ED 066 554

AUTHORTITLE

INSTITUFION

SPONS AGENCYREPORT NOPUB DATENOTE .

AVAILABLE FRCH

DOCUMENT RESUME

VT 0 14.707.

Stromsdorfer, Ernst W.Review and Synthesis of Cost-Effectiveness Studies ofVocational and Technical Education.Ohio State Univ., Columbus. Center for Vocational andTechnical Education.Office of Education (DHEW), Washington,. D.C.Inf-Ser-5772 --151p.Supewintendent of Documents, U.S. ,Government PrintingOffice, Washington, D.C. 20402

. .

EMS PRICE 17-30.65 BC-06.58. .

DESCRIPTORS. Bias;- community :Benefits; *Coat Effectiveness;Eaployment Programs; Manpower "Developnent ;*Methodology; Program Coati; Program Effectiveness;*Program Evaluation; *ResearCh'Reviews(Publieations); Sampling;: StatisticalTechnical Education; *Vocational Education'.

, IDENTIFIERS Minpolaer Development Training, Act; MDTA

ABSTRACTThe methodologY, to parting! benefit-Cost -studies

. exists, but''there is much ambiguity concerning the .implications ofsuch studies as a' result of using 'faulty methodology. ,For example,,control groups are a necessity, ae are adequate ,sampling prOcedtiresand adjustment for non-response\ bias' and self -selection-bias, and .allstliclies do.aot employ this methodology. Another limitation is thatcurrently a benefit-coe ti. study. will 'only show whether .or !not aprogram appears to. pay. Further judgments must be made on whether toexpand or cut back the:Program, -and What .to give up or .take :In.Spite of these and other limitations. .benef it-cost analysis ,.interjects4 systematic analysis inte; previOusly judgmental .area...The cost and methodological limitatiOns of'such.studies will .inhibittheir immediate widespread adoption. .This study,. written fornon-economists, surveys the various approaches to imprOving the v

,analysis of vocational education ahd related _manpower trainingprograms... (CD)

11;

UM. DEPARTMENT OF HEALTH.ECUCRTM1116 WELFAREOFFICE OF EDUCATION

THIS DOCUMENT HAS BEEN REPRO-DUCED EXACTLY AS RECEIVED FROMTHE PERSON OR ORGANIZATION DRIG.

. INATTNG IT, POINTS OF VIEW OR OPIN-IONS STATED DO NOT NECESSARILY

. REPRESENT'. OFFICIAL OFFICE OF EDU-CATION. POSITION OR POLICY,

Information Series No.57VT 014 707

REVIEW AND SYNTHESIS OFCOST-EFFECTIVENESS STUDIES OF

VOCATIONAL AND TECHNICAL EDUCATION

Ernst W. StromsdorferAssociate Professor of Economics

Indiana UniversityMoomington, Indiana

ERIC ClearinOonse on VOcatianal and Teehniad Edam*The Celt*. for Vocational and Technicel Education

. The Oldoltate Unbrersity1900 Kenny Road . Colombo, Ohio 43210

Most, 1972

The material in this publication *as prepared purivant to a contract with theOffice of Education, U.S. Department of Health, Edwatiord and Welfare. Con-tractors undertaking such projects under government sponsonhip are encouragedro express freely their judgment in professional and technical matters. Points ofview or opinions do not, therefcre, necessaHly represent' official Office of Educe-tion position or policy.

For ale by dui Superintendent ol Documents, Ui. .Ooaisnt Printing ohm Weddings, D.C. MN. . .

PREFACE .

The economics of vocational-technical education programs versus theeffects of a lack of such programs is considered extensively in this work.Cost-Effectiveness Studies of Vocational and Technical Education is designed toalert researchers and directors of education-related programs alike to the issuessurrounding this area.

. The author questions the validity of academic study as 'opposed tovocational-technical study in terms of future earning power for whites andnonwhites. He further relatei problems in this area to the length ofvocational-technical education programs..

. .

The bibliography is subdivided into four sections to allow the specialist toperuse those readings iiertinent to his field.

The profession is indebted to Ernst W. Stromsdorfer, Indiana University,for, his scholarship in the preparation of this review and synthesis report.Recognition is also due Michael E. BOrus, Michigan State University, J. RobertWarmbrod, The" Ohio State University, and Robert C7,-Young, The Center, fortheir critical review 'of the manuscript prior to final revision and publication. J.David McCracken, information specialist at The Center, coordinated thepublic ation's development.

Robert E. Taylor,Direc torThe Center for Vocationa! and

Technital EducttionERIC Clearinghouse on Vocational.

and Technical Education

INTRODUCTION

This study seeks a synthesis of the var!ous approaches to improving theanalylis of vocational education and related manpower training prograins. Thiswill be an economic analysis since the programs in question are designed toimprove the efficiency of the United States labor force and increase the overalllel of welfaie or well-being in the most general sense of this term. However,this study will also discuss the implications of noneconomic benefits to theseprograms*. as they relate to cost-benefit analysis. This study will achieve thissynthesis by investigatinibie current state of the art of economic evaluation ofthese educational prisgfinuN:,

. Both cost-benefit and cost-effectiveness studies of vocational educationand related manpower programs will be summarized to deterinine the current"state of the art" and to draw policy conelusions concerning the optimal letiel ofinvestment in such programs. LikeWise, the methodological literature will ,besummarized, to serve as a, reference point for judging the extant cost-benefitstudies as well as to serve as a general guide to the conduct of such flutireanalyses:

While this study, is intended' for noneconomists, economic analysis andtheory will be used; However, much of the methodology of dbst-benefit analysisis relevant to any rigorous social sciena to that, with appropriate developmentof ideas and economic concepts, the materials covered should be of use to theinterested noneconomist. :

As in any survey of the literature, this study draws very heavily on theWorks of others, ind the present author has only a marginal claim to originality.In this regard, therefore, j would like to express my appreciation to thosenumerous people on whose Works I have relied.

Specifically, I would like to thank the editors of Socio-Economic PlanningSciences for permission to use sectiont of My jointly authored article, "SPecialProblems in the Economic Analysis of Human Resources," in Chapter I. Thanksare due to Einar Hardin and_ Michael E. Borns for permission to repreduce datafroni their stildy; Economic Benefits and Costs of Retraining (Lexington, Mass.:Lexington Books,: D.C. Heath and Co., 1971). lAure M. Sharp likewise gavepermission to dupliCate data from her study, Graduates of Vocational-TechnicalProgranis in Junior Colleges. Also, rwould like to Thank Gerald G. Somers forhis permission tO use part of my jointly authored article in his book, VocationalEducation: Today and Tomorrow,

Thanks are especially due to my former colleagues, Jerome Moss of theUniversity of Minnesota, Teh-wei Hu of the Pennsylvania State University andMaw Lin Lee of the University of Missouri, who graCjously consented to allowme to reprint or adapt from portions of our previously published articles.

Finally, parts of Chapter III are based on analysis previously shown inReport of the Analysis Group, HEW Vocational Educoon Review Task Force.

I would like to reaffirm that I alone am responsible for any remainingerrors in this study, including any possible misinterpretations of authors' worksor ideas.

.Ernsi W. Stromsdorfir

:

%,

TABLE OF CONTENTS

PageTHE RATIONALE OF ECONOMIC ANALYSIS OF EDUCATION ASAN INVESTMENT 3

Optimum Allocation of Public Expenditures for VocationalEduca ti on 3Benefit-Cost Analysis and the Investment Criterion 11

Constraints Which Invalidate the Rate of ReturnCriterion 14Constraints Which Invalidate the Present ValueCrite rion 16Constraints Which Invalidate the Benefit-Cost

. Ratio Criterion 19Summary 20

METHODOLOGICAL ISSUES IN THE ESTIMATION OFCOSTS AND BENEFITS 21

Definition of Cost and Benefit 21General Considerations 22Practical Issues and Suggestions in the Measurementof Costs and Benefits . 30Summary ao

A REVIEW AND SYNTHESIS OF BENEFIT-COST ANDCOSTEFFECTIVENESS stpmEs 45

Summary of Results: Broad Program Effects 49Impacts of Vocational and Manpower Training onSelected.Sociodemographic Groups 69Effects of Vocational and Manpower Training by ProgramArea and Skill 87

SUMMARY, IMPLICATIONS, RECOMMENDATIONS 101

Summary and Implications 101Research in Progiess 104The Future of Benefit-Cost Analysis inVocational Education 105

BIBLIOGRAPHY 107I. COncepttial Issues 107II. Vocational Education 120III. Manpower Programs 128IV. General Works . 139.V. Bibliographies 141

4.7:7

LIST OF TABLES

1. Constraints on Decision Rules 16

2. Differential Shadow Price Estimates of the Value ofClassroom Space, Greater Los Angeles Area 32

3. Objectives and Population to be Served for SelectedFederal Manpower Programs 46

4. Student Characteristics by Race, Sex, Ar, and Program,in Percent, 1969 . 48

5. Comparative Analysis of Cost and Benefit Estimations of Select IdStudies of Secondary Vocational-Technical Education 50

6. Comparative Analysit of Cost and Benefit Estimates ofSelected Studies of Post-Secondary Vocational-TechnicalEducation and Junior College Education 54

7.j Comparative Analysis of Cost and Benefit Estimates ofSelected Studies of MDTA Institutional and On-the-JobRetraining and ARA Reiraining . . 58

8. comparative Analysis of Cost and Benefit Estimates ofSelected Studies of the JOBS and CEP Programs . ° 60

9: Comparative Analysis of Cost and Benefit Estimates ofSelected Studies of the Job Corps and NeighborhoodYouth Corps (NYC). ; . . . 64

10. Comparative Antlisis of Unemployment Reduction.of MDTAand ARA Institutional Retraining 66

11. Comparative Analysis of Percent of Time Employed forVocational Secondary Education 68

12. Net Effects on Earnings (in dollars) and Employment(in percentage points), Vocational Versus Comprehensive

_Graduates for Separate Regressions by Race and Sex . . yo.13. Effects of Racial Discrimination on Earnings and

Employment for Graduates from Selected SecondaryCurriculums, Three Northern Cities, 1959-60-1966 72

14. Dropout Rates of High School Students by Program andAbility Quartile 74 °

15. Present Value of Earning Streams for Males Age 17, byOccupation; yens of School Completed, and EthnicCategory, for the United States, 1960 75

...

