienced labor, force in residence. This 
 labor force often includes a large number of out-commutcrs formerly employ- 
 ed in a now defunct local industry, and can be identified through simple sur- 
 vey techniques. 
 
tected environmental situations, where a good labor force is available 
 and where other industries are leaving thereby reducing the competition 
 for this labor force and, of course, easing the pressure on wages. However, 
 firms located in larger towns, again in a declining situation, are less 
 likely to default on the loan. 
 
 Attention now turns to the analysis of the determinants of employment 
 change in EDA loan-recipient firms. The application of least- squares 
 
 regression to all variables in linear form yieldedpoor results, with an 
 
 2 
 
 R of only .48 as shown in columns 1 and 2 of Table 6. After several tests 
 
 with varying logarithmic transformations of the variables, the most 
 appropriate form of the equation was, found to be exponential (semi-log 
 transformation, log of dependent variable only) as in the preceedmg (AY) 
 equation. 
 
 Eliminating: three variables from the regression (SIC(2J., TIME, 
 
 2 
 
 and %FEM), an R of .72- was obtained. Once again, however, the scarcity 
 
 of observations (36) necessitated, further reduction: in the number of variables, 
 
 and another four variables were therefore deleted (APOP, MIG, average 
 
 2 
 
 wage and age). The resulting semi— lag regression provided an R of . 67. 
 
 The pattern of results in terms of the significance of the included variables 
 is similar to that found in the AY regression, with two exceptions: 
 
 1. Firm classification, rather than SIC-type has a significant 
 effect upon employment. New firms: receiving EDA loans contribute less 
 to employment objectives than other types. 
 
 5.16 
 
TABLE 6 
 
 Regression with AE as Dependent Variable 
 Number of Observations: 36 
 
 Variable 
 
 No. Transform 
 
 Semi- Log - Transform 
 
 Semi- Log Transform 
 
 coefficient 
 
 t- value 
 
 coefficient 
 
 t- value 
 
 coefficient 
 
 t- value 
 
 SIC (1) 
 
 0.099 
 
 +0. 1 
 
 . 572 
 
 +1.5 
 
 .388 
 
 + 1.2 
 
 SIC (2) 
 
 -0.075 
 
 -0. 1 
 
 
 
 
 
 ACT (1) 
 
 -1.968 
 
 -0. I 
 
 -1. 362 
 
 -2.2 
 
 -1.531 
 
 -2.6 
 
 ACT (2) 
 
 -2. 121 
 
 -1.6 
 
 -1. 157 
 
 -1.6 
 
 
 -1.4 
 
 ACT (3) 
 
 -1. 532 
 
 -1.3 
 
 -1. 237 
 
 -1.9 
 
 -1. 215 
 
 -1.9 
 
 TIME 
 
 -0. 126 
 
 -0. 1 
 
 
 
 
 
 EDATO 
 
 0. 153 
 
 +0. 3 
 
 .581 
 
 +Z. 3 
 
 .481 
 
 +2.0 
 
 EDAPR 
 
 3.706 
 
 +0.7 
 
 3. 022- 
 
 +1.3 
 
 2. 883 
 
 + 1.3 
 
 POP 
 
 -0.593 
 
 -Q.6 
 
 -1. 263 
 
 -2.. 8 
 
 -1. 127 
 
 -3.0 
 
 APOP 
 
 7.696 
 
 +0.8 
 
 5.858 
 
 +1.3 
 
 
 
 MTG 
 
 -7.537 
 
 -0.8 
 
 -7. 091 
 
 -1.5 
 
 
 
 M/ED 
 
 6.679 
 
 +1.7 
 
 5. 235 
 
 +2.5 
 
 5. 243 
 
 +2.6 
 
 %FEM 
 
 -8. 196 
 
 -0.9 
 
 
 
 
 
 M/AG 
 
 6.384 
 
 +0.8 
 
 4.784 
 
 + L3 
 
 
 
 M/WG 
 
 -0.729 
 
 -0.1 
 
 -2. 605 
 
 -0.8 
 
 
 
 %MG 
 
 8.331 
 
 +Z. 1 
 
 7. 395 
 
 +3.7 
 
 6. 654 
 
 +4.3 
 
 AINC 
 
 1.574 
 
 +Q..3 
 
 -5. 264 
 
 -2.3 
 
 -4.251 
 
 -2.1 
 
 %PVT 
 
 4.750 
 
 +1.0 
 
 3. 457 
 
 +1.3 
 
 4. 641 
 
 +2.5 
 
 CONSTANT 
 
 -10.046 
 
 
 -11. 054 
 
 
 -9. 251 
 
 
 R 2 
 
 .48 
 
 
 .7Z 
 
 
 .67 
 
 
 5.17 
 
2. The degree o£ poverty has a favorable effect upon the degree of 
 success in terms of employment generated by the program. This finding 
 is in keeping with the "protected environment" finding of the AY analysis. 
 
 C. Conclusion 
 
 Techniques of multivariate statistical. analysis were applied to the 
 identification of factors underlying (1) the probability of success and (2) 
 the degree of success. Three classes of determining variables influencing 
 both types of success were selected: characteristics of loan- recipient 
 firms; characteristics of the loan; and socio-economic characteristics of 
 the county in which the Loan -recipient firm was located. All three types 
 of variables w*r* foutti to significaLnily affeci success, although in differ- 
 ing ways. 
 
 The probability of success was found to be favorably influenced by 
 larger loans, larger populations, and lowerrwages, and negatively influenced 
 by conditions of economic growth, by time, and by such characteristics 
 of firms as new entry, and resource orientation. A discriminant function 
 was calculated to provide a means; for estimating the success likelihood of 
 future loan applicants. 
 
 The magnitude of success as measured_by income was found to be 
 favorably influenced by sixe of loan, by firms.that are resource-oriented 
 (an effect opposite that in the gu cmasi* probability case), and .by labor force 
 
 5.18 
 
quality characteristics including higher educational attainment and 
 higher proportions of manufacturing occupations. 
 
 Factors inhibiting the level of income generated by loan- recipient 
 firms include an environmental situation characterized by higher growth 
 rates, higher out-migration rates, and higher populations. This pattern 
 is repeated when the level of employment generated by firms is analyzed 
 with the addition that community environments characterized by a higher 
 degree of poverty are also a significant element of a highly favorable and 
 protected labor force climate for firms who can gain a degree of monopsony 
 with respect to hiring. 
 
 5.19 
 
VI. EVALUATION OF THE EDA BUSINESS LOAN PROGRAM 
 
 A. The Selection of Evaluation Criteria 
 
 Traditionally, program evaluation has been carried out using the 
 techniques of cost-benefit analysis. In such analysis., a program may be 
 said to be acceptable if the ratio of its social benefits to its social costs 
 exceeds unity. 
 
 In a recent study of the EDA Business Loan Program, * a series of 
 benefit-cost ratios were calculated (under/ various assumptions) and the 
 worth (benefit-cost ratio) of the program waa estimated, to be.norgreater 
 than 25 to 1 and no less than IZ to 1. The said study, which: was: extremely 
 thoughtful about contingencies arrived at these promising figures by employ- 
 ing a thorough cost-benefit procedure which entailed estimation of the cur- 
 rent value of the entire future income stream that could be expected to be 
 generated by the loan- recipient firms. These estimates wererthen divided 
 by the entire cost stream ahsorbed by the program in- terms -of EDA budgets 
 
 and the social opportunity casts of same. 
 
 Given such a worthwhile program and these high estimates of 
 
 net social benefits, it is somewhat puzzling that net private benefits 
 
 would not be at least as great as the interest rate, and high enough 
 
 to attract more private activity than exists. The problem, 
 
 * Stanley Miller, D. Gaskins, and Charles Liner, Evaluation of 
 the ARA-EDA Loan Program , Office of Economic Research, Economic 
 Development Administration, November,. 1.968. 
 
 6.1 
 
of course, is in the selection of items to be included in the benefit stream. 
 The Miller-Gaskins-Liner study included all expected payrolls and multiplier 
 effects discounted over an infinite time horizon. A question might be 
 raised as to whether all payrolls should be included or whether only those 
 paid to employees hired after the date of the loan approval are relevant. 
 Although many employees hired after: the loan would be as replacements 
 for pre-loan labor, the least that should be done would seem to be to ex- 
 clude personnel obviously not part of the loan- induced expansion. A number 
 of additional offsetting factors were considered in the MGLStudy, but these 
 were either heavily discounted or- dismissed. Among the mitigating factors 
 so considered were: 
 
 1. The likelihood that the firms: would have located in target areas 
 without the EDA loan and, in addition-, that even if the firm did not locate 
 there, some growth would have occurred anyway; 
 
 2. The evaluation of benefits; in terms of the achievement of specific 
 program goals, such as the income incident to poverty populations. 
 
 These two paints are indeed: critical to the evaluation; but having once 
 raised them, the MGL. Study dismisses the first as unlikely and presents 
 a. worthy but unsupported argument .against the second, allowing for as 
 little as zero or as much as fifty percent reduction in benefits, still leaving 
 a 15 to 1 or 6 to 1 benefit-cost ratio. 
 
 The first paint can be summarized as "the likelihood that growth 
 would have occurred even without the EDA loan. " The nature of the indus- 
 tries receiving loans is that theirr location decisions are almost always 
 
 6.2 
 
dependent upon supply rather than demand conditions at alternative locations. 
 This being the case, there must be an external, or national, demand that 
 these firms can serve from the candidate locations. A firm influenced 
 to accept a location in a depressed area as a result of an EDA loan, there- 
 fore, would likely have located in another area, with much the same output 
 level, without the loan. This would not be true if all EDA loans were, in 
 fact, loans of last resort. In this case, the loan can be considered a subsidy, 
 having the effect of reducing marginal costs of production and thereby increas- 
 ing aggregate output, depending upon the opportunity costs of the subsidy. 
 Treatment of this question, lies beyond the scope of the present study -- there 
 is no way to assess the likelihood that the- loans are loans of last resort. But 
 even if they are not, the success of the loan in influencing such firms to lo- 
 cate in depressed areas contributes to the specific goal of poverty reduction 
 in target areas. This, however/, does not satisfy the question of whether the 
 loans have generated a net increase in social benefits or^in^he words of 
 the MGL Study,, "pergonal income: streams foregone by society (relative) 
 to the personal income generated in. the designated area,"* Itmay be that 
 the loans contribute nothing to net social. product. In this case r according 
 to the MGL. Study criteria, the loan.program should be discontinued. But 
 many federal programs such as ED A are: not established to contribute to the 
 GNP but, rather, to a chie ve certain social. goals (even at the expense of the 
 GNP). It could be argued that the minimization of employment through govern- 
 ment programs would entail an inefficient use of national resources with 
 
 * Ibid . , page ZZ. 
 
 6.3 
 
respect to the "GNP maximization" criterion. The point being made here is 
 that legislated programs should be evaluated in terms of the achievement of 
 specific goals, whatever they may be, rather than in terms of the traditional 
 cost-benefit criterion. Thus, to influence firms to locate in depressed areas 
 -- toward the end of achieving a legislated goal -- may be worthwhile strictly 
 in these terms, even if those firms would have generated equivalent income, 
 without the loans, through location in other areas. The evaluation criterion 
 implied in a goal-oriented procedure is. similar to that of cost-effectiveness 
 analysis, where hi measures of effectiveness are stated along a different 
 dimension than are measures of cost. The approach essentially makes 
 program evaluation meaningless in the absence of alternatives against which 
 to draw comparisons, although, in fact, a review of the applications of the 
 more traditional cost-benefit approach suggests the same conclusion. Never- 
 theless, a body of evaluations and underlying data is currently being developed 
 in. the area of economic development programming,* andthe time is near at 
 hand when a consistent set at evaluational data will be available for comparing 
 a series of program alternatives;. The present study should.be treated in this 
 context. For this reason, the data, to: be. presented in this chapter will be 
 along a scale consistent with that used- in. the. Miller -Gas kins -Liner Study, as 
 well as along scales that are necessary forr several types of cost-effectiveness 
 analysis that may be necessary if several different types of programs 
 oriented toward economic development are to be compared. 
 