I

16. Impact of Manpower Training on Selected SociodemographicGroups, Percent of Time Employed in 18-Month Post-TrainingPeriod . 79. .

17. Impact of Training on Selected Sociodemographic Groups, 0

Total 18-Month Post-Training Earnings, in Dollars 83

18. Estimated Net Expected Internal Rates of Return to SelectedTrainee Sub-Groups 85oo r

19. Impact of Manpower Training on Weekly Earnings and HoursWorked per Week for Selected Disadvanttged Groups . . o .... . 87

20. Social Economic Benefits of Training bi Socio-EconomicGroups for Varying Lengths of Training, in.POilarll 88

21. Percent of Time Employed in Percentage Points ahd Averagebefore Tax Monthly Esmings for Nontollege AttendingVocational-Techriical High Schoo; Graduates ... 89

22. Regression Analysis of Starting Wage, Current Wage, AierageMonthly Earnings and Percent of Time Emplpyed, by ProgramArea, for Secondary Vocational Graduates , 91 .

23. Relative Hourly Wage Rates, by Progam Area, for JuniorCollege Graduates . . 92 .

Regassion Analysis of Wage Ratc on Fint Job, War Rate onLasr Current Job, and Average Monthly Earnings, by

, Education Level, for Separatekegressions by Program Arta . 93

25. Regression Analysis of Wage Rate on. First Job, War Rate onCurrent or Last Job and Average Monthly Earnings, by ProgramArea, by Education Level ' 94

26. Regression Analysl; on Wage Rate for First and Current orLast Job and Average Monthly Earnings, by Program Ares, forVocational High School Post-Secondary Vocational School andJunior College Graduates, Separate Regressions . 96

27. Impaet on Manpower Training on Weekly Earnings and HoursWorked per Week, Males and Females . , . 99

28. Multiple Regressions of Weekly Earnings 'sifter Training, andthe Components of Weekly Earnings after Training, on RespondentCharacteristics: Training Status Classified by TrainingOccupation. ... . 100

:

I.

LIST OF FIGURES

:-.:?

Figure Page

1. Hypothetical Benefit-Cost Curve for the VocationalEdu Cation Program . .' 5

2. Hypothetical Benefit-Cost Curves for VocationalEducation. and Maniower. Development Training Prop* . . . . .... . 6

3. Time Income Stream:4 an Individtial with MtiltipleCost Outlart ........ . : .. .... . . . . . 15

4. The Switching of InVestment Alternatives . . 17

5. Hypothetical BeforelAfter: Earnings Profile ofStructurally Unemployed ManPower Trainee . 39

Iwo

-:o

owAmaw .

1mm

z

No.

,' ;7.44,7.4747.:7-7.5c

. - . .

14407:77.0%777,-,,

171.

10

A

THE RATIONALE OF ECONOMIC ANALYSISOF EDUCATION AS AN INVESTMENT

.

Benefit-cost analysis or cosi-effecti4eness analysis is simply a popular termfor economic analysis of any program or action. This analysis can be part of akirger decision-making. strategy, sueh ,,as systems analysis or of programbUdgeting, oi it may be performed within its own narrower framework:1n eithercaseijt is quantitative analysis Whoselnient is to provide a criterion Or itandardfor decision-making.so as to allocate in a rational and optimal way a given set ofscarce reiources among nUmerous competing needs.

Thus, benefit-cost analYsis- is. a technique which concerns ittelf with the .optu ion ofnum allocat.resources. lt. is a . tool Of inalYsis *filch assesses the. .

alternative cOurses of action in order to help decision makers maximize thenetbenefit to society. The essence of this analysis lies in its ability to evaluate thetotal.Value of benefits against the total coits.

OPtinium Allocation oi Pisblic Expenditures for Vocational EducationA basic assumption in economics is that goods are scarce and that persons

prefer to have more goods rather thanks's. Therefore, it is geneially 'desirable to.employ .resources in those uses which have the highest productivity. Given thetotal .amount of:resources 'available for public and private education;Of all types,it is relevant to determine .the optimum allocation, of expenditures 'on thesedifferent prograna.

Thoretical Criterion. Assuming that the goal of society is, given its valuesand objediVes, to maximize its sociiil welfare,. which includes both economic andnoneConoinic cOmponents, it is ,possible to demonstrate the rule by which thisWelfare may be maximized.-Society has,a variety of goals and objectives, tome ofwhich are complenientery to each Other and some of which are competitive. Foreducational programs alone, there -are several goals. Theis are the goals of (1)economic efficieney-7eachieving' the maximitm outpUt for a given set of inputs,(2) immediate consumption and future consumptiOnthe enjoyment of theprocess of education arid the ability to achieve greater or more varied enjoymentin' the futUre due to,:cine's. education, (3) equitythe realization of a more*tally desirable dietribution a wealth, and (4).socializatiOnthe inculcation ofsocially-desirable values and behavior. . . .

These goals. Can be measitred by appropriate indices of output, Theteoutputs can be, Combined ,to ". represent : an overall measure of welfare orsatisfaction. Thus, we can specify a social welfare function either with respect tothe outputs of all soCial programs:including education, or we can specify a'morerestricted social welfare function which expresses only that part of social welfateaffected by a particular set of programs.

.1

1.2 / 3

Thus, the social welfare function can be written in the form:

l) W = w(gi, g2, - gn)

Where W represents social welfare (or can be denoted as social benefits) and theg's represent the output of different social programs, The maximization offunction (I) is subject to the constraint of the social budget, namely

2) B = (ai + cigi)i=1

Where_ ai is the fixed cost of the ith social program ci is the marginal cost of theith social program, and B is the total resources available to society.

The Lagrangian multiplier, is used to solve the maximization problem, that

3) w.(g1 g2, gn) A E cigi) 13] z

where, ). is the Lagrangian multeplter. Differentiating this expression with respectto gi, then: .

Aci=0

awwhere wi = is the marginal benefit of the ith program. From this itfollows that: ail'

5) wi = ci 6, j = 1, 2, .. njci

and also that:

6) wi X

Thus, in equilibrium, as shown in equation (5), the maximization *Of _socialbene fits is achieved if the ratio of marginal benefit in this example of two' .

goierninent programs* is equal to the ratio of the marginal Cost of these. programs.; that is, the nginal benefit is proportional tO the marginal cost.

(Marginal means the incremental increase in total cost Of benefit due to addingone more Unit of output to a program.) . . .

',An application Of- this principle to die optimum. allocatiOn of publicezpenditures on vocational education versus say manpower development trainingis to spend resources on each prograin to the *point where the marginal,benefit-marginil cost ratio of vocational education is eqUal .to the marginalbenefit-marginal cost ratio of manpower development training. In other words,other conditions being equal; such as the population of persons being screed, if"'the ratio of marginal benefits to Marginal costs of Vocational education is higherthan that of manpaler development training, then the government shouldincrease its expenditures on vocational education Up In the point where the tworatios are equal. This can be done within a flied budge: by duffing funds front

'manpower, training to vocational education or by . expending any mxtra publicfunds on vocational ed9cation . as 'additional fundi beCome available.. More

4

explidtY,Ilte iiiiiicuitfimountof-public_expenditures for vocational educationand manpower development training is at the point wiliiiiike-additional benefitsfrom an additional dollar spent on these two educational progrimt would beequal.

This analysis points out the necessity of coptristing the marpfnal benefitswith the marginal costs of competing educational programs in order to discoverwhich among a set a alternatives is relatively -more..desirable. That is, theadditional benefits of adding one more unit Of output (a student) must be .

. contrasted with the marginal or extra cost of that unit of output (i student). 7

Marginal cost-benefit calculations; are not sufficient, however, .to mike acomplete. decision With respect to investing in social programs such as vocationileducation. In addition to the relative effectiveness of a program as measured bymarginal benefits and costs, it is often important to know What the absolute.level of effectieeness of a program is; that is, in the long run-does the programoperate in the black. To make this detenninaliox of absolute effectiveness, ameasure of average costs and benefits must be performed, for a program ouldbe relatively more effective than some set of alternatives, yet it may .not becovering its long rim average costs. Such a condition will be reason for rejectingthe educational program if other non-efficiency gonls do not intervene. Thus,average costs, thai ia, total cost -divided by the totil 'units of outPut, must becompared with average benefits, total benefits dieided by. the total units of

FIGURE I .

HYPOTTIETICAL BENEFIT-CM CURVE FOR THEVOCATIONALEDUCATION PROGRAM

-Benefit

01110 =1, aim Output Expansion

., Path of ProgramI

VMM IMM 0 .11110

0

.4Coat

output, And, the present value of net average benefits (benefits milui costs)should be zero or positive. . ,.

A Diagrammatic Exposition These principles can, perhaps, be bestillustrated by means of graphs. (See Glennan, 1969, I.) Given that .a decision isto be made on whether to spend an additional sum-of social resources on either .

vocational education or manpower deielopmenflriitting;-the prOblem its :tochoose 'between the two..That IS, Wlich will yield the greater addition tObenefits for the alloted additional resources? Figure 1 showa. the differencebetween average costs and benefits and -marginal coats slid benefiti. Figure 2contrasts the marginal and average cOsts and benefits of the two progiams. Thediipams are hypotheticiL.

In Figure 1, assume the vocational education program is operating at alevel where total cOsts are equal to-Oci and total benefits are equal tO ob1 . The .

average benefit-cost ratiois given by. This ratio is also equal to the slope of.oc

the line segmentoa 1., . .

lSuppoSe the program is expanded Out-to point a2 by adding resonrces

equal to c1C2. Then c1c2 represents marginal costs, the addition to total costs;'and bib2. represents Marginal benefits; the increase in total benefits, due to the

. . .

increase in costs. The marginal benefit-cost ratio is equal to1c2

and slope ofcbib2

the ate ala2.

FIGURE 2HYPOTHETICAL BENEFIT-COSV CURVES FOR VOCATIONAL.

EDUCATION AND MANpOWER DEVELOPMENT TRAINING PROGRAM

Benefit

b '

2

b2

. b.'

6

-

Vocat ional Education

Output ExpansionPath

Manpower TrainingOutput Expansion Path

,

Cost

*Thus, it can clearly beeeen that marginal and 'average costs and benefitsand hence, their ratios, usually differ, the difference.depending on the level ofoutput. . . . . . . . -

In 'Figure. 1, given that costs and benefits are Measured in the same units,average benefits are °greater than average costs, as depicted in the graph. The nextquestion is to compare two posterns-vocational edOcation and manpowerdevelopment trainlog. . _

r_i_Figiie.2 mit the current funding level of oc1, manPawer training has a

higher average benetit-cost ratio,_.°111 ' than- does vocational education with an ...oc 1. ,ob .

.atierage benefit-coit ratio of ---1l -. However, if we expand both programs by theoc

same amount of increased resources, crit,'2the incremental or marginal.

i

benefit-cosi 'ratio:of Manpower training .,12'IL-1? Thus, since both programs are-

benefit-cost ratio of vocational is than the marginal. . .

1

.

covering their average costs, i.e., each is operating in the black, the extra. -S resources should be applied to the vocational education program and not to the

manpower development training program in this hypothetical example. To applythe extra reiources, ,c1c2, to- manpower training rather .than to vocatiPnaledmition would result in a smaller addition to totel output.,'' N.

A Model for Cost-Benefit E:iduai:on ofe Program.. An evaluative mo'del is .

needed to achieve the estimates of costs and benefits to perform tha analysis-above. This evaluative model should have several Components. First, it shouldexamine the nature, of the outpUt processes of competing programs which aredesigned tO fulfill a given set of objectives for a target population Seeond, theModel ihould determine which program and its output procesi is most efficient.As suggested above, this type of evaluation has 'several. major gtaracteristics.Firit, it is mmilitative. There must be Some estimate of cats and benefits.Usually, but not necessarily, these costs aitd benefits are expressed in monetary. ,.termr.s. Second, the evaluation must be .direcily ielated to the specific purposesbeing served by the program. The appropriate specification of the objective orset of O!..)jectives of the program is crucial id the evalintion. Aii impreperspecification of objectives as well as an ill-Conceived choice and construction ofindices to measure the attainment of objectives will result in an invalid'eialnation of the program. Third, the .beiefit-coat evaluation must link benefits-with cats. Treatment of either benefits Of costs in isolation.cannot provide valid

1 information in- making choices among social programs. Vocational education .isnot less efficient Or less 'desirable simply because it costs more, both on theaverage and in margnal terms, to educate a student in a vocational program thanit does td educate hhn in an Academic program in a comprehensive high school.

_ .

In summary, an mippropriate model t evaluate arty program within. .

vocational, education or a similar social p m in education' should 'have:the following steps: ...

1) The 'program objectives Of desired program outcomes must be specified.

-It

2) The processes or activities used to implement the propam must be.specified. In economic terms, this means that the production functionOf production process mist be specified whereby the output of mygiven activity is related to.a relevant set of inputs to that activity.

5) A cost function or cast relationship bard on the production functioMgiven for each activitY must be speCified. .

4) A benefit function (s) must be specified bued on an approrriate indexor set of indices designed to Mauro program outputs. .

5) Costs and benefits mud be cOMpared

Program Objectives Ed Outpost. The objectives of a social pMgrem such esvocztional education must be made explicit. If objectives are stated in terms ofall-encompassing goods nich . as "the improvement of happineee," the programcannot be evaluated since there is no wty to Measure ant a broad outcome si"happiibmis," let alone define it with clarity.

Vocational and technical education are, however, more efficiency orientedand lend themeelves to a benefit-cost framework more readily then other typesof education eitcept manpower trainhly However, the objectives of vocationaleducation are still multklimendonal and the specification of a single functionelrelationship which uniquely encompasses all of these Mulhouse objectieee isextremely difficult and remains, as yet, to be done. It is for this rearm that theestimation of program benefits is generally so much mote difficult then the..estimation of provam costs, although, as we *ell see below, some of the relativedrnplicity in the estimation of costs is more apparent then real, sines costs andbenefits are simply two sides of the sem: coin. Costs ars motive benefits endbenefits are negative costs. The same reral economic principle of foregonealternatives or opportunities governs the:libncePtttal. identification of each t.

, Neverthelese,:without a singe index of benefit: (and Cods) to Joanne themultiple dimension of objectives(andlusth econosnic ard non-economic costs),the practice his been to estimate a Ingle dimension, nth as aarnor wanerate per hoUr, and treat it Is an index of the objective of "efficiency." Thus,..wage rates Of a dmilar unique measure implicitly cotes other dimeadone of theefficiency concept such as the reduCtion of unemployment or At potential ...increase in output due to Misused labor mobility or job athfaction.

To continue; the output of vocational, technical or manpoiVer training isthe acquisition of rertain behavioral capabilities. The objectives of these types ofeducation, whose fulfillment depends in a . functionslly related way on theacquisition of these capabilities, have been emmverated above but bear repeating..These objectives are:

1) Economic efficiency (h1)2) Consumption (h3)

3) Equity (h2)4) Socialization (h4)

The program objectives (W for welfare or well-being) can be expreseed as: ,

W w(h1 , h2, h3, h4)

17

, .

'

A specificition of the relative weights of eich component of welfare, theirgeneral functional form;, that is, whether they are linearly or non-linearly relatedto welfare, as well as knowledge . of the interactions, complementaritiei andconflicts among them would complete the specification of this "objectivefunction.", The next step would be to minimize the total value of this function,given one's limited reieurces. .

However, fa repeat, the prOblem is that we do not have a unique index toMann' W and, in fact, we do not even have a unique. index to measure thecomponents of W, such as h1 or h2. Also since certain affects of these objectivecomp on e nts are almost surely- jOintly created, that is, a given inputsimultaneously creates more than- one type Of Output, the choice of an index torepresent I component which iijointly deterinined with another component willlikely result In liiError in ascribing costs. .

..Production and Cost Punctiont Cost functions can be estimated directly

from cost and output data Without performing the intermediate itep ofspecifying the production function, the proceu whereby the-Program producesthe desired output: However- without an understanding of the- productionprocess, that is, the way in Which program inputs are related 'to program outputsand any interactions among the.inputs, it is very difficult to estimate-the impactof a program. This is a crftical stage in the evaluation procesi since al yet there isno widely accepted theory es to, how vocational capabilities are imparted 'endwhat variables are critical to the efficiency and effectiveness of the learningprocess.

. Therefore, the production process is usually one which in practice isspecified through trial and error by attemptkig to statistically "fit" variousempirical relationships. This, is unsatisfactory, however, since the availableindependent variables which can UfVe U candidates to explain a given learningprocess ire limited only by the .researchir's ingenuity at generating additiOnalvariables, as is demonstrated by more than one study discussed below..

However, leaving these problems aside, if educational administrators act sothat they try to maximize a set of objectives f vocational education, then theproduction process can be specified as:

Yi Xi2, Xin) .

where Y1 is a complex index of output performance of vocational education forthe ith student, and the Xi's are the input used to produce the output ofvocatiOnal edUcation for the ith student.

If inputs can be ravened in money. terms, costs can then be expieised asa functiOn of the production recess, ai_follows:. Z/ f(Vi) where Z is totalcosts, V is program enrollment, i stands for the ith program of a given type. Thiscost function could be expressed ifi linear er nonlinear form; and variables Otherthan enrollment could be added to .the function to account for coit-influencingfactors whose ...fleets one may wids to hold constant. The readt of estimating atotal cOst function will be an estimate of marginal cost.-4he extra cost of trainingone additional student.

The Linksie ot Costs and BenefittIf benefits are nonmonetary, then for aparticular program the achievement of a target level of program performance at

:1.9

,

2.

.;

the -lowest cost (both. monetary...1nd "ndnmonetary) identifies the desirableprogram. Or, .a 'given cost can be set and that' 'program whiCh achieves tilegreatest increaent of 'improvement in output performance is the desirableprogram. For situation; Where costs and benefits aro in moneteri terms, the*general economic rule is to maxingize the presentvalia of net benefits. However,several investment. criteria 'exist to achieve this, . Such is the internal rate of .

return, the benefit-cost ratio, 'or net present value. In the real World, constraints .

usually exiat which invalidate each of these criteria to a degree. Thew problemswill be discuseed below. .

. The Generality of the Model.: This simile Model Outlines the generalapproach One would take to evaluate the efficiency of vocational and technicaleducition programs". Given thst objectives ire dearly.. apecified and thatpefformance indices to ,measure .the achievement of 'the objectives can bedeviaed, then. alternative projects to achieve the objectives can beinvestipted.Input combinations betWeen alternative projects will likely vaiy. Also, inputcombinations. can be varied within a given project. The effects of both types of.variation can be noted.%) both output ind input costs. Ideally, the combinationof inputs for a given cost. which will maximize a given type of output can bediscovered and (Waal educational efficiency can be hnproied.

The Specification and :Measurement of Inputs. The specification andmeasurement of inputs into the process by which voCational Capabilities areimParied to', students suffer from the lack of a widely accepted theory oflearning. In the absence of.in unambiguously acceptable theory, the problem ofspecifying the input variables becomes more complex. There are; however, three .

broad classes of variables to %madder, and,- of course, there are unknowninteractions 'amain These three sets of variables con be dallied asstudent inputs, educational Prom Mputs, and socioeconomic influences.

. The educatiOnal prOcem starts with students, each of whom differs withrespect to characteristics*which affect his ability to learn at the time he enters :the partk,ular vocational program. Students differ in relevant aptitudes,achievement, motivation, and health which create variation in their ability tolearn. One must adjust for :he effects of these factors on anticipated programoutputs.

The educational `Process' In which the students arc engaged hascharacteristics which provide the kerning experience: Students *re encouragedto respond in particular ways, all under the guidance of an instructor, withcertain. characteristics. Finally, the activity takes place in particular phydcal andpsicho-socW learning environments.