 * Office of District and Area Planning, Economic Development 
 Administration. 
 
 6.4 
 
At this point, attention may be given the treatment accorded by 
 the MGL Study to the second question raised above, that of program evalua- 
 tion in terms of goal achievement. In the present program, the legislation 
 states that the reduction of unemployment and of poverty in depressed areas 
 are specific goals to be pursued by EDA, although it broadens the goal set 
 by stating that the achievement of a general employment increase toward 
 "economic development" is also an objective. * The MGLStudy notes that 
 if the criterion is redefined as "income increases to the unemployed and under- 
 employed, " then it "should not change matters if one assumes that jobs 
 vacated by -workers formerly employed and now working in ARA and EDA- 
 assisted plants will, inmost cases, be filled by the underemployed, un- 
 employed, or by workers moving internally or from other firms, and thus 
 
 vacating jobs for the unemployed, underemployed, and non-labor force 
 members. "** 
 
 This argument, which, i-sr certalrdy reasonable, provides. the analyst 
 with a rationale for assuming that those program impacts reaching any 
 groupwill be equivalent to the impact thai reaches the target group. Thus, 
 for example, suppose that, out of a total. wage bill of $500, 000, only 
 $100,000 in annual earnings were paid by loan-recipient firms to the for- 
 
 * Public Law #90-103, The Public Works and Economic Development, 
 Act of 1965, as amended. 
 
 ** Miller, Gaskins, and Liner., ape, cit. , page. 21.. 
 
 6.5 
 
merly unemployed, with the remaining $400,000 being paid to people formerly 
 employed in other firms. In this case, then, instead of crediting the loan 
 program with only $100, 000 in direct current benefits, the assumption 
 that the trickle-down effect will operate to generate jobs for the unemployed 
 equal in number to the jobs provided by the loan- recipient firms to the 
 formerly employed makes it possible, to credit the loan program with the 
 entire $500,000. Given, that the gain is 400%, the assumption is certainly 
 worthwhile. But, before examining the empirical evidence on the size of the 
 gain provided, by the assumption (the MGL Study pegs it at 100%), there 
 should be considered the empirical evidence concerning the assumption itself. 
 
 In a recent CONSAD study, * data on unemployment, ** employment, 
 and labor force changes in a large number of rural counties over the period of 
 1963-66 were analyzed, taking account of local factors related to migration and 
 labor force participation, to determine the average county-wide unemployment 
 change associated during this period with a unit employment increase. It should 
 be noted that in a closed economy with no labor force change, a unit increase 
 in employment would result in a unit reduction in unemployment. The analysis 
 showed that a unitincrease in employment has been associated with an average 
 
 * CONSAD Research Corporation, A Study of the Effects of Public 
 Investment, prepared for The Economic Development Administration, 
 1969, Chapter IV. 
 
 ** Unemployment includes only those counted as participating in the 
 labor force. The data used was obtained by State Bureaus of Employment 
 Security. 
 
 6.6 
 
of . 1 1 decrease in unemploment, migration and labor force participation con- 
 sidered, or that approximately ten new jobs are required to cause county- 
 wide unemployment to drop by one. Clearly, for the MGL hypothesis to 
 hold, it must be assumed that 1) new jobs absorbed by new labor force par- 
 ticipation are just as valuable to EDA as the jobs that would otherwise have 
 been absorbed by the local unemployed --an assumption which may or may 
 not be warranted, but that in any case begs the question -- and 2) jobs in 
 other counties, vacated by the in- migrants who absorb the new local jobs, 
 do not disappear but instead go to the unemployed of those counties. 
 
 To consider these assumptions, let us review the data shown in 
 Chapter II of this study. Of the 1, 817 new jobs provided by the 40 loan re- 
 cipient firms, 196 or 11 percent went to people who were formerly unemployed -- 
 afigure precisely matching that found in the earlier CONSAD study. But here, 
 data is available for explicitly considering the direct migration effects. Only four 
 percent of the previously unemployed had in-migrated from another country; 
 only six percent of all full-time employees were in- migrants. Evidently, 
 in-migration accounts for but a small proportion of the immediate employ- 
 ment effect; and, given the superior visibility to a dispersed labor force 
 of a new plant being constructed in a rural area, it would seem that this 
 in-migration rate would be higher for the immediate effect than for the 
 local firms involved in the secondary trickle-down (or "musical-chairs") 
 effect. Yet, it is conceivable that the "musical- chairs "effect, even with 
 a declining in-migration rate (combined with the 11% unemployment reduc- 
 tion rate)could result in an unemployment reduction equal to the immediate 
 
 6.7 
 
employment increase. Far more likely, however, are the two other 
 possibilities, the second of which must be stressed: 
 
 1. Many of the jobs will be taken up by new labor force entrants; 
 which are not counted. 
 
 2. Many of the jobs will disappear due to the resulting scarcity 
 of labor, i. e. , the musical-chairs effect will be truncated. 
 
 The first of these, of course, accounts for a significant portion 
 of the immediate employment in the loan-recipient firms (11%), and while 
 most of these jobs benefited people who are not heads of households (82%), 
 they might well be considered a contribution to goal achievement even 
 though preferred program incidence would: be upon previously unemployed 
 heads of households. 
 
 The second effect is an absolute contraction in employment in the 
 local firms that lost labor to the new plant. This effect has long been a 
 
 focus of interregional trade theory, although one generally ignored in 
 regional growth economics. 
 
 To see this effect, consider Figure 1, in.which the local .labor demand 
 function (DD) is shown to have shifted rightward (D'D 1 ) as a result of the 
 entry of the EDA-assisted plant. 
 W| 
 
 _ L 
 
 o H 
 
 Figure 1: Effect of New Entry Upon Local Wage Rate 
 
 6.8 
 
Now, assume that a local wage change affects the local labor sup- 
 ply function (SS) by affecting labor-force participation decisions only 
 and, further, that migration has the effect of shifting the labor supply 
 schedule, as shown in Figure 2. 
 
 W 
 
 out-migration 
 
 I in-migration 
 
 -L. 
 
 Figure 2: Effect of Migration Upon Local Labor Supply Function 
 
 Next, assume that migration is. also a function of the local wage 
 rate. Then, as the local wage rate increases assuming wages in other 
 regions to be constant, the local labor: supply function shifts j-ightwird, as 
 shown in Figure Z.. Itisithus possible to trace the cross -cut migration- sensitive 
 
 Figure 3: Migratiori-Sensitive Labor: Supply Function 
 
 6.9 
 
labor supply curve S S . Attention is directed to the elasticity 
 
 mm ' 
 
 of this cross-cut supply curve. 
 
 If tb« migration response to wage rate shifts were perfectly respon- 
 sive, then the cross-cut supply function would be infinitely elastic. An 
 increase in labor demand of.AL. due to new entry would result in an 
 equilibrium employment of L + AJL, or L,., as shown in Figure 4. 
 
 AL 
 
 Figure 4: Equilibrium Local Employment Assuming Infinitely Elastic 
 Migration -Sensitive L»abor Supply Function 
 
 But in depressed are as where strong out-migration patterns exist and 
 ■where perfect mobility in-response to economic conditions is not typical, it is 
 unrealistic to assume that the in-migration response will.be such as to 
 a*sure a perfectly elastic cross— cut labor supply function. A more 
 credible assumption is that of a less-than-perfectly elastic supply function, 
 
 S r S r of Figure 5. Their,. a.siriff:in-the labor demand function due to new 
 
 m m 
 
 entry of a firm employing AL workers will result in an equilibrium employ- 
 
 6.10 
 
ment level of L ? < L. . 
 
 Figure 5: Employment Equilibrium With Less -Than- Perfectly Elastic 
 Labor Supply Function 
 
 If the new entry is able to retain its. entire employment requirement 
 of AL as the wage rate increases from W to W., then employment in 
 local pre-existing firms will contract by an amount L. - L-. 
 
 Evidence for this effect is to be found in Chapter IV of this study, 
 in which the degree of program success in generating employment and 
 income increases was shown to be adversely affected by local conditions 
 acting to increase the competition for county labor. 
 
 From this discussion, it should be clear that the "musical-chairs" 
 assumption cannot be accepted at face value. In the same vein, the assump- 
 tion that unemployment reduction among the target population will rise to 
 equal the aggregated immediate employment impact is likewise unaccept- 
 able. Instead, it will be necessary to explicitly consider the incidence of 
 program benefits upon each target population and with respect to each goal. 
 
 6.11 
 
The evaluation of the Business Loan Program will be guided by this 
 requirement. However, for purposes of comparison, the assumptions 
 of the Miller- Gaskins -Liner Study will also be given, following the 
 evaluation stated in terms of goal-achievement. 
 
 B. Program Evaluation with Incidence Criteria 
 
 Objectives of the Economic Development Administration include the 
 reduction of poverty and of unemployment in depressed counties that meet 
 the designation requirements (which include an unemployment rate of at 
 least six percent or median family income of less than 40 percent of the 
 national average). Program benefits in this section of the evaluation will 
 therefore be based upon the degree to which these objectives are being 
 achieved. 
 
 Three types of measures for program evaluation will be shown 
 following the estimation of program benefits: 
 
 1. Current value of the gross increased income stream* to all 
 groups; 
 
 2. Current value of the gross increased income stream incident 
 to the poverty population divided by the current value of the cost stream, 
 generated by the program; 
 
 * A preferred measure would be the current value of the net income 
 stream incident to the target population generated by the Loan Program, 
 divided by the current value of program costs, wheje net income is the 
 increase in earnings less opportunity costs. 
 
 6.12 
 
3. The number of people for whom program goals are achieved 
 (previously employed or earning less than $4000 per year) relative to 
 the discounted value of the cost stream. 
 
 1. Benefit-Cost Evaluation in Terms of Income Impacts. 
 
 a. Benefit Estimation Based Upon Gross Income Change to 
 all Groups - The total annual income earned in the 40 
 loan-recipient firms that -were surveyed amounted to 
 $7,747,600. Considering earnings from other sources 
 both before and after hire, this represented an aggre- 
 gate increase to all employees of $2, 530, 900. If the 
 "musical-chairs" effect is entirely discounted, the 
 direct aggregate benefit then amounts to the $2. 5 mil- 
 lion figure. The remaining $5. 2 million being treated 
 as the amount that would have been earned in the absence 
 of the Business Loan Program. If, however, it is 
 assumed that the "musical-chairs" effect is 50% effec- 
 tive, * then the direct aggregate benefit would amount 
 to $5, 039, 250, as shown in Table 1 . 
 
 Added to th*se direct income impacts are the indirect 
 income impacts that result from the multiplier process. 
 Using sectoral earning data for 90 counties** in which 
 EDA and ARA Business Loans have been made (including 
 the 40 firms surveyed), it was found that an exogenous 
 change in county income of $100 would result in an addi- 
 tion of $86 earned in internal sectors (including county 
 sectors serving consumer and interindustry demands). *** 
 This represents a multiplier of 1. 86. Applying this 
 multiplier to the aggregate change in direct earnings, 
 an aggregate direct and indirect earnings estimate of 
 
 * Section B of this Chapter treats the contingency of 100% 
 effectiveness. 
 
 ** The sectoral earnings data was obtained from county estimates 
 for 1962 and 1966 by the Office of Business Economics, Department of 
 Commerce. 
 