In addition 10 the student -characteristics and. the specific educationalprocess which is to be evaluated, the act of learning is affected by otherexperiences and conditions in the students' environment which could influencethe proper identification and meuarement of net educational outcomes. Thetaexperiences can- take place at any time after the educational process begins andbefore the outcome fa mleaured. For example, students, Might take a variety ofother courses which. differentially alter their ability to learn the content of thegiven educational process which, is to be evaluated. Or, for instance, economicconditions could alter the availability of particular kinds of 'jobs after

10 19

graduation. Of course, if one is able to ttructure an experimental model with aproperly .foimed control group, the last set of influences may not be too seriousan obstaCile.

In summary, dozens of variables can be used to account for the three typesof influences noted above, and thus far, little conceptual guidance exists, todictate their choice or their functional form.

Specification' and Measiweinent of -Educational Outputs. The generaldifficulties involved in constructing a properly specified index of output havealready been discussed. However, additional practical problems eitist.

A benefit can be defined as any result of the vocational ethication process.that increases individual or social well-being or welfare. This increase in welfarecan be either economic or non:economic. With rcipect to economic welfare,benefits occur either 'directly by increasing prodle:theity or indirectly by freeingresourcel for alternative vies. With respect to non-economic welfare, theeducations', inoceu 'resuits in an increased level of latisfaction for thoseparticipatingin the educational process.

. The problem tif selecting and weithting relevant output indices becomeseven moiti complex when programs with varied mixes Of "general" and"vocational"' components 'are compared. ! Typically tie output indices chosenare apPropriate to vocational objectives hut slight. the intended outputs of thegeneral education component; this raises serious questions about the validity ofthe resultant program comparison. In i'inore generalized sense, it epitomizes thetype of bias that can result from judging any program on the bails of a narrowly

'conceived set of outputs, without regard for the program's concomitant,effect(positive or negative) upon other desirable outputs.

. Conceptual difficulties also arise when the amount of education isconsidered as a relevant Variable. When holding power o. amount of furthereducation, for example, is utilized as a dependent variable,' education is beingtreated as an end in itself. In other instances, the education variable (like holdingpower) might be considered an independent variable, and its ultimate and actualeffect upon othu outputs Measured.. The choice of treating the amOunt ofeducation' u a dependent or' independent variable changes with the evaluationcontext and ritionale, but making the Choice cannot be ignored,'

Finally, greater attention must be paid to. the iPecification andmeaiurement of developmental outputs. The effect of educational processesupon career patterns, as one illustration, should be determined.: Longitudinaldata are therefore required.

Benefit-Cost Analyais and the Investment Critesion

Given that costi and benefftrhave been successfully estimated, there aretwo additional elements to benefit-cent analysis: time and the interest rate used,to discount costs and benefits to their present value. Both the costs aid benefitsof inveinment in -vocational and other forms of education occur through time.Different investment alternativei are likely to have different time profile:. The

.. Purpose of diacounting is to attach relative weights to them cost and benefittime profiles in order to account for the productivity of investment and social or

private time praerence. (For further, discussion, see Prest and Turvey, 1965, I;Eckstein, J961; 1;'Hirshleifer, 1960,1: Solomon; 1959,11.)

. .Discounting is theoretically justified for I number. of reasons. Thelhat isthat the interest rate _Used in discounting represents the opportunity cost ofinvestment funds; that is, invested wealth usually earns a positive rate of return.Thin, "Y" dollars invested today will yield "Y + X" dollars at some time in thefuture due to the productivity of - the investment. Therefore, reversing theprocess, to relate this frotwe income to itspramt vilue, one must discount thefuture incOme itream to the present time when theinvestment decivion is beingcontemplated. Second, future income is alued lesi than present income. Peoplehave a positive time preference, that is, they dislike postponing consumption.(See Baumol, 1961, 1, pp. 410-413; for a brief theoretial rationale of timepreference.) .

Investwit Criteria A variety of investment criteria is available to theeducation decision-maker. At the simplest level of analysis benefit differentialsand cost differentials can be estimated. The piy-backLperiod an 'also be..estimated. The net expected present value, the benefit-coat ratio, and theexpected internal rate of return can be calculated. Under certain conditions,these last three measures are equivalent and provide the mine guidance todecision-making. The conditions are noted below and exceptions to these ..

conditions comprise the bulk-of this discuidon. .

The Correct Oiterion. There . Is condderable confusion 'oar . whatjconstitutes a "Correct" investment criterion. Thb is due to the fact tat there is

confudon between specification of the appropriate investment rule as distinctfrom the criterion to achieve the goal of the rule. The appropriate investmentrule in benefit-cost analyds is to maximize- the nit present value of benefits.Depending on the nature of constraints, ani,of the last three criteria above mayachieve this rule.. However, there are both practical and theoretical conditions .

which either commonly exist or can be devised whic demonstrate that no instilinvestment decision criterion le theoretically correct for all lnvestnient

..situationi. (See Hirdileifer, 1958, I, pp. 329-352, and Bailey, .1959, pp.476-488.) This di:cation concentrates on only three of the above criteris:theexpected internal rate Of return, the expected net present velue; add 'thebenefit-cost ratio.

Cost end Benefit Differentials t, Cost and benefit differentiali represent a+, necessary but incomplete stage of analYsis. These differentials show the

configuration of the 'data and provide the inputs to the proper (for a given nit ofconstraints) investment criterion. Alone they are not a useful guide todecision-Making, yet,, one commonly perceives misunderstanding of this fact.For instance, a given education program A, costing X dollars more than analternative education program B, is averred (by, its advocates) to be of "higherquaiity"..or (by its detractors) to be "too. costly." But "Walter quality" or-"toocostly" in what sense? Both these statements, taken by themselves, irenoLilense in terms of the efficiency and effectivense of the program. Costs and.benefits must always.be related to each other.. .

Me Pay-Beek Period. The pay-beck period is a dmple ratio of total costs,C, to constant marginal benefit, b, with the constant benefit measured over a.

..given time unit such as a month or-year. Thus, b/C eqUds the pay-back period. .

1221.

This simple index relates costs end benefits to each other and different prograinscan be crudely judged as to their relative effectivcnesi. The criterion is to selectthe investment with the shortest p0-back period.

The pay-back criterion suffers from a variety Of conceptual flaws. First, itignores the fact that costs and benefits of competing investments are distribuVd_through time and have differeit time profiles. Education yields its benefits overan entire lifetime. Diseounting is necessary to make the different benefit-Costprofiles commensurable. Second, the absolute size of net benefits betweenalternatives may differ but the use of the ratio wJlMbscure this. Thkd., as With. I

the internal rate'of-leturn, the pay-back crit n breaks down completely inthose cases where 'investment' alternativesfr mutually exclusive. Thus,' the .

pay-back period criterion has serious Conceptual limitation:Asa decilon-makingtool, and is not highly recommended. .

A Consideration -of Three 'Otte*. The net expected present value'criterion can be stated es follows:. Given the apPropriate interest rate by which.to discount, one should adopt any. program for whiCh the present value of the,discounted .stream of net benefits is greater than zero. Or if More than oneprogram has net discounted benefits greater than zero at the given tate .ofinterest, adopt the program with the highest -present value of net benefits.

The benefit-cost ratio tells the decision-maker to invest in Mose programsfor which the ratio of the present 'value of benefits to the present value of costs .

is Unity or greater..

The result of calculating a rate of return is a simple percentage which callbe compared against that interest rate which rePrerents an acceptable .rate ofsocial or private investmentrethrn. Briefly defined, the internal rate of return isthat interest rate Which makes the discounted value of costs equal to thediscounted value of. benefits.(See Blaug, 1965, I, p. 155.)

A .CWtiqbe of the Three Otte*. Much controversy exists over whatconstitutel the proper investment criterion. The discussion in the literaturecenters aroUnd a critique of the present value and the internal rate of return

_criteria. The benefit-cost ratio is not widely considered. This latter fact isespecially significant in light of federal government practice to employ thebenefit-cost ratio as an investment criterion.

Many. Writers argue thav: the present value rule is most correct since itautOmatically assures that the present value of benefits is at a matmum. Asnoted, previously, this pOsition is taken becanie of a confusion between whatidentifies the correct \maximand as against the appropriate criterion to achievethat maximand. How er, to repeat, both the 'present value' and internal rate ofreturn criterion Wilt re t in-the proper and identical investment decision giventhat: capital markets aèç perfectly competitive, investment alternatives are notinterdependent, all releiant investihent choices are'completely divisible so thatmarginal adjustments cn be -made, and all net returns are reinvested at theoriginal rate of return r higher up to the end of the Project with the longestbenefit stream: (Blaug, 1965, 1, p. 168.)

In this context both are correct and neither is to. be preferred over theother. However, It is unlikely. thut. 'these conditions will pier be met

13

La

.;.%

, . .

. simultaneously. Tne mil world imposes constraints such that each of these ridescan, at times, give advice which will result in the decision-maker's notmaxinizing the present value of net benefits. The following sections' considerthese constraints in turn.

veiConstraint.; Which In Mite the Hate of Retirn Criterion. -

Interdependency. Where two programs are mutually exclusive, the uie ofthe rate of return criterion breaki down At Ii possible under this condition. ofinterdependency to invest "pan activity which has a higher internal tate of returnbut lower present value than an alternative project. This criticism is quite relevantfrom the view of an individual Contemplating an educational investment in him-self. When an individual makes a decision that cominits him to an educationalprocess for a specified period of time, he ear:inter, all other eduCational actionshe may have taken for. the period which Is subst4uently comMitted. In addition,'once he has undergone training during that time period, that exnanse of time isirretrievably hist. Thus, if he decidei to take training as a carpenter, he usu. laily can:not simultaneously take training as a psychiatrist: The two investments are mu-tually exclusive and thus, interdemident-taking one course of action affects theability to take the Other course of action. In short, In this ciintext one can thinkof the human as a site.or locus upon which, in general, only one type of trainingor edncition can occur at a given point in dine. Thus, educational or

! :oecupational investiments in specific human beings have the general Characteristicof being mutually exclusive.

This* criticism of the internal rate of returnAjuet as binding from thesocial standpoint but the relative magnitude of the consequences stemming fromit are probably not as serious. For exaMple, if the Lenstruction of acomprehensive high school on one end of town proves to be an ecomimicmistake, one can always constrat an area vocational-technical school on' adifferent .site in 'another part of town. Or, an incorrect investment in IndividualA does not preclude a correct investment decision to be made With respect toIndividual II, since While one indtridual ia not divisible, a group of individualsfrom society's standpoint is diVisible.

Suceessire Cost Outlays. More than one Cost outlay.occurring Over timecan result in . more than one rate of return beingrestimated for *the lamebenefit-cost stream. The same number of rates can exist is there are inflectionpoints where the cost stream switches to a benefit stream and vice verm. No onecf these rates is necessarily correct. . 1,

From the private or individual standpoint the *occurrence of multipk Costoutlays , is a theoretical possibility due to the risk of unemployment. The

\ 'individual can perceive at least part of the -expenditure necessary to Maintainhim during periods of long-term cycliCal unemplajmeat, as costs incuried tomaintain- his productive capacity in a given skin.. Thus, lie -nay have. a timestream of benefits and costs as appears in Figure 3. Here, as many is fourinternal rates of :return conld exist. However, it is nOt likely that thort-terMcyclical or seasonal unimployment would result in any measurable illdeterioration. Long-terM structural unemployment could; hoWevert .

14

\..

.,

FIGURE 311M2-INCOME STREAM OF AN INDIVIDUAL WITH MULTIFLE

COST OUTLAYS

NetBenefit

OrCost

\ , . . .

,

.The uneniployinent eximple is Similar 'front the standpoint of society.Although one could argUe -that society. in any case is commitved to keeping .

,.. its members alwe, or .st least.a certain number of them, in order to assure its ownContinuity, society may still incur differential . costs which are uniquely 'associated with maintaining a given skill structure and level of ability. These .

costs shoiild be counted as necessary costs to assure the viability of the originalskill leveL..... . , . . ..

,

-FinallY, from both society's and.the private or individual viewpoint; if the2 person had to reinveit in himself due to the fact that technological change had .

destroyed the economic' relevance of his .previous skill, this new investment costand the benefits flow* from it should be trealed &Sari entirely nevibenefit.Costsequence. . . ..

,

Chasing Rate of Interest. Investment hi licational education over time' Will likely change the dittribution of inconie and hence, other things equal, wilalso' change the social opportunity costOf inviitment funds Which depends; inpart:' On the distribution of income. In ads case,X uniqyely calculated rate of

. return becomes conceptually:jzielevsnt shice it does not reflect the' changingsocial opportimity cost rate o investment funds. .

Constraints Which Invalidate the Present Value CritedonMultiple Interest Rates. An individual may invest in himself by using

personal savings, borrowed funds, or by reducing his current consumption. Adifferent private interest rate may be relevant to each of these sources of funds.Assuming the individual did not use "some weighted rate of interest to representthe interest rates which apply to personal savings and borrowed funds plus therate of time preference he attaches to foregone consumption but instead choseto discount the stream of costs and benefits of .different alternatives by eachrate, the ranking of alternatives at one rate may differ from the ianking ofalternatives at the others. it is then unclear u to which relative ranking is thecorrect one.

TABLE I

CONSTRAINTS ON DECISION RULES

a

, Difficulties Dow With.

Oresent Value ., . Internal Rate' of Reties' When

. , ,

i. Different discount rates areused to evaluate a set ofprojects with dissimilartime-benefit streams.

.fResult:. Different rankingsmay occUr for each discountrate. .

,2. Discontinuities occur such-'. that project costs become.. large relevant to current

.I.A.:.

.-,tf4fte., resource*: .

Reault: Adoption of a given-,. pioject on the basis of its.., ...higher present .vilue may

preclude the adoption 'of twoor more smaller projectsAue summed present value.is larger than the original"project.

.

. Budget constraints occur.-

Result: This is 'a Variant ofthe discontinuities constraintand, again, the likelthood

, may be that failure tomaximize present value willOCCur. . -

,

1.

.

2.

.

3.

,Pr oject s a.re mutuallyeapt_ive.Result: A hth rate of returtrproject muy be adopted.which preclUdes thepossibility of maximizing netpresent value.

The market interest rattvaries over the life of theproject.

...

Result:' The single computedrate of- return becomesconceptually irrelevant since'all time periods are treated ona par. This it the: mast,fundamental conceptual'failure of the rate of retamrule.,

.. .

More than one cost outlayoccurs over time. . .

Result: a) Mutliple rates ofreturn are Computed no oneof . which is conceptuallycorrect;, b). Problems of .,

mathematical estimationbecome extremely difficult.

16

25

,

.

In addition, th many practical situation when .a angle unambiguous ratecannot be donut; advice is. often given that more than one rate of intarest*mad be wed in order .to provide a range-of eathnnes of discounted costs andbenefits. This again. may result in a switch in the differential rankings of .

- alternatives vis-a-visythe different rates. The fault will be betweenin vestment alternatives wdl become indeterminate if oes at pts. to employ

. both renkinp. ; .

. A aggeited solution to this twitch* problem involves the 'diction ofthat interest rate whiCh makes the net premat value of the set of alternatives all

! equal. This- rate then ones is the ecut-off point in selectincthe appropriateranking, and hence, the appropriate investment. Int Flows 4, the present valuesof education progresni A and B are equal at an interest sate of 5. If the social.t) preference rate is dways liiiiiiWc-pethaps-rets-c-thewthiepresent-value-ofprogram A is always &eater than programa end program A thould be chnen inpreference to S. But, if. dthgreement Wits as to what is the proper acid thnepreference rite, (for butane, ."is its value to the left or rigu of. e?"), then one isno better off than before. The dilemma remains:

FIGURE 4MS SWITCHING OF INVESISIENTALTERNATIVES

NetPresent

Vilpe

'Interest

Rate

Budget amstraint. The present endue rule will sometimes prove to beinvalid when a Lbudiet-constraint_Or_investment -dikontinuities face the

ciecision-Maker, If one follows fin adviee to iinest first in thit activity which hasthe highest present yob., it may still.be that some alternative combination of .

lineaments *ill prove :possible, each of which requires a smaller investmentoutlay but which, when taken together, yield a summed present value greater ..Alum the single larger investinent. The proper strategy when budget constraintsor discontinuities occur, then, as long as the alternatives are not mutually 7

exclusive, is to exhaust the budget by choosingthi set of alternatives with thehighest internal rates of return. This Will actually maximize present value for theset orinvestmenti. .

Such a constraint is a major problem from the standpoint of the individualseeking to educate or otherwise iniest in himself. As kweitors; etudents have

----Ihnited-ebeess-to-investmentsources-Alsootildents are relatively unproven in-the labor market so that there is a great deal of risk and uncertainty concerningthe benefit stream of an investment in them. Capital markett are relativelyimperfect in the area of human resource inveitineni due, in part, to theunwillingness of neditors to accept a person's own self as loan'collateral as wellIs the quasi-illegality of indenturing oneself.' The capital created by theinvestment in educatioxis real, 'but it is embOdied in and cannot be separatedfrom the human agent. It cannot be used as collateral in the same way thatphysical capital can. It cannot be sold. High risk and liquidity premiums wbuldhave to be charged in addition to the' opportunity cost rate of 'capital if thecapital market were to make funds prierally- available to investors in this area. .

(Becker, .1964, I, p.. 35.),Institutional constraints are such that these very highinterest ratet are not charged. Instead, lower rates are set and 'the pool ofinvestment funds is rationed aMong thole projects which qualify st the lOwerinterest rates. As a result, investment funds are not generally, available to fmanceone's self-investment in education and similar human capital investment.. .

Perional loans are made strictli on a person's representation thathis actualor expected income stream and, hence, his expected pipits! value, is of suffidentsize and certainty of being realized that he can pay back the loan. Thus, in suchcases the loan iemade on the basis of accepting the person's expected capitalvalue as ecilisteral, but this practice occun normally titer and not before theperson seeking the loan has created the Capital valise which is embodied in him.In Ilie with this, most student loans ma6e by banks are offered mainly' as apublic service and are made on the basis of the parents' expected incOmestreaniand not on the basis of the great expectatkins of the student seeking the loan.

Hence, the individual is genera* faced with inveament budget constraintswhich do not allow hint perfect choice aMong all 'possible investmentalternatives. He may have access to sidficient funds tO contemplate training as icarpenter bill not as an electronics technician or as a medical doctor.

Investment budgets are also constrained :froth the standpoint ofgovernment, such as a schOol district, though disogreement etists as to the exactnature and seriousneis of this conetraint. Legislative knits are set for specified 'periods upon amounts to be .spent by school district and other governmentalunits. Even though nevi fuqds are voted for new bUdget periodi and the budget

.6 .

18 1

4)

periods continue" through time, a" short-nm constraint exists which can be: repeated indefinitely. ---

Only in the broadest sense doe: n constraint exist for the economy as a. whole in thil area .. of education for it is :difficult to xonceive of a givenkwestment which would be so large as to absorb a significant proportion of thegrosinational product.

Constraints Which Invalid* the Beeefit-Coit gado CfilefiOR

The ItenefjOrnsrratio has some of 'the operational shortcomings. of .both_the expecteil net present valisegule'antthe expected internal rati of return. Like..

the expected net preeent value rule, its use will cause problems it more than oneinterest rate is used to discOunt, That is, the '.choice of the most efficientinvestment -alternative may switch. However, if budget constraints or

Of return, is preferred -over the present _value criterion. Given theinterest rateused to discount, choice of those hive. stments with the highest ratios willmaximize net present benefits.. But, it investments are mutually exclusive,. theUse of the benefit-cost ratio, RS with the expected internal rate of return, 11111)give an incorrect result Unless the returns from the investment *are reinvested atan interest rate at least as high as that yielded by the next best alternative and atleast through that time period represented "by the invest:tent .alternative havingthe longest time profile of costs and benefits.. .