 *** See equation (3 1 ) in Appendix B of this study. 
 
 6.13 
 
Table 1 
 
 Direct Income Change Impacts 
 
 
 "Musical-Chairs" 
 Ineffective 
 
 "Musical-Chairs " 
 50% Effective 
 
 Aggregate Direct Change 
 Direct Incidence to Pool 
 
 i 
 
 $2,530,900' 
 
 $5,039,250 
 $1,5*2,010 
 
 6.14 
 
 ( A 
 
$4, 704, 700 is obtained. Assuming the existence of a 
 partial "musical-chairs" effect, the direct and indirect 
 aggregate impact would be $9, 368, 000. These figures 
 are shown in the first row of Table 2. 
 
 Insofar as these figures measure increased earnings to 
 the unemployed, underemployed, the poor, and, in a 
 broader sense, to the underutilized, they may be inter- 
 preted as the gross immediate contribution to program 
 goals. 
 
 b. Benefit Estimation Based Upon Gross Income Change 
 
 Incident to Poverty Population - Of the $2. 5 million 
 in direct aggregate benefits, 31 percent or $773,300 
 was earned by employees with incomes of under $4,000 
 per year. If this same proportion is applied to the $5. 2 
 million figure, the benefits incident to the poor amounted 
 to $1, 542,000. These figures are shown in columns 1 
 and 2 of Table 1 . 
 
 A question now arises as to the incidence of the direct and 
 indirect effects upon the poor. It is tempting to apply the 
 ratio of poor to aggregate earnings (.31) to the indirect 
 effects and thereby to obtain the low income multiplier effect, 
 but to do so would be highly misleading because, as further 
 reflection would indicate, the incidence distribution of the 
 indirect effects would depend largely upon the structure 
 of the county economy and only to a small extent upon the 
 distribution of the direct benefits by the loan-recipient 
 firms. In fact, since it might be guessed that the ratio 
 of entrepreneurs to total employees in the consumer sectors 
 would be relatively high, it would seem that the higher 
 income groups of the county economy would gain a larger 
 proportion of the indirect impact than of the direct. 
 Partitioning the county income groups into low (less than 
 $4,000), high (over $10,000) and middle income, a matrix 
 multiplier was developed to explicitly treat this effect. * 
 
 * The incidence multiplier is described in detail in Appendix A. 
 
 6.15 
 
Table 2 
 
 Direct and Indirect Income Impacts 
 
 
 "Musical-Chairs" 
 
 "Musical-Chairs" 
 
 
 Ineffective 
 
 50% Effective 
 
 Aggregate Direct and 
 
 $4,704,700 
 
 $9,368,000 
 
 Indirect Change 
 
 
 
 Direct and Indirect 
 
 $1,157,460 
 
 .$2,332,600 
 
 Incidence to Poor 
 
 
 
 6.16 
 
While the high income group earned 12 percent of the 
 direct impacts, it earned an estimated 31 percent of the 
 indirect impacts,* or 2-1/2 times its direct impact 
 rate. Faring less well in the outcome was the low in- 
 come group, which earned but 16 percent of the indirect 
 impacts, or one-half its direct earnings rate. Applying 
 these derived percentages, the direct plus indirect 
 program impacts incident to the poor amounted to $1 , 157, 460. 
 Assuming the partial "musical-chairs" effect, the direct 
 and indirect impacts to the poor would be $2, 332,600. 
 These estimates are shown in row 2 of Table 2. 
 
 These estimates provide the annual income impact gener- 
 ated by the EDA loan program. However, it may be assumed 
 that these benefits will be generated by the EDA loan- 
 recipient firms into perpetuity. 
 
 In the Miller- Gaskins- Liner Study, the effects of the real 
 growth of the economy upon the benefits generated by the 
 loan recipients was considered by assuming that these 
 firms would experience a growth rate of either zero or 
 two percent per year, and that a discount rate of either 
 4, 6. or 10 percent rr.:ild be =.£ ruined. Here are considered 
 only the two percent growth rate (x) and a low discount 
 rate (4) of 6 percent and a high of 10 percent. (4% dis-. 
 count rate is currently unrealistic. ) Inserting these para- 
 meters into equation 1, two alternative current values of 
 the future benefit stream (Y ) may be estimated for the 
 figures in Table 2. The results are shown in Table 3. 
 
 «oo (x - r)* ** 
 
 y K = s b Y fc dt 
 
 These benefits mustnowbe divided by the current value of 
 the cost stream arising from the Business Loan Program 
 to obtain the measure of program performance. 
 
 * These percentages are obtained from Chapter II of this study. 
 They are applied to the matrix multiplier (Appendix A) to obtain the 
 incidence distribution of the multiplier effect. 
 
 ** . In the MGL Study, the benefits generated in the first six years 
 were ignored due to lack of data. 
 
 6.17 
 
Table 3 
 
 Current Value of Future Benefit Stream 
 
 Discount Pate 
 
 6% 
 
 10% 
 
 "Musical- Chairs" 
 
 None 
 
 50% 
 
 None 
 
 50% 
 
 Aggregate Benefit 
 
 Benefits Incident to 
 j Poor 
 
 $85, 507,922 
 $21,036,836 
 
 $170,263,400 
 $ 42, 395,005 
 
 $35,405,322 
 $ 8,956,425 
 
 $72,489, 584 
 $18,049,658 
 
 6.18 
 
c. Estimation of Program Costs - Program costs are 
 comprised of (1) losses due to default on loans, (2) ad- 
 ministrative costs, and (3) the social opportunity costs 
 of pre-empting the private sector from seeking its own 
 highest use of the funds. As in their benefit estimates, 
 the Miller-Gaskins-Liner Study provides several cost- 
 estimates -- low, high, and reasonable. Since the low 
 estimate made use of a 4% discount rate, which is un- 
 acceptable, only the assumptions used in the high and 
 the "reasonable" cost estimates need be considered here. 
 
 The high, or most conservative, cost estimate as calcu- 
 lated in the MGL Study is based upon an estimated default 
 loss rate of 18 percent of the disbursed loans, where returns 
 from liquidated assets are included in the estimate. In 
 addition, a discount rate of 10 percent is assumed. 
 
 Applying these assumptions to a total of $23, 180, 000 in 
 
 loans to the 40 firms, default losses would amount to 
 
 $4,172,400, administrative costs would amount to 
 
 $2, 874, 102, * and there would be an interest cost of 
 
 $6, 082, 769. #* The combined discounted value of the 
 
 cost stream generated fey the EDA Business Loan Program 
 
 in financing the 40 surveyed firms is therefore $13,129,271. 
 
 The "reasonable" cost estimate is based upon an assump- 
 tion of a ten percent default loss rate and a six percent 
 discount rate. Under these assumptions, default losses 
 would amount to $2,318,000, administrative costs would 
 amount to $3, 921, 795, and the interest cost would be 
 $2,857,373, to give a combined total discounted program 
 cost of $9, 097, 169. 
 
 d. Benefit Cost Ratios Based Upon Income Change Impacts - 
 It is now possible to calculate the benefit-cost ratios for 
 program evaluation purposes. However, where benefits 
 
 * Administrative costs were obtained by deriving the ratio of loans 
 held by the 40 firms surveyed in this study to the loans analyzed in the 
 Miller-Gaskins-Liner Study, and applying this ratio to the administrative 
 costs obtained in the MGL Study. 
 
 ** Prorated in the same manner as in deriving administrative costs, 
 
 6.19 
 
are discounted at ten percent, costs are discounted at 
 the same rate. Since costs, when discounted at ten 
 percent, include an 18 percent default loss assumption, 
 the benefit- cost ratio assuming a ten percent dis court 
 rate will be considered to be a conservative estimate. 
 "Reasonable" estimates assume a discount rate of six 
 percent, and a default loss rate of ten percent. For 
 those reasons, four benefit-cost ratios are estimated 
 based upon income change impacts incident to all groups, 
 and four are based upon income change impacts incident 
 only to the poor. These ratios are shown in Table 4. 
 
 For the case where benefits include impacts incident to 
 all groups, the most conservative benefit -cost ratio 
 estimated is 2. 7, and the least conservative is 18. 7. 
 
 For the case where benefits only include impacts incident 
 to the poor, the most conservative benefit-cost ratio 
 estimated is . 7, which is less than unity, and the least 
 conservative estimate is 4; 7. 
 
 Since, for the first case, the benefit-cost ratio exceeds 
 unity, the evaluation is clear. The benefits generated 
 by EDA Business Loan Program exceed the costs. 
 
 For the second case, the most conservative estimate shows 
 that costs exceed benefits. However, the traditional 
 criterion that the benefit-cost ratio must exceed unity must 
 not be taken at face value, since only benefits incident to 
 the poor are included in the estimate. This should be taken 
 to mean that impacts to the poor are accorded much greater 
 value than impacts to other groups, since non-monetary 
 social considerations are included in the valuation. Viewed 
 this way, these benefit-cost ratios are seen as not truly 
 comparable with ratios that do not include incidence criteria 
 (as is true in the first case). Therefore, the criterion that 
 the benefit-cost ratio exceed unity should not apply to the 
 case where incidence of benefits is restrained. Instead, 
 such ratios should be used for comparing alternative 
 programs, using comparative goal achievement per unit 
 cost as the criterion. 
 
 6.20 
 
Table 4 
 
 Benefit-Cost Ratios Based Upon Income Change Impacts 
 
 Discount Rate 
 
 6% 
 
 10% 
 
 "Musical Chairs" 
 
 None 
 
 50% 
 
 None 
 
 50% 
 
 
 
 
 
 
 Benefits to all Groups 
 Conservative Estimate * 
 
 "Reasonable Estimate** 
 
 9.4 
 
 18.7 
 
 2. 7 
 
 5. 5 
 
 Benefits to Poor 
 Conservative Estimate* 
 "Reasonable" Estimate** 
 
 2. 3 
 
 4. 7 
 
 . 7 
 
 1.4 
 
 ♦Conservative Estimates assume a default loss cost of 18 percent on disbursed 
 loans, and a discount rate applying to both costs and benefits of 10 percent. 
 
 ♦♦"Reasonable" estimates assume a default loss cost of 10 percent on disbursed 
 loans, and a discount rate applying to both costs and benefits of 6 percent. 
 
 6.21 
 
2. Evaluation Based Upon Number of People for Whom Program 
 Goals Are Achieved. 
 
 In the preceeding evaluation, program benefits were measured in 
 terms of income change and the distribution of income to low income 
 groups (earning under $4, 000). But this criterion is weak when goal 
 achievement is being evaluated. Although it measures income delivered 
 to the poor, EDA program goals are stated in terms of the reduction 
 of poverty and of unemployment. Therefore, a more appropriate measure 
 of program goal-achievement would be the number of people taken out of 
 poverty (income increasing above $4, 000 due to the loan) and off the un- 
 employment rolls. 
 
 Of the 1, 817 new jobs made available in the 40 loan-recipient firms, 
 a net of 289 jobs raised the income of employees formerly in the poverty 
 group over the $4, 000 line. In addition, 196 jobs were obtained by indi- 
 viduals who were formerly unemployed. However, 139 of the latter jobs 
 were obtained by individuals who previously were both poor and unemployed. 
 Therefore, 346 jobs, out of a total of 1,817, either raised individuals out 
 of poverty or out of unemployment, or both. Thus, the EDA program re- 
 sulted in 346 cases of direct achievement of program goals, a success rate 
 of 19%. 
 
 To the 346 jobs provided by the loan-recipient firms (the direct 
 impact) must be added the jobs created by the indirect impacts generated 
 by the multiplier process. 
 
 6.22 
 
Applying the employment (and unemployment) multiplier described 
 in Appendix B, a total of 2, 601 jobs would be created, directly and in- 
 directly as a result of the multiplier pTocess, of which 294 would be ob- 
 tained by the unemployed. Thus, the indirect impacts of the loan program 
 would provide 98 jobs to the formerly unemployed, in addition to the direct 
 impact of 196. Assuming that the ratio of total direct impact (on the 
 poor or unemployed) to the direct impacts on the unemployed (346:196) 
 will also apply to the indirect impacts, the total direct and indirect im- 
 pact of the EDA Business Loan Program upon goal-achievement (number 
 of people taken out of poverty or unemployment) would be 250. 
 