The resolution to the above :Witching problem under condition otbudgetconstraint is to discount at only one interest. rate. Notothat this single rate ii notnecessarily the social or private interest rate representing the opportunity cost ofcapital. The proper rate is the highest marginal rate of return on that set ofinvrnment projects which just exhausts the, investment budget. Then thoseprojects in the chosenlet which are discounted at this rate must have a presentyalue Of zero or greater:- Any" project with a present.value of less than zero whendiscounted at this marginal rate should be excluded. The method fOr finding theinvestment set with the highest Marginal rate of return is to discount the array ofinvestment alternatives at differeat interest rates until that set Of investmentalternatives is found Whickjust exhausts. the investment budget. (See Hirshleifer,1960, I, Appendix to Chapter VII; also McKean, 1958, I Chapter:5 and 7.) Onethen chooses the set with the higheit rate.. However; this .tec!.?nique can becumbersome and impraaical if there are a large number of alternatives andLaterdependency exists among them. With interdependentY, an entremelY largenumber Of possible combinations of these. alternatives can exist; all of whichmust be tested. Such a situation is likely to occur in. investments in humanbeings. .

. It is important to note that -the bulget could conceivably be soconstrained that the nuMber of investment 'projects "Would be insufficient to:include those-which would lower the marginal rate -of retum down to the-socialor private opportinity cost rate of capital. If the Social rate n used Ina situation ,

where it is less' than the marginal internal rate of the contemplated set ofinvestments, then the projects wth likely be adopted which will not result inmaximizing net present value. . .

19

r, Hirthleifer points out that even this nali, while e nodal andplausible (Me under conditions of capital rationing or budnet constraint, is notstrictly correct. First of all, thimargind project may not have an unambigutms

. rate of return. Second, even if there is ah unambiguous internal rate of return,one may choose the wrong course ..of artiontifedem Consideration is Msde of theearning value of resources yielded by each Project al well as the market rate of .

interest by which intertemporal shift of benefits of a given benefit stream cin beundertaken. (Hirahleifer,..1960, I, p.,I71,) r

In luminary, given the qualifications above, when there is capital rationing(and this is a coMmon situation for an individual contempleting investmenthimself), the -benefit-cost ratio' ia the proper 'criterion for ineestntent'making; since-by- choosinr the set-of investmanti aritiltheldgliett ratios hi willthereiq maximize net present value. When there is no budget constraint, and forsociety (not a governmental unit, including the, Anal goveinthent as a whole)this is usually the case, adopting those projects with the maximum net preeent

'value is the proper course of action.

Sentrany

Both marginal and average costs and benefits must be estimated. Theaverage benefit-cost ratio indicatei the abtelute level of performance of a socialprogram. The Marginal benefit-cost ratio indicates the relative level ofperforrnmice of a social Program.

Given that the average benefit-cost ratios of tem comPeting programs areboth greater than unity, welfare Will be maximize:I by Stiffing funds to thatprogram with the higher marginal benefitcost ratio, uptil the ratios of the twocompeting programa are equal. .

An arcurate evaluation of a serial program . requkes an accuratespecification of objectives, a specification of desired outputs, appropriate indicesto measure these outputs, and an appropriate Model which verifies the mannerin which program inputs create the desired program outputs.

A program cannot be evaluated in my efficiency sense by comparing costsor benefits . separately. High costs imply neither a "high quelity" nor en"expensive" program. High benefits do not normal:0y Inplya moithwhileProlirant

The appiopriate investment decision rule is to invest in thstsociel programwith the highest net present. capital valise. Three different investment criteria.exist to achieve' this efficiencY rule: the net present value male, the internal rate ofreturn, and the benefit-cost ratio. Under appropriate conditions, all thme rulesyield the same result. However, the real world imposes constraints, which in anygiven case, can invalidate any one of these criteria..

0

METHODOLOGICAL ISSUES IN ME ESTIMATION

OF COSTS AND BENEFITS

Definition of Cost and Benefit

Costs ire defined in- theirmongeneral sense-aropportunitfoitS:-Thiris,the cost of doing anything is the. Value of the next best opportunity oralternative which has to be foregone because of the particular course 'of actionone has taken. Thus, in the most general terms,- tbe cost for sn individual to.invest in vocational or technical education is the cost of (I) not being able towork simultaneously ht the labor market or (2) the cost of foregone leisure or(3) the inability to engage in production at home.

There is often a confusion in the literature since costs are sometiMestreated as being conceptually -different when, in fatt, what differs are theproblems involved in measuring them. Thus, some writers will categorizeeducational. costs into direct outlays, indirect mitlays, opportunity

. costs-meaning wages foregone-and capital costs. Yet, it bas to be rememberedthat.all costs'are opportunity costs and one should not consider cost elements asconceptually different simply because they may occui at different points intime, accrue to different individuals or groups, or take different institutionalfornis such ai wages or tuition.

Benefits are just the reverse of costs. They ire opportunities gained u aresult Of engaging in some activity. Thus, they can represent (1) increases in thevalue of labor market production or activity or (2). increases in the value ofconsumption or leisure or (3) 'increases in the value Of non-Market Or homeproduction. In short,-they represent increases in the productivity of market andnon-market production Sad constimption.

To- *Void errors in underestimating or overestiMating costs and benefits,they shOuld ideally be messuredin terms Of utility lost and gained. Then; therewould be no confusion that the complex multiproduct -nature of- educationalinvestMents:were being crassly subsumed under money returns andcosts alone.Miuurement of utility, however; is a counai of perfection. It cannot be done,given the state of the art. And, in an imperfect wOdd, it is improper to considermoney costs and benefits as meisures of utility oreven as good indices or proxiesfor it. Given the Complex nature of educational investment, both in terms Of itscosts or benefits, it is best to simply indicate the cOMponents oInosts or benefitsbeing measUred and not claim wider validity for them as prOxies of utility. Thisdiscussion brinp us too general-consideration of-the methodological- issues inthe measurement of costs and benefits: In short, the measurement of Costs is justas difficult as thc, :measurement of benefits, mild previous statements by someinvestigaton concerning the greater ease of cost estimation tic based on atimPlistic concept of cost. (See Judy, 1969, I.)

30. 21

General Considerations

:'

A. foray into the methodoligical issues surroUnding the distinction of costsand benefits of invistment in social programs designed to improve humanwelfare brings a variety of basic conceptual problems to the fore. Thr bestsummary discussion of these Problems to date has been perfonned by LesterThurow in his book Investnwnt hi Human Gepital. (See Thurow, 1970, I,especially Chapter 8.) The bask issues he discusses can be outlined as follows:

.1) Eamingmaxinfizafion versus utility maximization;2) Complementsrity in production and consumption;3) Joint costs of production, consumption and investment;4) No-w-wi-iirk-etion and consumption; .

5) Change in perferenees due to the act of educating ortraining;6) Risk dueto lumpinen of kinitinenti; and7). Complementarity, substitutability; and inseparability of human skills

and abilities.trarow lista seviral other points, but these above are of most interest with

respect tO investment in .vocational and technical education or other train541progrein since they impinge direttly on the measurement of costs and benefits.

. tin addifion to these points, one shmild conskler:8) Externality;9) Income redistribution effects as theyinfluence the determination of

costs and blnefits;10) The kifluence of unemployMent on the determination of costs-and

benefits; and11) The problem of the control group.Each point will be considered below . with specific reference to vocational

and technical education. And, where applicable, each point will be consideredWith respect to social; individual or governmental estimation of costs andbenefits. ,

Earnings versus utility niudiniration. Even thOugh vocational and :teChnical education is well as manpower 'training have a more iinmediite labor,market orientation than do other forms of education, such as a .liberal artscollege eduntion or the pursuit of. a pneral cUrriculum. in high school, it isdangerous to evaluate the former types of educaikm.onlyin .teitni 'of 4arninpmaxháizition Earninp are only one of the elemeAts which comprise one'sutility. One of the elements of utility one gains besides earning are directconiumption benefits during the educatiOnal process itielf as vidl as improvedpossibility for the enhancement of consumption after education. If persom arerational in their pursuit of utilitY.or welfare maxanization, they will gravitate.to .

those kinds of educaition and Occupations which give them direct consumptionbenefits along with increased earninp. This is the crux of the matter wheneducaton, economists Ent others,. seek to evaluate the degree of "job .

satisfaction" involved in Career Choke.Of course, if all persons are raticinal, hicludin those : who. pursue the

college preparatory or general: curriculum in high school, there is no necesoryreignn after the fact to assume a priori that vocational or technical students will .

have greater job satisfaction than other types:of student's. Presumably, each

22 : .31

n.3,

. -group gravitatet-to-that-type-ofintiningwhich:wilimaximizeitrexpected futurejob Satisfaction. Thus, lob satisfaction and other charicteristics of a person'spost-training situation which are measures of psychic well-being and the degreeof consumption benefits being received on the job Must be directly measured.HoWever, if different kinds of persons gravitate to differentdprograms;there"remains the difficult task of establishing unambiguous scales to measure thesedirett consumption and plychic benefits. Different eleffients may comprise theconsumption and receipt of psychic benefits by, different groups. Thus, even ifyou ask the same kind of question of these different groups, seemingly uniformand consistent responses ;may -.have entirelydifferent_ meaninp- andbeincommensurable: ,

_Complenieniarity -in-Production-and-Consumption Since one's humancapital. is inseparahle from oneself, in the act of producing one also consumes.This occurs siMultaneously and failure tO account for this phenoMenon can leadto an incorrect measure of behefit. Other thing, equal, if a person dislikes hisjob, one may tend to overeatimite the benefit to the indieidual person. However,if he likes his job a peat 'deal, other thinp equal, one may tend to underestimatethe total benefit receved. There is no reason, however, to assume that one typeof curriculum automatically 'has a greater Tier (or under) estimate of measuredbenefits due to this phenomenon.

Joint Costs of Produe don, Consumption, and Investment. Thus, reductionand .consumption on the job are joint due to the fact that any economic activitybased..on human capital. can't beieparated from the human agent. Likewise, theinvestment itself is joint, producing both production and consumption.capabilities. This fact complicates 'the estimstion of coats considerably. Anexcellent example of this is the Job Corps. Here, participants enpge in trainingat residence centers. They simultaheOudy produce, consume and invest inthemselves. Their maintenance costs support all three, of these activitiesimultaneously since the activities are joint. Even tliough 'society or theparticipants' familiei won't have to maintain these participants, Ihey would in:luny cises. be msintained at lower levels were they not presently in the JobCorps. Thus; the question becomes, why isn't any Measured increase in..consumption treated as a sociallenefit of the program? Or, should it bctreatedas a transfera bevefit received for which 'no reciprocal service ot benefit isrendered-and hence not counted as a sociaflikin?

. Btit even if one agrees from., the consumption standpoint to treat theincreased level of maintenance as a transfer, the higher level of nutrition, medicalserviCes, clothing, ssid the like, should centribute simultaneously to increasedproduction' in the Job Corps Center as Well as improve learning while in thegter. Hovi can one sort out the immediete consumption component from the in-vestment compOnent Of the higher level of maintenance? Assuredly, the hiiherlevel of maintenance is not all transfer -payment 8Utsince the three activitiesare a joint. Outptit ofmaintenande,-theY Cannot ba separated

, ,

Similar kinds of problems exist with cooperative vocational education. Thework component of a coop, program is jointly' production, consunaptinn andinvestment:is the wage, rate. the.. student receives a "measure of the student'sproductivity net othis-on-llie4ob training or investment? Economic theory

23

would arbue that it is not likely _to be completely net of the on-the-job trainingcoststhatis, the student will not pay for all of his on-the-job training costs via areduction in his wage rate. To the extent that a coop student will have a job with'components that are peculiar to the finn'sown operations, we could expect thefirm to pay this cost to cover this firm's specific component of the job.However, to cut down turnover and loss of his investment in the student, theemployer is likely,. to share both the costs and the returns tO the finwspecifictraining component of this job with the coop student. Conceptually, these costand benefit componenti should be separatedoutbitthis is often difficult to do.No one, to date, has attempted an empirical resolution of this issue. Thejointniss of these -activities renders a separation most difficult from in empiricalitandpOint,(Problems-of-prorning-jOint ceets-will-be discuned-further below.) .

Non-Market Production and Consumption. Non-maiket prOducilop andconsumption is a major consideration .in any Complete evaluation *coT theeffectiveness of voCational and technical education or manpoWer training.Obviously, persons trained in vocational and technical skills, such as electricians,auto mechanics, engineers and the like, are in the position to provideconsiderable non-market preduction for themselves since the skills ,themselvesare in the high &teas of demand: craftsMen and semi-technical, and professional.This production should be imputed as a return to the education, but as yet, no

. effort tO do so has been made in benefit-cost or cost-effectiveness studies.Likewise; women who have learned vocational skills' nuy be in a pontion toprovide higher valued services as housewives than those with winch training.With respect to the 'ghettoized, poor, andother disadvantaged groups, courses inhOme economics and consumer education may yield very high nomarketreturns if the assertions concerning the instability and leek of Parental guidanceand know-how to' provide homemaking Md economic skills among poor familiesis true.

Change in Preference& It is difficult enough to evaluate *consumption, benefits When one 'spumes that a penon's tastes and preferences staY the sane,

thus assuring that the relative weights one attaches to a benefit Or cost d,o notchange. However, .the purpose of 'ediication, including iOcational ind technicaleducation, is to change a person's preferences, tastes and attitudes. Severalpoints ire of concern here. (See Wiseman, 1965, .1.)

First, since it can be' assumed that . education changes tastes and.preferences, the value and inightsthat is, relative prices one will put onconsumption, home production and leisure activities before one undergoes aneducation program is likely to be different from the valuation one altos tothese economic activities after one has completed his education..Which set ofvaluations or prices is the correct one? Should we add or deduct any differences

. in valuation betnent the two perkith? ShouldWthe valuation and prices afterthe educational process be comidered even. though the valuation tncl prices wereCreated by the education process itself?

Next, this change in tastes and prefftences may alter one's taste for leisure,work, and investment in education. Persons with higher levels of education_ .

generally work longer hours so the marginal value of leisure-time May be higher

.y4 ,t

47

/

for this type, of peison. In, any case, if worideisure-intestment preferenceschange, this will change the measurement ot opportunity costs at -.ell as benetits.

. Risk and Lumpy Investments. Human life is finite. Investments in humancapital often extend overionger periods of time than investments in physicalcapital. And, persons cannot otdinarily train for more than one occupation oroccupational Set at a time. If a mistake is snauethat is, one takes training in askill which proves incompatible with one's. needs or which has its demandeliminated by changing technology or tastes or competition, then, there may bevery little time left to recoup one's lossesor to retrain in a new occupation area.The only meaningful alternative, as with many older displaced Appalachian coalminers, for instance, may be to drop out of the labor fotce altogether..

Two observationi on vOcational and technical education and"manpowertraining are pertinent at this point.

'First, while it is conceptually reasonable to train for the "job oftomorrow," OUI manpower forecasting techniques are not accurate enoua% topermit this type of educational strategy. Hence:Ilse focus on quick jobplaCement and training-relatedness is a proper one in voCational education eventhough to date, indices to measure training relatedness are atill too crude to beof much assistance in guiding investment decisions in the area .of occupationaltraining.

Second, given the flexibility Of manpower training and the general shortduration of such training, the gestation period of this investment strategy istelatively short and hence the opportunity costs, especially due tosthe tisk ofsnaking a mistake in occupational choice, Lie relatively low so that manpower'

'4.: training has the flexibility to overCOme the general lumpiness of human capital

,investment. .

Next, this lumpiness of human capital argues for a shortening of 'thegestation period whenever possible. There is no ironclad.reasim, after all, whyhigh school must last four calendar years or whysummer vacations must occur.Thus, the irelative cost position of vocational, and technical education canimpiove- via-a:Via -its close competing substitutes, such as the general or collegepreparatory program, if efforts are nude to shorten the training periods. In thisregard,_ also, _ cooperative_ vocational education may have a relative costadvantage over other types of education including straisht vocational; sinceopportunity costs of foregone wages &Telesis, the students often work and aschool all year round, and job placement may be more quickly achieved.

In short, while secondary vocational am technical education generally costmore than the general or college preparatory curriculum, this cost differentialcan be narrowed significantly by appropriate educational plannina. Sinceforegine wages are a major cost of education even at the high school level, coopprograms-and- propams designed to shorten the calendar time-in-school may--represent appropriate.educational strategies.

. .

-Complementarily, Substitutability, and Inseparability of Skills. Thisphenomenon arises from the fact mentioned earliet that it is impossibk toseparate one's human capital from his person. As a corollary, it is diffscult to'estimate the separate net effects of different kinds tof human capitalsimultaneously embodied in the human agent and thereby determine the

. 1-

contribution of each to one's earnings or welfare. This problem is especiallyignificant in the area of vocational and technical eetication due to the presence

of on-the-job training. It is important to measure the contribution of one'sgeneral education courses, his vocational courses and, since we are usuallydiscussing a followup period of employment, his on-the-job training in order to .

make appropriate judgments as to the optimum relative mix of each kind oftraining. The problem is further compounded by the fact that much on-the-jobtraining is infonnal rather than formal. It is possible to sort out these separateeffects statistically, but the average effects of the investMent elements aredifficult to estimate with any precision where they interact jointly. Jatob Mincer(1962,01) did estimate the amount of on-the-job training costs by essentiallyworking backwards from estimated rate of return differentials between groupswith different amounts of education. His methodology is useful where directmeasurement of on-the-job training is not possible. However, Audies usinginterview data can collect the necessary information on wage differentials amongskill levels within the same occupation to. arrive at cost estimates and time spentin on-the-job training. Such cost and time estimates can then be entered

. appropriately in a regression model to control for the effects of .on-the-jobtraining.

Finally, cooperative training carries with it the same measurementproblem. Namely, how muchof the measured benefit is due each to the general,vocational and work experience components of the educational program? Theseseparate costs and effects should be identified in order to make decisions as tothe optimal mix among them in the training strategy.

Externality. An externality is an econornic effect caused 'by an economicagent which bestows economic costs or benefits on secondary parties. Thesecondary party has no control over the receipt of these costs or benefits, butthey influence his own economic behavior in positiie or negative ways. On theother hand, the individual creating the externality is indifferent with respect towhom or where the cost or benefit _finally resides. By its very nature, the.

. _

externality cannot be priced and hence, rationed among possible recipients.. As aresult, the creator of the externality is indifferent to its existence, and the factthat he may be creating costs or benefits elsewhere in the economy hiJes notenter ink, his own investment decision.

The standard example of an externality, is air pollution. In the area 'ofvocational or technical iducation, an example wGuld be the existence ofcomplementarity between ; given skilled technician ansi the remaining membersof a research team such that the technician's productivity raised Of reduced theproductivity of the remaining member; of the 'team To the extent that theother members' productivity rose' (fell), their wage rates would rise (fall), but.there would be no way that the technician could request .(of be charged) aportion of the other parties' pin (or loss) in wages due to his role as a teammember. To some extent, the entrepreneur whO brought the research teamtogether would captm-e these external benefits. His role is to internalize themwithin the company.. But he captuies the, benefigind not the worker, whoseactivity results in the external benefit. . 0;1

2635

With respect to a given skill, such externalities should be accounted for inany complete accounting of costs and benefits, but this is difficult to do forseveral reasons. First, because there is no market mechanism (though one couldoften be established) to price and ration these benefits, their quantities andrecipients are indeterminant. As a corollary, the very pervasiveness ofexternalities makes many of them take on the characteristic of a pure publicgood, so that in the case of a benefit, the consumption of this externality by oneindividual does not deny the use of any plot of that benefit by other individuals.Since the externality is not rationed and since different persons weigh the valueof it to them differently, in the absence of prices, one simply cannot estimateihe folnl quOlity Of Benefit bestowed onindividuilliir society.