 It is conceivable that a two percent growth rate might be applied 
 to this figure with the future resulting in addition to £oa.l achievement 
 being discounted as in the preceeding income analysis; but, for the sake. 
 of credibility, this procedure will be avoided. The 520 figure will be 
 used as the program effectiveness measure. 
 
 In order to maintain the success rate into the future, the total future 
 costs must be considered. The same two cost figures estimated for the 
 preceeding evaluation will be used here --a reasonable estimate of $9. 1 
 million and a conservative estimate of $13. 1 million. Based upon these 
 figures, the cost-effectiveness ratio for the EDA Business Loan Program 
 is estimated at $17,500 per man or, more conservatively, at $25,200 per 
 man. These costs may be broken down into components, as shown in 
 Table 5. 
 
 6.23 
 
Table 5 
 
 Analysis of Program Costs per Unit Effectiveness 
 
 Component of Cost 
 
 Reasonable Estimate* 
 
 / 
 Conservative Estimate** 
 
 Default Loss 
 
 Administrative 
 
 Interest 
 
 $4,458 
 $7,542 
 $5,495 
 
 $8,024 
 
 $5,527 
 
 $11,698 
 
 Total 
 
 $17,495 
 
 $25,249 
 
 ♦Cost estimate based upon 6 percent discount rate and 10% default loss rate. 
 
 ♦♦Conservative cost estimate is based upon 10 percent discount rate and 18 
 percent default loss rate. 
 
 6.24 
 
It should be made clear, at this point, that these cost-effectiveness 
 ratios may be used only for comparing alternative programs intended to 
 achieve the same goals. There is no way of determining whether $17, 500 
 per man is too. high a cost to pay or whether it is a bargain; since criteria 
 for determining the "red zone" benchmark have not been presented. One 
 way of obtaining such a benchmark, which unfortunately lies beyond the scope 
 of this study, is to determine the dollar savings to society for taking an in- 
 dividual out of poverty or off unemployment. To do this would require estima- 
 tion of the current value of the future cost stream saved society in terms of 
 welfare payments, public housing, health programs, education, crime, 
 mass disorders and property damage, and the like, for each individual 
 taken out of poverty or unemployment. Also, new economic productivity of 
 these individuals that had previously been denied the economy must be added 
 to the other cost savings. If these estimates were obtained, the cost savings 
 to society generated by the program on a per person basis would almost surely 
 exceed the $17, 500 per person in program costs. 
 
 Another measure of program effectiveness might be shown by noting 
 that it took $23, 180, 000 in business loans to permanently reduce unemploy- 
 ment or poverty by 520, a loan cost of $44, 500 per person. 
 
 C. Program Evaluation in Terms of Aggregate Impacts 
 
 In this section, a cost-benefit evaluation based upon the discounted 
 total cost and total benefit streams will be shown. The purpose of this 
 
 6.25 
 
approach is to provide a result comparable with that obtained in 
 the Miller -Gas kins -Liner Study. In this sense, this section can be 
 viewed as an extension of the MGL Study, the main difference being 
 that 40 observations, rather than eight, are used for deriving the results 
 and that this study excludes employees hired by loan- recipient firms 
 prior to the date of loan approval. 
 
 The total direct annual impact of the EDA Business Loan Program, 
 as measured by the income paid to the 1, 817 employees of the 40 firms who 
 were hired after the date of loan approval, was $7, 748, 000. Although 68 percent 
 of the new employees were previously employed, and although only one- 
 third of the $7.7 million represented an income increase, it is believed 
 that the "musical chairs" effect will be complete -- the jobs released by 
 the previously employed will ultimately be filled by members of target 
 groups. Therefore, the entire $7.7 million will be counted as the direct 
 program benefit. 
 
 Applying the multiplier of 1. 859 to the direct benefit figure, the total 
 annual direct and indirect benefit is estimated as $14,420, 900. Assuming 
 a two percent annual growth rate, and a discount rate of 6 percent, the 
 current value of the infinite income stream is $262, 099, 858; with the same 
 annual growth rate, but a discount rate of 10 percent, the discounted bene- 
 fit stream is $111,449,400. These figures are in Table 6. 
 
 6.26 
 
Table 6 
 
 Total Cost and Benefit Impacts 
 (Musical Chairs 100% Effective) 
 
 Direct Impact $ 7, 748, 000 
 
 Direct and Indirect 14, 402, 900 
 
 Discounted Benefit Streams: 
 
 at 6% rate 262, 099, 858 
 
 at 10% rate 111,449,400 
 
 Discount Cost Stream 13, 129, 271 
 
 Benefit-Cost Ratio: 
 
 benefits discounted at 6% 41.26 
 
 benefits discounted at 10% 7. 19 
 
 6.27 
 
It can be seen that the benefit to cost ratio calculated by this 
 method is extremely high: the lowest, at an assumed growth rate 
 of two percent and discount rate of ten percent is 8, while the high 
 (discount rate of 6%) is 20. 
 
 D. Conclusion 
 
 In this evaluation of EDA's Business Loan Program, it was found 
 that: 
 
 1. By almost all counts, the program appears to make a worth- 
 while contribution toward achieving such economic development goals 
 as the amelioration of poverty and unemployment. Approximately 
 one-fifth of the jobs created served either to raise the income of an 
 individual above the poverty level or to take an individual off unemploy- 
 ment. 
 
 2. While the cost of achieving this goal may appear high ($17, 500 
 per individual formerly poor or unemployed), the several benefit-cost 
 ratios that were obtained all exceeded unity, with one exception, and that 
 being the estimate for the case where only income earned by the poor 
 was counted as benefit and very conservative assumptions (ten percent 
 discount rate, 18 percent default loss, no "musical-chairs" effect) 
 were used. Therefore, this evaluation indicates that the EDA Business 
 Loan Program is generating positive net social benefits for society. 
 
 6.28 
 
VII. OPTIMAL- SIZE OF LOAN 
 
 A major question for EDA policy is the size of loan that should be 
 given. If the total loanable funds available in any one year to the EDA 
 Office of Business Loans is fixed, then a policy of making small loans 
 implies that a large number of loans would be made during the year, 
 presumably over a widely dispersed geographic landscape. But if the 
 policy is to provide large loans, then a fewer number of loans will be 
 made during the year, although in a more concentrated geographic pattern 
 in order to generate more jobs per loan. 
 
 Now, although jobs per loan may increase as size of loan increases, 
 the annual number of loans that could be made under a fixed budget will 
 be reduced. For this reason it is not at all clear that by increasing 
 the size of loan, the total number of jobs generated by the fixed annual 
 budget would be increased. Actually, if the number of jobs increase 
 slowly as size of loan increases, then it may very well be the case that 
 a policy of larger loans per firm would reduce, and not increase, the 
 total annual number of jobs generated by the Business Loan Program. 
 In fact, as shown in Table 1, the ratio of total annual wages paid out 
 by the EDA loan recipient firms to the size of loan they obtained from 
 EDA declines as size of loan increases (this is seen in the negative sign 
 of the coefficient to EDATOT obtained through regression analysis of 
 the 40 EDA loan- recipient firms surveyed in this study). This finding 
 
 7.1 
 
Table 1 
 
 Dependent Variable - Income Per Dollar Loan* 
 
 Log of VAR 
 SIC1 
 
 EDATOT 
 60-POP 
 MIGRAT 
 MEDED 
 FEMALE 
 M-WAGE 
 LABOR 
 CH-INC 
 
 Coefficient 
 
 CONSTANT 
 2 
 
 • 
 
 513 
 
 m « 
 
 621 
 
 • « 
 
 953 
 
 -1. 
 
 706 
 
 3. 
 
 327 
 
 9. 
 
 409 
 
 -5. 
 
 ,206 
 
 _£ 
 
 ±i k 
 
 -6.681 
 
 6.005 
 .6305 
 
 T- Value 
 
 1. 
 
 47 
 
 2. 
 
 72 
 
 2. 
 
 48 
 
 2. 
 
 26 
 
 2. 
 
 47 
 
 2. 
 
 17 
 
 1.62 
 
 ■3 r.r\ 
 
 3.01 
 
 * Definition of variables, see Table 1 of Chapter 5, 
 
 7.2 
 
is strong evidence that as EDA loan size increases, the size of payoff as 
 measured in number of jobs or total payroll increases proportionately 
 less. However, this does not mean that smaller loans are preferred -- 
 only that at some point, which may be greater than the present loan size, 
 diminishing returns will set in. It is to the task of identifying that point 
 that this analysis is addressed. 
 
 It is our hypothesis that the income generated by an EDA loan- 
 recipient firm increases as size of loan increases. In addition, as loans 
 become increasingly large, the rate of increase in income diminishes. 
 Such a relation is shown in Figure 1. * 
 
 Income 
 per Loan 
 (Y) 
 
 XT — »/T \ 
 1 - J.\J-</ 
 
 inflection point: d B _ 
 
 dL, 2 
 
 Loan Size (L) 
 
 Figure 1: Benefits Per Loan as a Function of Loan Size 
 
 * Alternatively, these conditions maybe stated as follows 
 Y = f(L) 
 dY 
 
 d L >0; i£ < 0; 
 
 where Y is income per loan, and L is size of loan. 
 
 7.3 
 
From the analysis of Chapter 5, it was found that aggregate program 
 benefits (B), incidence not considered, could be stated in terms of the 
 aggregate direct wage and salary income generated by each loan- 
 recipient firm. This is done by first expanding the direct income 
 impacts (Y) by the multiplier (0) to obtain direct and indirect impacts, 
 and then expanding the result to account for impacts in all future 
 periods, properly discounted. If our hypothesis that income is a function 
 of size of loan (L) is correct, then aggregate program benefits may be 
 stated as: 
 
 a) 
 
 B = / 0f(L)e (x ~ r) dt, 
 
 
 where x is the expected annual growth rate of loan recipient firms, 
 r is the discount rate, and t is time period. 
 
 Another relation to be considered is that between program cost (C) 
 and size of loan. It is assumed that program cost can be estimated j.s 
 a linear and homogeneous function (a ) of loan size. This linear cost 
 function is shown together with the benefit function in Figure 2. 
 
 B,C 
 
 Figure 2: Program Benefits and Costs as a Function of Loan Size 
 
 7.4 
 
In identifying the optimal size of loan, the objective will be to 
 
 locate the point at which net benefits (n ), or the difference between 
 
 program benefits and program cost is at a maximum. This may be done 
 
 by stating the function for net benefits, in equation (2), 
 
 (2) n = f /9f(L) e [X T) dt -aL, 
 
 
 and locating the highest point on that function, L*, as shown in Figure 2, 
 
 and again in Figure 3. 
 
 n 
 
 = o 
 
 n = f( L ) 
 
 L* L 
 
 Figure 3: Net Benefits Stated in Terms of Loan Size 
 
 This may be done mathematically by taking the derivative of 
 equation (2) with respect to loan size and setting the result equal to 
 zero, thereby to obtain the point at which the slope of the net benefit 
 curve is stationary -- neither rising nor falling. However, care must 
 be taken to be certain that this is the maximum net benefit point (L*), 
 and not the minimum point (Lo) by assuring that the second derivative 
 is negative. 
 
 The task now confronting the analysis is to estimate the function 
 Y = f(L). 
 
 In Chapter 5 income impacts were analyzed in terms of three sets 
 of causal factors: characteristics of the EDA loans, characteristics of 
 
 7.5 
 
the firms, and characteristics of the locality in which the firms operate. 
 It was learned that size of loan did indeed have a positive and significant 
 impact upon income impacts. But it was also found that characteristics 
 of local economies, particularly those indicating that a loan-recipient 
 firm would have protected access to a good labor force, also significantly 
 favored the size of program impact. This, together with the observed 
 low but significant r between the two variables (. 20) provides suspicion 
 that the degree of the impact of the size of loan will depend upon factors 
 characteristic of the locality. It is tempting simply to make use of the 
 regression coefficient for the loan size variable as shown in Chapter 4, 
 since the multiple regression, using an exponential form, corrects for 
 confounding influences (so long as multi-collinear -ty is not present). 
 But the coefficient of this relation implies that impacts are increasing at 
 an increasing, rather than decreasing, rate as size of loan increases. 
 It is therefore necessary to focus upon the size of the coefficient and the 
 functional relation between size of loan, specifically, and income impacts. 
 
 To test the hypotheses indicated earlier in this chapter, it was 
 decided to make use of ordinary least squares regression upon alternative 
 logarithmic transformations of the variables. Two such transformations 
 were considered: 
 
 (3) Log Y = a + b Log L + C j 
 
 and (4) Y = a_ + b Log L + C 
 
 z 2 2 
 
 7.6 
 
Equation (3) can be restated as: 
 
 a, + C bi bi 
 
 (3«) Y= e 1+ 1 L^AL 1 
 
 So long as 0< b< 1, and is significant, and A > 0, the hypothesis 
 that income impacts increase at a decreasing rate as loan size increases 
 is confirmed. In equation 4, the logarithmic transformation serves 
 as part of the function itself, and follows a path that increases at a 
 decreasing rate. 
 
 Both functions were estimated first via step-wise multiple 
 regression, for the purpose of identifying the influence of local socio- 
 economic factors upon the coefficient b . It was found that b varied 
 
 1 1 
 
 considerably as new variables were entered into the analysis, 
 although well within the defined range (. 29<b£.60). The highest t- 
 ratio for b in the step-wise regression was obtained with 15 variables 
 included, with b 1 having a value of . 56 and a t- ratio of 3.5. Since the 
 instability occurred upon entry of the variable "median age in county" 
 it was decided to estimate a simple regression between Log Y and 
 Log L, with all observations deleted having a median age greater than 
 the average for the entire sample leaving only 15 observations. This 
 
 gives a b value of . 597, a t- ratio of 2. 0, and an r of .23. Because 
 
 1 
 
 of the scantiness of observations in this test, it was decided to 
 estimate the coefficients to equations 3 1 and 4 on unstratified data, 
 providing 33 observations for the analysis. The results were as follows: 
 
 7.7 
 
(l-bi)6 
 (3') Y = I AL M10 u 
 
 ■[-'■Tv... 
 
 (2.99) 
 
 r - 6 1 ( 
 
 = a + b log(L10 )\ 10 
 
 (4) Y = a + b log(L10 
 
 = I . 337 + .100 log(L10 )J 
 
 < 2 - 97 ' R E = .22. 
 
 The results for equation 3' provide estimates of b substantially- 
 similar to those obtained in the smaller sample, stratified tests, and 
 to those obtained in the step-wise multiple regression. Therefore, it 
 was decided to use the coefficients estimated for equations 3' and 4 as 
 shown above. * 
 
 The next step in the analysis was to estimate the linear cost 
 coefficient (u). In Chapter 5 of this study, two program cost estimates, 
 a reasonable and a conservative estimate, were provided based upon the 
 analysis made in EDA's Miller, Gaskins, Liner Study. * The reasonable 
 estimate of program cost for loans amounting to $23,180, 000 was 
 $9,097,169, while the conservative estimate amounted to $13,129,271. 
 Dividing each program cost by the total loans disbursed gives the 
 
 * Here, the size of the ratio of the control variable (size of loan) is the 
 statistic of interest, and t = 3 is acceptable. The size of R^ is not a criterion 
 for acceptance of the model. The R2 of . 22 merely indicates that "size of 
 loan", as single variable, accounts for 22 percent of the variation in income 
 (Y). Many other variables, including the growth rate, wage, and quality of 
 labor supply as the firm's location, account for much of the remaining 
 variance of Y, as discussed in Chapter V. This is to say that while the 
 model cannot be used to forecast Y for an applicant firm, it can be used to 
 forecast, with confidence, the effect of "loan size" on Y 
 
 7.8 
 
the following cost ratios as estimates of : 
 
 Reasonable Cost Estimate: a = . 3925 
 
 Conservative Cost Estimate: a = . 5664 
 
 The estimated parameters for deriving the optimal loan size are 
 now available. Since two sets of coefficients have been provided for the 
 benefit function and two for the cost function, four estimates of optimal 
 loan size will be derived. 
 
 Inserting the coefficients of equations 3' and 4, and taking the 
 integral from t = 6* to t = oo t equation 2 may be stated as follows: 
 
 (2.1) Hj = .£l.859[.2411/ 541 10 2 ' 75 ]e (2 - r)t dt - a L 
 
 r 00 T -6 T 6 (2-r) fc 
 
 (2.2) n = J 6 1.859 1.337 + .100 In (L 10 ) 1 10 e V ' dt - aL 
 
 Taking the derivatives of equations (2. 1) and setting them equal to 
 
 zero, the optimal loan size may be derived by solving for L: 
 
 r°° - 459 2 75 (2-r)t 
 
 (2.1') a = X 1.859(. 541)(.241) L 10 e v ; dt 
 
 (2. 2 1 ) a = S, .1859 L 10 e* ; dt 
 o 
 
 The discount rate, r, will be set at the conservative value of 10 
 
 percent when the conservative cost estimate for a is to be applied, and 
 
 * Rather than using a lower bound of t = 0, the procedure will 
 be followed that was applied in the MGL Study of ignoring the first six 
 years of program benefits due to lack of data. 
 
 7.9 
 
at the value of 6 percent when "the reasonable" cost estimate for 
 is to be applied. 
 
 The estimated loan sizes L* that maximize net benefits 
 according to the four alternative sets of benefit and cost coefficients 
 are shown in Table 2. 
 
 Table 2: Estimated Optimal Loan Size 
 
 .Cost Function 
 Benefit Function"" 
 
 Reasonable (. 393) 
 
 Conservative^ 566) 
 
 Exponential (n ,) 
 Logarithmic ( n ) 
 
 $194,000,000 
 $ 8,637,500 
 
 $13,618,000 
 $ 2,548,400 
 
 The average loan size during the 1962-1967 period surveyed in this 
 study amounted to $579, 500. However, recent policy has been to approve 
 loans of a much higher amount - - $1 million or more per loan. Table 2 
 indicates that the recent policy is a step in the right direction. 
 Even admitting of the crude nature of this analysis, the most conservative 
 estimated optimal loan size is $2.5 million. 
 
 7.10 
 
APPENDIX A 
 CALCULATION OF INCIDENCE MULTIPLIER 
 
 The income directly received by income group i_ in time t is 
 
 equal to receipts from exogeneous sources (E) plus receipts from the 
 
 fch 
 local group J_ (assumed to be a constant proportion (p. .) of the j group's 
 
 income (Y) in the preceeding period (t-1)). This can be stated as: 
 
 (1) Y.(t) = Zp Y (t-l) + E.(t) 
 j 
 
 i - l,...,n 
 
 which is equivalent to: 
 
 (2) Y (t+1) = Zp Y.(t) + E(t+1) 
 J 
 
 or, in matrix terms: 
 
 (2. 1) Y(t+1) = . PY(t) + E(t+1) 
 
 We assume that E(t+1) = E(t) = E, a constant. The purpose of 
 
 this assumption is to represent the case in which the local economy 
 
 experiences a permanent increase in the export sector of a constant 
 
 amount such as might result from effective public development activities, 
 
 In the second time period, equation 2. 1 gives: 
 
 (2.2) Y(t+2) = PY(t+l) + E 
 
 = P [PY(t) + E] + E 
 
 = P Y(t) +PE+E 
 
 = P 2 Y(t) + [P+ l]E 
 
 A-l 
 
Similarly, 
 
 (2.3) Y(t+3) = PY(t+2) + E 
 
 = P [P Y(t) + (P+ I) E]+ E 
 
 = P 3 Y(t) + [P 2 + P]E + E 
 
 = P 3 Y(t) + [P 2 + P+ l]E 
 
 By induction, it is seen that: 
 
 (2. 4) Y(t+n) = P n Y(t) + [P n + P n_1 + . . . + p + i]E 
 
 As n— °° , equation 2. 4 reduces to: 
 
 (2.5) Y = P°°Y(t)+[Z p' + ljE 
 
 t=l 
 
 Further, the infinite series in the bracketed term is solved as: 
 
 00 t r t r "~ n-1 * 
 
 Z P + I = [i - P°°] [i - P] 
 
 t=l 
 Equation 2. 5 can be restated as: 
 
 (2.6) Y = P°°Y(t) + [I - P°°][l - P]" 1 E 
 
 which solves for the income vector Y after the multiplier process has 
 converged. Attention now focusses upon the value of P 00 . If all 
 columns of P contain elements p.. such that 0< p. . < 1 and Zp.. < 1, 
 then P reduces to zero. It will be shown below that these requirements 
 on the elements of P will be satisfied. This allows equation (2. 6) 
 to be reduced to: 
 
 (2.7) Y = [i - P]" 1 E 
 
 * This can be derived by "multiplying the term [i + Z P J by [i - P] 
 [I-P][I+P+P 2 +. . . P n ] = I-P n+1 t=1 
 
 Multiplying both sides by [i - p] and letting n =co , 
 
 1+ IP 1 = [I - P°0[l - p]" 1 . 
 
 t=1 A-2 
 
We now consider the derivation of the matrix P. 
 
 Let the expenditures of income group i for commodity k be C., , 
 i = 1, . . . ,n and k = 1, . . . t s. 
 
 If these expenditures are a constant proportion of income then: 
 
 ik ' Y. ' \ i= 1. .... n 
 
 i J k = 1 , . . . , s 
 
 Further, suppose that the proporting of goods imported from other 
 
 regions ( m ., ) can be identified: 
 
 IK 
 
 m, = £ I k - 1, .... s 
 
 k 
 
 Over a small growth span, these import coefficients will be rela- 
 tively constant, although as the region becomes large, import substitu- 
 tion will take place, reducing the size of the import coefficients. This 
 effect will not be considered here. * 
 
 Subtracting m from one and multiplying by c , the propensity 
 to consume locally is obtained: 
 
 L c ,k = c ik (1 - m k ): \} '- ! n 
 
 |k = 1, .... s 
 
 Next, the percent income distribution for each industry directly 
 
 receiving the expenditures of local consumers (a..) is obtained: 
 Y. k 1J 
 a.. = „ J ; jk = 1, .... s 
 1J Y k j = 1 n 
 
 * Another consideration is that import coefficients will vary by income 
 group. Higher income groups will have higher propensities to import. This 
 suggestes that the import coefficients should be differentiated by income 
 group: m. k = M. k /C lk . 
 
 A- 3 
 
This states that the probability that income group j^ earns income 
 
 from industry k is constant, equal to the earnings of income group j 
 
 employed in industry k divided by total earnings in industry k. However, 
 
 not all of an industry's sales are spent on local wages and salaries -- 
 
 much of it is spent on purchases of interindustry commodities. It is 
 
 assumed that all interindustry purchases are imported -- supplied from 
 
 outside the region. * Hence, to obtain the probability that a dollar of 
 
 receipts in industry k will be earned by employees in income group J_, the 
 
 a coefficients must be adjusted downward to reflect the probability 
 
 of interindustry purchases. To do this, the ratio (V ) of value added 
 
 to total purchases in industry k is obtained and multiplied by the a : 
 
 o . — a . v , , K— i , . . . , s 
 
 kj kj k . 
 
 Next, to obtain the probability ( p..) that a dollar spent by income 
 
 L. lj 
 
 group i for local consumer goods is earned by local employees of income 
 
 group j , the c for a given income group ^_in each consumer sector 
 Xj iK 
 
 are multiplied by the T a, . coefficients and for income group j employed 
 
 L. kj 
 
 in each of the corresponding sectors, and summed over all sectors: 
 
 8 
 
 p.. = Z c a ; i, j = 1, . . . , n 
 
 L ij , , L lkL kj 
 
 * This simplifying assumption will not hold in industrialized and 
 urbanized regions. 
 
 A- 4 
 
or in matrix terms: 
 
 Finally, noting that the . p.. represent the probability that a dollar 
 
 L ij 
 
 spent locally by income group i is locally earned by income group j rather 
 than the reverse, we simply transpose the P matrix to obtain the pro- 
 babilities required to implement equation (2. 7), the probability that income 
 
 group j earns a dollar spent by income group jL. 
 
 T 
 (4) P = L P 
 
 By substituting equations (3) and (4), equation (Z. 8)now reads: 
 (2.8) Y = [i- ( L C L A) T ]" 1 E 
 
 which gives, as required, the total direct and indirect income earned 
 by each of the various income groups of a region resulting from an increase 
 in economic activity in the export sector. With this tool, the incidence of the 
 regional income impacts of public programs may be estimated. The data 
 required for estimating equation (3) is shown in Tables 1 through 6 as 
 follows. The [i - PJ inverse matrix derived from this data and to be 
 
 applied to equation (2. 8) is as follows: 
 
 r, , ^ a>T-|-1 Tl.116 .100 .072 
 
 [I - ( T C _ A) J 
 
 i Ti.iii 
 
 .38 
 |_ .23 
 
 L L ' .385 1.329 .240 
 
 230 .196 1.144 
 
 A- 5 
 
TABLE 1 
 
 Percent Distribution of Consumer Expenditures 
 by Commodity Type and Income Group* 
 
 Commodity"! 
 
 
 
 
 
 
 
 
 
 
 Inc^m>-^^P 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 Tot- 
 al' 1 ' 
 
 Group 
 
 
 
 
 
 
 
 
 
 
 $3, 999 or less 
 
 . 31 
 
 .19 
 
 .09 
 
 .08 
 
 . 03 
 
 . 11 
 
 .03 
 
 .16 
 
 
 $4,000 - $9,999 
 
 .23 
 
 . 14 
 
 .08 
 
 .08 
 
 .05 
 
 .08 
 
 .03 
 
 . 15 
 
 
 $10, 000 - over 
 
 .14 
 
 . 11 
 
 .07 
 
 .08 
 
 .05 
 
 .07 
 
 .03 
 
 . 11 
 
 
 * Source: Survey of Consumer Expenditures, 1960-61, U. S. Dept. of Labor, 
 ** Commodity groups are as follows: 
 
 1. Food, tabocco, alcoholic beverages, 1/4 allocation of gifts and con- 
 tributions. 
 
 2. Shelter, fuel, light, refrigeration, water, 1 /2 allocation of gifts and 
 contributions. 
 
 3. Household furnishings, operations, equipment. 
 
 4. Clothing, clothing materials and services, 1/4 allocation of gifts and 
 contributions. 
 
 5. Personal insurance. 
 
 6. Personal and medical care, education and reading. 
 
 7. Recreation. 
 
 8. Transportation, other expenditures. 
 
 * Totals do not add to 100 percent. The remainder consists of taxes and 
 savings. 
 
 A-6 
 
TABLE 2 
 
 Import Coefficient -- Percent Distribution of Response to Question: 
 "Indicate What Percent of Purchases are Made in the County in which You Live, 
 
 and Calculation of Average Percent* 
 
 ""^-^^^^ Percent 
 
 
 
 
 
 
 Weighted 
 
 Class ^""^"^^^^ 
 
 00 
 
 25 
 
 50 
 
 75 
 
 100 
 
 Average 
 (1-M. k ) 
 
 Food 
 
 
 
 
 
 
 
 Low Income 
 
 .05 
 
 .17 
 
 .12 
 
 .12 
 
 .55 
 
 .74 
 
 Middle Income 
 
 .04 
 
 . 15 
 
 .08 
 
 .12 
 
 .62 
 
 .78 
 
 High Income 
 
 .08 
 
 .08 
 
 .12 
 
 .12 
 
 .60 
 
 .77 
 
 Clothing 
 
 
 
 
 i 
 
 
 
 Low Income 
 
 .07 
 
 .24 
 
 .12 
 
 .19 
 
 .37 
 
 .64 
 
 Middle Income 
 
 .06 
 
 .19 
 
 . 13 
 
 .19 
 
 .43 
 
 .68 
 
 High Income 
 
 .08 
 
 .20 
 
 .16 
 
 .28 
 
 .28 
 
 .62 
 
 Appl. , Furn. , and 
 
 " 
 
 " 
 
 
 
 
 
 Auto 
 
 
 
 
 
 
 
 Low Income 
 
 .15 
 
 .19 
 
 .10 
 
 .14 
 
 .42 
 
 .62 
 
 Middle Income 
 
 .15 
 
 .12 
 
 .10 
 
 .13 
 
 .49 
 
 .67 
 
 High Income 
 
 .20 
 
 .12 
 
 .08 
 
 .20 
 
 .40 
 
 .62 
 
 Other Purchases 
 
 
 
 
 
 
 
 Low Income 
 
 .15 
 
 .20 
 
 .11 
 
 .15 
 
 .38 
 
 .60 
 
 Middle Income 
 
 .09 
 
 .18 
 
 . 14 
 
 .16 
 
 .44 
 
 .67 
 
 High Income 
 
 .20 
 
 .16 
 
 . 12 
 
 .28 
 
 .24 
 
 .55 
 
 Shelter 
 
 
 
 
 
 
 1.00 
 
 * Source: Survey of employees by EDA loan recipient firms by Chilton 
 Research Services, Philadelphia, Pennsylvania, 1969. 
 
 A- 7 
 
TABLE 3 
 
 C Matrix -- Probability that one Dollar Earned by Group (i) 
 is Spent on Commodity (k) and Within County 
 
 ^"^^C ommod it y 
 
 
 
 
 
 
 
 
 
 Income\Group : 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 Group """"""-^^ 
 
 
 
 
 
 
 
 
 
 Low Income 
 
 .229 
 
 . 190 
 
 .056 
 
 .051 
 
 .018 
 
 .066 
 
 .018 
 
 .099 
 
 Middle Income 
 
 . 179 
 
 . 140 
 
 . 054 
 
 .054 
 
 .034 
 
 . 054 
 
 .020 
 
 . 101 
 
 High Income 
 
 . 108 
 
 . 110 
 
 . 043 
 
 .050 
 
 .028 
 
 .039 
 
 .017 
 
 .068 
 
 * See Table 1 for detailed explanation of commodity groups 
 
 A- 8 
 
TABLE 4 
 
 Percent Income Distribution by Industry and 
 Value Added as a Percent of Purchases* 
 
 ^»»^^ Commodity 
 
 
 
 
 
 
 
 
 
 ^^■^ Group* 
 
 Group ^>>^ 
 
 * 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 Low Income 
 
 .16 
 
 . 10 
 
 .16 
 
 .16 
 
 .07 
 
 .16 
 
 . 12 
 
 .48 
 
 Middle Income 
 
 .61 
 
 .44 
 
 .61 
 
 .61 
 
 .45 
 
 .43 
 
 .58 
 
 . 34 
 
 High Income 
 
 .23 
 
 .46 
 
 .23 
 
 .23 
 
 .48 
 
 .41 
 
 .30 
 
 . 18 
 
 
 
 
 
 
 
 
 
 
 ■t r - i . . - .1 .1 ' 
 
 V CklUG CtUUfcJU 
 
 to purchases 
 
 .724 
 
 .666 
 
 .724 
 
 .724 
 
 . 560 
 
 .681 
 
 .584 
 
 .247 
 
 ratio (VU*** 
 
 
 
 
 
 
 
 
 
 * Source: Table 230, Census of Population, I960 , U.S. Bureau of Census. 
 ** Commodity groups are based on two-digit SIC sectors, and correspond 
 as nearly as possible to the commodity classes of Tables 1, 2, and 3: 
 
 1. Retail and wholesale trade. 
 
 2. Real estate and rentals; electric, gas, water and sanitary services. 
 
 3. Retail and wholesale trade. 
 
 4. Retail and wholesale trade. 
 
 5. Finance and insurance. 
 
 6. Professional and related services. 
 
 7. Entertainment and recreation; hotels, personal and repair services (exc. 
 auto). 
 
 8. Transportation; business and repair services. 
 
 *** Source: Table of Interindustry Transactions, 1958, Survey of Current 
 Business, September, 1965, page 38. 
 
 A- 9 
 
TABLE 5 
 
 A Matrix -- Probability that one Dollar of Receipts 
 
 in Commodity Sector (k) is Paid as Earnings to 
 
 Income Group (j) and Within County 
 
 "*"* Income 
 
 
 Middle 
 
 
 Commo^ --^^Group 
 
 Low income 
 
 Income 
 
 High Income 
 
 dity Group *"" -*^^ 
 
 
 
 
 1. Food 
 
 . 116 
 
 .442 
 
 .167 
 
 2. Shelter 
 
 .067 
 
 .293 
 
 .306 
 
 5. Furnishings 
 
 .116 
 
 .442 
 
 i 
 
 .167 
 
 4. Clothing 
 
 .116 
 
 .442 
 
 .167 
 
 5. Personal Insurance 
 
 . 037 
 
 .252 
 
 .269 
 
 6. Personal Services 
 
 .109 
 
 .293 
 
 .279 
 
 7. Recreation 
 
 .07© 
 
 .339 
 
 .175 
 
 8. Transportation 
 
 .119 
 
 .084 
 
 .045 
 
 A-10 
 
TABLE 6 
 
 P Matrix -- Probability that a Dollar Spent by Local Income Group (i) 
 is Earned by Local Income Group (j) 
 
 ""^^-^ TO 
 
 FROM ^^^ 
 
 Low Income 
 
 Middle 
 Income 
 
 High Income 
 
 Low Income 
 
 .0726 
 
 .2423 
 
 .1451 
 
 Middle Income 
 
 .0631 
 
 .2673 
 
 .1227 
 
 High Income 
 
 .0452 
 
 . 1506 
 
 .0913 
 
 A-ll 
 
APPENDIX B 
 
 CALCULATION OF INDIRECT IMPACTS UPON 
 EMPLOYMENT AND UNEMPLOYMENT 
 
 The change in income in a county is taken to depend upon exogenously 
 determined growth in local primary (AP) and secondary (AS) industry, 
 which stimulates growth in " tertiary" industry (AT). This is essentially 
 the economic base theory of regional growth, and may be expressed as 
 follows: 
 
 (1) AY = AT + AS + AP 
 and 
 
 (2) AT = a + b. AS + b AP 
 
 Upon substitution, equation (1) becomes:* 
 
 (3) AY = a + (b + 1) AS + (b + 1) AP 
 
 Using OBE** estimates of income by 1 -digit SIC for 90 counties 
 
 in which EDA business loans were granted, equation (3) was estimated 
 
 as: 
 
 (3 1 ) AY = 5284. + 1.8589 AS + 1.2782 AP - R 2 = . 79 
 
 (13.28) (6.39) 
 
 * Weiss, Steven J. , and Edwin C. Gooding, "Estimation of Differ- 
 ential Employment Multiplies in a Small Regional Economy, " Land Economics , 
 Vol. XLIV, No. 2, May, 1968. 
 
 ** A description of this data is to be found in Robert Graham and Edwin 
 Coleman, "Metropolitan Area Incomes, 1929-66, " Survey of Current Business , 
 August, 1968. 
 
 B-l 
 
The predictive power of this estimated equation is good, and the 
 t-value of the regression coefficient for secondary industry is extremely 
 high. 
 
 Since the EDA business loan porgram would cause a change only in 
 secondary industry, the multiplier to be applied to the observed direct 
 income impacts of EDA loans (AS) is 1. 859. 
 
 The resulting estimate for AY was then converted into an employment 
 dimension via the following procedure: 
 
 1. Secondary income change was subtracted from estimated total 
 income (AY - AS) to give total tertiary industry change (AT). Using a 
 ratio of wholesale, retail, and service income to total tertiary income of 
 
 .4808 (obtained from OBE estimates for the 90 counties), estimated tertiary 
 income was split into "wholesale, retail, and service" income, and "other 
 tertiary" income. 
 
 2. The ratios of U. S. employment to payrolls for the "wholesale, 
 retail and service" sector for the "other tertiary" sector, and for the 
 secondary (manufacturing) sector were obtained from 1964 County Business 
 Patterns , U. S. Department of Commerce. These ratios were corrected 
 for price changes (1. 17) during the period 1964-1969 using annual ratios 
 
 of real-to-money-GNP obtained from the Economic Report of the President , 
 January, 1969. 
 
 The resulting ratios were further corrected to reflect rural technology 
 by obtaining the actual employment-income coefficient observed in the 41 
 loan recipient firms using a simple linear regression. This coefficient. 
 
 B-2 
 
2 
 
 (.0002271; R = . 98) was divided by the national secondary sector ratio 
 
 to obtain a rural productivity correction factor (1. 565) which was multi- 
 plied by the two tertiary sector ratios to give: 
 
 E 
 
 — ; wholesale, retail, and service: .0002116 
 
 — ; other tertiary: .0001519 
 
 3. The two estimated tertiary sector income magnitudes were then 
 multiplied by the respective employment/income ratio and summed together 
 with actual secondary sector employment (employment in surveyed EDA 
 financed firms since date of loan) to obtain total estimated employment 
 change, {AE}. 
 
 (4) AE = [(.0002116) (.481) + (.0001519) (1 - .481)] 1.859 AS 
 
 = .00033577 AS 
 
 4. Based upon a recent study of the impact of employment change 
 upon county unemployment (CONSAD Research Corporation, for EDA, 
 1969), the following relation was applied to estimate the impace of employ- 
 ment change upon county unemployment change (AU): 
 
 (5) AU = (.887 - 1) AE 
 
 or, substituting from equation (4): 
 
 (5 1 ) AU = (.887-1) [(.0002116) (.481)+ (.0001519) (1- .481)]l. 859 AS 
 
 B-3 
 
Applying equations (4) and (5) to the estimates of total annual 
 earnings of employees from 41 EDA-financed firms and of change in 
 annual earnings of those employed in EDA-financed firms, two sets 
 of unemployment change estimates based upon the multiplier process 
 were obtained: 
 EDA Total Earnings Change in Earnings 
 
 AS = $7,747,643 AS = $2, 530, 795 
 
 AE= 2,601 AE= 850 
 
 AU = 294 AU= 97 
 
 B-4 
 
APPENDIX C 
 COMMUNITY LEADERSHIP SURVEY 
 
 Exhibit A is a copy of the Community Leader's Opinion Survey 
 given to 100 community leaders of 40 counties where EDA loan recipient 
 firms were located. The interview questioned leaders on job opportuni- 
 ties in the community, community services and facilities, and growth 
 trends. Exhibit B is a list of the distribution of leaders by position. 
 
 Section A reports the responses of the community leaders and Section 
 B includes an analysis by geographic area and by population growth rate. 
 
 A. Opinions of Community Leaders 
 
 The community leaders were asked if they felt that more job oppor- 
 tunities were available in the county now than four years ago. The majority 
 (88) felt that there are now more job opportunities. Eight felt that job 
 opportunities were the same and four felt that there were fewer job oppor- 
 tunities. Table 1 summaries their estimates on how many companies 
 had moved into the area or had expanded facilities in this time period. 
 Most of the companies which had moved into the county or expanded their 
 facilities within the county were manufacturing or construction companies. 
 
 Because of increased job opportunities most of the leaders (79%) 
 thought that the people of the county were better off, mainly because of 
 higher wages and more available jobs. Other reasons mentioned included 
 
 C-l 
 
expanding industry, better housing and educational facilities and improved 
 community services. 
 
 Table 2 summarizes the opinions of the leaders about who has done 
 anything to improve the number of job opportunities. 
 
 Private industry, the state government and the federal government 
 were all felt to be active in improving job opportunities. Community leaders 
 cited the expansion of facilities creation of more jobs and job training 
 programs, higher salaries and hiring the hard core unemployed among the 
 activities of private industry. 
 
 The state government was credited with helping to bring in new in- 
 dustries and advertising for new business, in addition to sponsoring job 
 training programs and building new highways. Activities of the Federal 
 government which were mentioned included helping to finance local busi- 
 ness, providing EDA loans, grants for better sewage and water systems, 
 and poverty programs (e.g., CAP, OEO, Job Corps, Head Start). 
 
 The majority of local leaders (84%) felt that the people of the county 
 were generally better off because of new and expanding industry, more job 
 opportunities, and a positive business outlook. They mentioned these same 
 reasons for expecting a continued increase in job opportunities and standard 
 of living. 
 
 One of the questions in the survey asked whether the leaders felt that 
 more families have moved into or out of the county in the last few years. 
 Over 80% responded that there had been a shift of population into the county. 
 
 C-2 
 
Reasons cited for movement into the county were better job opportunities, 
 recreational facilities and "to get away from the city, " Those community 
 leaders in counties with a larger out migration felt that it was due to the 
 lack of jobs and a decline in agriculture. Only 15% of the leaders felt that 
 the size of the population in their county was not changing. 
 
 Several of the survey questions dealt with community services such 
 as schools and police, fire and first aid squads (Table 3). 
 
 Over 70% felt that the school population of their county was growing; 
 of those with an increasing school population, 34 reported that school 
 facilities were adequate while 36 thought the facilities were cramped and 
 overcrowded. 63% said that school facilities were being expanded by build- 
 ing new schools or adding onto current facilities. Other types of improve- 
 ments mentioned included expanding the curriculum, consolidating school 
 programs, expanding recreational facilities and building vocational 
 schools and community colleges. 
 
 About two-thirds of the community leaders felt that civil services (police, 
 fire, etc) for the county were expanding and most of them felt that this trend 
 would continue because of increasing population and planned new construction 
 of facilities. Those who thought that services would .not continue to grow 
 or would remain the same, felt that present services were adequate for the 
 county. 
 
 Table 4 presents the community leaders 1 estimates of new retail 
 business in the county. 
 
 C-3 
 
Less than half reported that new supermarkets had moved into the 
 county and only 24% said that the county now had additional large depart- 
 ment stores. 
 
 More than 70% reported increased new highway construction and 63% 
 reported improved highway maintenance in their county. 
 
 Two-thirds of the community leaders said that cultural, social and 
 recreational facilities in the community had increased and mentioned addi- 
 tional parks, swimming and boating facilities, libraries and community re- 
 creation centers. 
 
 B. Analysis of Geographic Area and Population Change 
 
 Responses of community leaders to the questionnaire were analyzed 
 by the geographic area and population change of the county. 
 
 For the first analysis four geographic areas were used to sort the 
 
 survey responses for 99 community leaders. (See Table 5) 
 
 2 
 
 The X (Chi-square) test was used to determine responses signifi- 
 cantly influenced by geographic location. Significant results were obtained 
 for the following survey questions: 
 
 1. Generally, as a result of the increase in job opportunities, 
 
 do you think people in this area are better off, about the same or not as 
 
 well off as they were previously? Most of the leaders in each area thought 
 
 that people were better off. In the South, 81% of the community leaders 
 
 responded this way; 80% of the leaders in the Northeast, 75% in the West 
 
 and 64% in the North Central States. 
 
 C-4 
 
2. In your opinion has the state done anything within the past 
 four years to improve job opportunities within this county? About 
 
 73% of community leaders in the Northeast felt that the states had had a role 
 in increasing job opportunities; 56% of the Southern leaders, 43% in the 
 North Central region and 37% in the West. 
 
 3. Were the educational facilities adequate to handle the increased 
 number of children attending school or did the school become cramped and 
 overcrowded? (This question was asked only of community leaders who 
 thought their county's school population was increasing). 50% of community 
 leaders from the Northeast thought that their schools were adequate to handle 
 increasing population, 43% thought that the schools were cramped and over- 
 crowded. In the West 37% thought that schools could adequately handle the 
 increase; 62% did not. In the South 39% did not know if schools could handle 
 increasing population; 29% thought that they could and 29% thought that they 
 were cramped and overcrowded. Of community leaders in the Northeast, 
 47% felt that schools were cramped and overcrowded, 33% thought they 
 were adequate and 20% did not know. 
 
 In the analysis by population change, community leaders responses 
 were sorted into quartiles according to the rate of growth or decline of 
 
 population in their county (See Table 6). 
 
 2 
 
 Significant X results were obtained for responses to the following 
 
 questions: 
 
 C-5 
 
1. Do you feel the people in this county are better off, about the 
 same or not as well off as they were say four years ago? More of the 
 community leaders with a rapidly declining population felt that people 
 were now better off (96%) compared with 87% of those with a slowly 
 declining population and 75% of those with a rapidly growing population 
 in their county. 
 
 2. Were the increased educational facilities adequate to handle the 
 increased number of children attending school or did the schools become 
 cramped and overcrowded? About 48% of community leaders in counties 
 with moderate population growth thought schools were adequate while only 
 25% of those with rapidly increasing population thought schools were not 
 overcrowded. 
 
 3. Are schools expanding any of their present facilities ? 80% 
 
 of the community leaders of counties with net in-migration said schools 
 were expanding facilities while only 45% of leaders in counties with net 
 out-migration reported expansion. 
 
 4. Have civil services been increased within the past few years? 
 In counties with rapidly declining population only 35% reported increases 
 in police, fire and other civil services. In counties with growing popula- 
 tions, 74% (of those with moderate growth) and 79% (of those with rapid 
 growth) thought these services were increasing. 
 
 5. Would you say you have more, about the same, or less adequately 
 maintained roads in the county than you had four years ago? Over 70% of 
 
 C-6 
 
those leaders in counties with population gains reported better road main- 
 tainence while only 48% of those with losses in population reported im- 
 proved road maintainence. 
 
 C-7 
 
Table 1 
 
 County Growth of Job Opportunities 
 
 NoT^-^A ctivitie s 
 
 Of ^^\^ 
 
 Companies *^^^^ 
 
 Companies which 
 Moved into Area 
 
 Companies which 
 Expanded facilities 
 
 
 
 5 
 
 5 
 
 1-4 
 
 37 
 
 53 
 
 5-14 
 
 23 
 
 14 
 
 15 or more 
 
 9 
 
 8 
 
 do not know 
 
 26 
 
 21 
 
 C-8 
 
Tabic 2 
 
 Job Opportunities 
 
 
 Imp rove me nt 
 
 in Job 
 Opportunities 
 
 Has -- Done Anything to Improve 
 
 Number of Job Opportunities 
 
 Private 
 Industry 
 
 State 
 Government 
 
 Federal 
 Government 
 
 Yes 
 
 88 
 
 83 
 
 1 
 
 56 
 
 71 
 
 ■ No 
 
 12 
 
 8 
 
 31 
 
 22 
 
 Do not know 
 
 — 
 
 9 
 
 13 
 
 7 
 
 C-9 
 
Table 3 
 
 Community Services -- Expansion 
 
 
 Schools 
 
 1 " 
 Civil Services 
 
 Population 
 
 Facilities 
 
 Are Expanding 
 
 72 
 
 63 
 
 64 
 
 Are Not Expanding 
 
 25 
 
 36 
 
 32 
 
 Do Not Know 
 
 3 
 
 1 
 
 4 
 
 C-10 
 
Table 4 
 New Stores 
 
 
 Large Retail 
 Food Stores 
 
 Large Dept. 
 Stores 
 
 Other Retail 
 Stores 
 
 Yes 
 
 47 
 
 24 
 
 33 
 
 No 
 
 49 
 
 74 
 
 62 
 
 Do Not Know 
 
 4 
 
 2 
 
 5 
 
 Oil 
 
Table 5 
 
 Number of Community Leaders by Regions 
 
 
 Number 
 
 1 
 
 % of 
 
 Region 
 
 Surveyed 
 
 Total 
 
 1. Northeast 
 
 15 
 
 15 
 
 2. South 
 
 62 
 
 63 
 
 3. North Central 
 
 14 
 
 14 
 
 4. West 
 
 8 
 
 8 
 
 Total 
 
 99 
 
 100 
 
 C-12 
 
EXHIBIT A 
 QUESTIONNAIRE 
 
Table 6 
 
 Population Change 
 
 Quartile 
 
 Number of Respondents 
 
 
 1 
 2 
 3 
 
 4 
 
 26 
 23 
 27 
 24 
 
 rapid decline 
 moderate decline 
 moderate growth 
 rapid growth 
 
 C-13 
 
Chilton Research Services 
 Philadelphia, Pennsylvania 
 April,. 1969 
 
 Study #9138 
 (1-4) 
 
 Int. if 
 
 (5-7) 
 
 Co. // 
 
 (8-9) 
 10 - rej. 
 
 Good afternoon, I'm calling long distance from Philadelphia for 
 
 Chilton Research Services. We are doing a study of community leaders like yourself to 
 
 get their opinions about past and future community growth in county. We 
 
 would appreciate it very much if you would give us your (name) 
 thoughts on a few subjects. 
 
 Time Int. Started A.M. P.M. 
 
 A.M. 
 
 Time Int. Ended 
 
 P.M. 
 
 1. Do you feel there are more, about the 
 same or fewer job opportunities in 
 
 county today than there were 
 
 say four years ago? 
 
 SKIP TO Q. 9 
 
 SKIP TO Q. 8 
 
 Q. 9 
 
 More 
 
 About the same 
 
 Fewer 
 
 11-1 
 
 2. As far as you know, how many companies, 
 if any, have moved into this county in 
 the past four years or so? 
 
 IF NONE OR DK SKIP TO Q. 4 
 
 Don't Know 
 
 3. What do these companies do or manufac- 
 ture? 
 
 a. (IF MORE) Why do you say that? 
 
 More companies have started 
 operations in the past few 
 years 
 
 #1 
 
 n 
 
 12-1 
 
 Companies have expanded 
 their facilities. 
 
 Other (PLEASE SPECIFY) 
 
 IF NONE OR DK SKIP TO Q. 6 
 
 5. What do these companies do or manufac- 
 ture? 
 
 01 
 
 n 
 
 #3 
 
 t=rr 
 
 4. As far as you know, how many companies 
 in this county, if any, have expanded 
 their facilities during this period of 
 time? 
 
6 
 
 Generally, as a result of the increase 
 in job opportunities, do you think 
 people in this area are better off, 
 about the same, or not as well off 
 today as they were previously? 
 
 9a. 
 
 (IF "YES") 
 
 What did they do? 
 
 
 
 Better off 
 
 13-1 
 
 • 
 
 About the same 
 
 2 
 
 
 
 Not as well off 
 
 3 
 
 
 
 SKIP TO Q. 9 
 
 Don't know 
 
 V 
 
 10. 
 
 In your opinion, has the state govern- 
 
 7. Why do you say that? 
 
 ment done anything 
 years to improve jo 
 within this county? 
 
 within the past four 
 b opportunities 
 
 
 
 Yes 
 
 15-] 
 
 
 SKIP 
 
 TO 
 Q. 11 
 
 No 
 
 2 
 
 Don ' t Know 
 
 V 
 
 ! 
 
 10a. 
 
 (IF "YES") What did they do? 
 
 8. (IF FEWER JOBS ASK) Why do you feel 
 that way? 
 
 11. 
 
 In your opinion, has the Federal govei 
 ment done anything within the past four 
 years to improve job opportunities 
 within this county? 
 
 
 
 Yes 
 
 16-1 
 
 
 SKIP ' 
 
 TO 
 Q. 12 
 
 No 
 
 2 
 
 ASK EVERYONE 
 
 Don't Know 
 
 V 
 
 9. In your opinion, has private industry 
 done anything within the past four 
 years to improve job opportunities 
 within this county? 
 
 11a. 
 
 (IF "YES") 
 
 What did they do? 
 
 
 ' 
 
 Yes 
 
 14-1 
 
 1 
 
 SKIP 
 
 No 
 
 2 
 
 
 
 TC 
 
 Q. 
 
 t 
 10 
 
 Don' t Know 
 
 V 
 
 
 
 
 
 C-15 
 
12. Do you feel the people in this county 
 are better off, about the same or not 
 as well off today as they were say 
 four years ago? 
 
 14a 
 
 Why do you fee 
 
 1 this way? 
 
 
 
 Better off 
 
 17-1 
 
 
 About the same 
 
 2 
 
 Not as well off 
 
 3 
 
 Don ' t Know 
 
 V 
 
 13. Within the forseeable future, do you 
 think the number of job opportunities 
 in this county will increase, remain 
 about the same, or decline? 
 
 
 
 15. 
 
 Within the past few years, have mo 
 families moved into the county or 
 more families moved out of this co 
 
 re 
 
 have 
 
 unty? 
 
 
 Increase 
 
 18-1 
 
 Remain the same 
 
 2 
 
 Decline 
 
 3 
 
 
 
 
 
 
 More moved in 
 
 20-1 
 
 
 SKIP TO Q. 14 
 
 Don' t Know 
 
 V 
 
 More moved out 
 
 2 
 
 13a. Why do you say that? 
 
 SKIP TO Q. 16 
 
 Don ' t Know 
 
 s for 
 
 - 
 
 15a 
 
 . What would you say are the reason 
 this shift in population? 
 
 14. Do you feel t 
 this county w 
 the same or c 
 few years? 
 
 he standard of living for 
 111 increase, remain about 
 ecrease within the next 
 
 
 Increase 
 
 19-1 
 
 Remain the same 
 
 2 
 
 
 
 Decrease 
 
 3 
 
 
 SKIP TO Q. 15 
 
 Don' t Know 
 
 V 
 
 
 
 
 
 C-16 
 
 
16. 
 
 Now, we would like to ask you a few 
 questions about i_he schools in this 
 county. Do you feel that the number of 
 children attending grammar school and 
 high school has increased, remained 
 about the same or decreased within the 
 last few years? 
 
 SKIP TO 
 Q. 17 
 
 Increased 
 
 Remained the same 
 
 Decreased 
 
 Don' t Know 
 
 21-1 
 
 16a. 
 
 (IF INCREASED) ASK: Were the 
 educational facilities adequate 
 to handle the increased number 
 of children attending school or 
 did the schools become cramped 
 and overcrowded? 
 
 Adequate 
 
 Cramped and overcrowded 
 
 Don't know 
 
 22-1 
 
 1/. is the community building any new 
 schools? 
 
 Yes 
 
 No 
 
 Don ' t know 
 
 23-1 
 
 18. Is the community expanding any of 
 their present school facilities? 
 
 SKIP TO Q. 19 
 
 Yes 
 
 No 
 
 Don ' t know 
 
 24-1 
 
 18a. (IF YES) What are they doing? 
 
 19, 
 
 What about services such as police, 
 fire, medical and first aid squads. 
 Have these services increased, remained 
 about the same or decreased within the 
 past four years? 
 
 SKIP TO Q. 20 
 
 Increased 
 
 Remained the same 
 
 Decreased 
 
 Don ' t know 
 
 25-1 
 
 19a. Do you feel this trend will continue? 
 
 SKIP TO Q. 20 
 
 Yes 
 
 No 
 
 Don't know 
 
 26-1 
 
 19b. Why do you say that? 
 
 20. Do you know if any large retail food 
 supermarkets moved into this county 
 within the past four years? 
 
 Yes 
 
 27-1 
 
 No 
 
 2 
 
 Don't know 
 
 V 
 
 C-17 
 
Do you know of any large department 
 stores that have moved into this 
 county within the past four years? 
 
 Yes 
 
 28-1 
 
 No 
 
 2 
 
 Don ' t know 
 
 V 
 
 22. Have any other large retail stores 
 moved into the county in the past 
 A years? 
 
 Yes 
 
 No 
 
 Don ' t know 
 
 29-1 
 
 23. Do you think there has been any increase 
 in highway construction, outside of 
 ordinary road maintenance, over the 
 past four years? 
 
 Yes 
 
 No 
 
 Don ' t know 
 
 30-1 
 
 24a. Would you say you have more, about 
 
 the same or less adequately maintained 
 roads in this county than you had, 
 say four years ago. 
 
 SKIP TO Q. 25 
 
 SKIP TO Q. 25 
 
 More 
 
 About the same 
 
 Less 
 
 Don ' t know 
 
 31-1 
 
 24b. (IF MORE OR LESS ) Why do you feel that 
 way? 
 
 25. 
 
 In t 
 the 
 and 
 were 
 
 his county are there more, about 
 same or less recreational, social 
 cultural facilities now than there 
 , say four years ago? 
 
 
 , 
 
 More 
 
 32-1 
 
 
 
 TERMINATE 
 
 Same 
 
 2 
 
 
 
 
 Less 
 
 3 
 
 
 
 TERMINATE 
 
 Don' t know 
 
 V 
 
 25a. 
 
 (IF MORE OR LESS) Please 
 me any illustrations. 
 
 give 
 
 THANK YOU VERY MUCH 
 IT WAS A PLEASURE TALKING TO YOU 
 AND GETTING YOUR OPINIONS. 
 
 Respondent 
 
 Title 
 
 County 
 
 Co. # 
 
 State 
 
 Interviewer 
 
 C-18 
 
EXHIBIT B 
 
DISTRIBUTION OF INTERVIEWS 
 
 FOR 
 
 OPINION LEADERS PHASE 
 
 CATEGORY TOTAL 
 
 // 
 School (Education) 
 
 Superintendent 6 
 
 Principal 13 
 
 President — Board of Education 1 
 
 Teacher 1 
 
 Bank (Executives) 
 
 President 11 
 
 Vice President 3 
 
 Assistant Vice President 1 
 
 Manager 2 
 
 Banker 1 
 
 President — Savings and Loan Assoc. 1 
 
 Chamber of Commerce 
 
 President 7 
 
 Vice President 1 
 
 Secretary 1 
 
 Business (Executives) 
 
 President 3 
 
 Vice President. General Manager 1 
 
 Owner, Manager 1 
 
 Owner — Department Store 1 
 
 Owner — Newspaper 1 
 
 Retail Merchant 1 
 
 City/State Government 
 
 State Senator 1 
 
 State Legislator 2 
 
 Councilman, Assemblyman 2 
 
 Mayor 10 
 
 Town Clerk 1 
 
 Law Enforcement 
 
 Chief of Police 8 
 
 Sheriff 1 
 
 Local Organizations 
 
 President — Lions Club 1 
 
 President — Merchants' Association 1 
 
 Director — Board of Freeholders 1 
 Retired President of Home Owners' Association 
 
 & Presently — Aid to Library Conan. 1 
 
 Judicial 
 
 Circuit Judge 1 
 
 President of Police Jury 1 
 
 Attorney 5 
 
 C-1Q 
 
DISTRIBUTION OF INTERVIEWS 
 
 FOR 
 
 OPINION LEADERS PHASE 
 
 Other 
 
 (continued) 
 
 Pastor 
 
 Accountant 
 
 Physician 
 
 Voluntary Fire Chief 
 
 Social Leader 
 
 Total 
 
 TOTAL 
 # 
 2 
 1 
 2 
 1 
 
 100 
 
 C-20 
 
HSiiif