In addition, it is difficult to identify externalities and areal possibility fordouble counting and, hence, overestimating costs or benefits exists. For instance,Burton Weisbrod (1964, I.) lists socially desirable attitudes and behavior as anexternal nonmonetary benefit of education. Is this really an externality or just adirect noneconomic benefit of socialization? It is, in part, both. My sociallyappropriate behavior will' yield direct economic and psychic returns to me. Tothe extent that my behavior is appropriate and predictable, other individuals,benefit from a more stable, predictable environment. Due to my behavior, theirlevel of security and haPpiness will rise as well as their wage rate or earnings, yetthey, will not compensate me for this improvement in their well-being. Forsociety, part of the externality, then, is directly measured by. the second parties'increased earnings, but how much? Of course, the rise in happiness eludesmeasurement at this state Of the art.

In the recent past, education has had a good press partly due to presumedlarge external benefits. The extent of these benefits of course is unknown, sincethey are, by their nature, unmeasurable in most cases given existing ecOnomicinstitutions and market structures. Recently, however, the presumed largeexternal benefits to education, especially post-secondary education, are beingvigorously challenged in the literature. (See Hansen and Weisbrod, 1969, I.)

Income Redistribution. Income distribution changes present a majorproblem in the estimation of-the benefits and.costs to an educational prograni.The issue is as follows: A given benefit-cost analysis must take as a given orconstant the distribution of income before a given educational program isimplemented since the distribution of income is a major determinant of prices,wages, interest rates and rents. However, the very purpose of educationalprograms, including vocational and manpower training programs, is to alter thedistribution of income in favor of some target population, such as youth, thedisadvantaged, blacks, or Appalachian 'coal miners. Thus, relative prices willchange if the program has any noticeable impact at all, and the problem becomesone of choosing which set of prices to use inevaluating the inVestment value ofthe program. The before and after states are non-comparable, especially if theprogram is a large one, such as a rutionwide expansion of cooperative vocationaleducation, area" vocational-technical schools, or two-year community orpost-secondary technical schools. Thus, the logical basis on which to make theinvestment judgment is lost.

27

. A less serious problem is the direct income redistribution effects which can ,

occur ai a result Of a even educational program. Thus, en. areavocational.technical School may flood a labor market with welders to the extentthat the increase in supply reduces the wage rate of the existing journeymenwelders in the market. This represents a capital loss to the existing journeymen-welders Who undertook the expenses of their training tioder-the_aupliption ofreceiving the . higher Wage rate necessary to yield them a profitable rate-of return .

on their investMent. An awareness of the impact of vocational programs on thesupply of skills results in craft unions not taking kindly to the attempt4of public .

,education officials to expand their apprenticeship programs motherwise traindeservins groups, such as blacks, in their skill areas, .

The Problem of Unemployment. Concerns such as the above existelienwhen there is full employment in 'the econOmy. When cylical or-deficient-demand employment' exists, the problem is Compounded mainly- beciUse theimplications of income redistribution become so much more direct. In situationsof less than full, employment due to deficient aggregate deinind, theie is alwaysthe very strong possibility .that a retrained worker from a manpOwer programwill simply displace an equally deserving worker who is not.tormally trained. .

Here, .the concern is not only one of income redistribution;bist" oneof the ,realizition that there inky be no net Increase in national product While valuableresources have been expended, thus resulting in a het loss for society and a gainfOr one group of individuals that may not even totally offset th . losses in welfare

. of the displaced group. (See Borus, 1966,111, where he terms this phenomenon. 'the "displacement effect.") Of course, even under full employment, if theie isincome redisjgbaition due to a program, one can, strictly speaking, .fluke no .

judgment as to whether social welfare has improved because of the change in thestructure of relative prices and the .theoretical inability to make interpersonalcomparisons among People concerning theif relative losses or gains of utility dueto the change. Where, then, .does this leave us? Possible income redistribution .

effects should be taken account of and measured. To date.no beneflt-cost study'does much more thin proiide lip service to this issue .

The existence of less than full employment compounds die. measurementproblems of beneilt.eost analysis in other ways.. For instance, as the level OfunemploymUnt as well as its distribution among occupational classei changes, .

the value of the embodied human capital represented by these acquired skillsamong .occupationil groups changes.-Thit no unique capital 'value for a giveit

exists. The expected capital value fluctuates for reasons, independent of anyfundamental underlying demand skill.

The question is, should One allow his measures of the value of human ...capital created by 'an educational program' reflect the 'Phenomenon Of cyclicalunemployment? From a private standpoint, earnings benefits as well as foregoneearnings should reflect unemployment experience. However, it is not certain thatthis type of adjUstment should be made for an estimation of' social Benefits orsocial opportunity costs: For 'the social . case, one wishes to. know whatalternatives were foregone in a real sense-what society amid hive produced. A'moment's reflection will indicate the arbitrarinesi Of making an adjustmeni foe...unemployment for society when yOu try to estinute social Opportunity costs of

28

C.

4

education in, say, 1932, as opposed to 1944. (See Bowman, 1966, p. 431, I; .

Haveran ind Krutilla, 1967,1.) Fiscal and monetary techniques exist for the useof government to control the level of...employment. A given educationalinvestment should not be made to reflect the vapries of a price level, or income

nd emPkiiment policy whose social and political impetus may have nothing todo with the educationA policy. .

An additional issue is linked with the-unemployment problem. With the. .

.existelice of, unemployment, the question arises as to which it a better measureof pMductivity-wage ratel Or earnings? It id contended thin wage rates are less

.to reflect the vagaries of unemployment and, hence, do not penalizeechicational programs due to the effeCts of fikal and monetary poliCies whichare irrelevant to the purpoees of OdUcation. In short, wage rates pre a more stahleMeasure of the productivity of educational investment than are earnings in an

/environment . of cycliCal unernploYment. ,Yet, to the extent Mt. wages are/ it flexible downward (end this is only slightly), they, too, will reflect the 'impact of

/ j. unemployment. To the ektent that they are not fleiible downward, the'validity' f wager as measures of productivity is brought into 'question. Thus, the use of

earnings becomes MOIC meaningful as a measure of relative prothktivity in laboimarkets characterized by sticky wilsaund structural UnemploymentOn such.

. markets a person may bndergo continuing. cycles Of employment and .

unemployment because his productivity is less than the wage rate at vihich-he ishired. Once it becomes apparent to the employer that a man's prodectivity ,is

less than his wage. rate, lie is laid off. Manpowet retraining can serve to increase aperson's productivity up to the point where it equals the going Wage rates,When

. this retrained person is compared, against a comparable person in a controlgroup, no difference in wage rates may be discerned, but the trainee will.experience more stable. employment and higher earnings. It would be incorrectto argue in such a case, is do Earl D. Main (1968, III) and David Sewell (1969,III), that there are, no necessary benefits to the training program since wage rateshave not risen.

In line with this general problem Of -Utiemployment is the problem ofestimating - the coits of foregone Wages in a 'labor market where structuralunemployment exists.

/' If unemployMent is tompletely structural, there are no opportunity costsduring the, training process. The worker cannot perform the existing jobs .z.t allwithout the refraining: Likewise, once ha is retrained, the structural assumptionimpfies that the person's entire earnings be ascribed to the benefits of the trainingprogram. ,flowever. as an empirical matter, it is difficult to accept theseassumpliont Niihich ascribe no opportunity costs during training' and treats thetotal amount of earnings efter training as. a benefit, In' the first case, the

bribmeni 4.1444 the trabieehad no economic alternatives before him. In the/ extreme, this' implies that' his marginal revenue product (productivity of a

/ margreskUnit of labor times the price of the marginal unit of labor's output) iszer0. Neici.,(by counting Ole entire wage bill as a benefit; one is assuming that thetrainee's' marginal revenue product was zero at the time he mitered training andthe probability of untained workers filling that job slot wis zero.

. 38

I.

,

0

However, "the evidence in all these retraining studies is that trainees didforego earnings since members, of a cOntrol group had earnings' daring the

' training period. Members of the control group got jobs in the same areas IItrainees Thus, a zero probability of employment in these jobs Ify both thetrainee and the control group does not exist. Therefore, it is incorrect, to treatthe entire posttraining wage as 'a benefit, or opportunity costs during retrainingas zerd. The reason is that no market is ever completely dominated by structuralunemployment.

In short, person's expected earnings are almost never zero even at highlevels of cyclical unemployment. Also, it is almOst never the case that purestructural unemployment exists. Unemployment will usually be a mixture of thetwo typesa mixture which cannot be theoretically or empirically untangled.

But what if there are high levels of cyclical unemployment? If a workerbegins retraining, he is eliminated for a time from the labor market. Theprobability. that remaining unemployed workers may now become employed isat least the san.. and may how be higher, since'the supply of labor in the marketis reduced. If the probability that rethaining workers in the labor force will beemployed increases such that the zero likelihood of employment by the workerbeing retrained is exactly compensated for, then no social opportunity costsexist in terms of foregone earnings. There has simply been an income.redistribution. However, private opportunity costs do exist for the worker beingretrained since a positive expectation of employment now becomes zero duringthe training process,.This is the "vacuum effect" of Borus (1966, Ili).

Practical Issues and Suggestions in the Measurement of Costs and Benefits

Identification of Costs under Conditions of Matching Grants. TheVocational Education Act of 1963 'and its Amendments as well as suchmanpower acts as the Economic Opportunity Act set up-conditions whereby thereceipt of federal funds is contingent on the establishment of matching shares orpartial cost sharing by the grant recipient.

Two broad problems exist when one attempts to measure the social costsof the Neighborhood Youth Corps' (NYC) program, vocational education orsimilar social legislation involving federal-local cost sharing provisions. The firstdeals with the problem of measuring the social value of the sponsor share whenthe social program may be only partially 'funded by federal monies. The secondproblem deals with federal reimbursement of the sponsor for the use of certainsponsor facilities. These are common issues in any matching grant case.

The Sponsor Share. The federal eipenditure represents an actual outlayfor the federal government and is a cosi to the federal government. HOWevtfrom the standpoint of social economic cost, there is some question as to thevalidity and accuracy of the cost measure of the sponsor share. There are three,problem involved here.

1) First, if the sponsor, often a school district, has excess physicalcapacity, the use of which is restricted to the school district, the cost tothe sponsor for using this excess capacity is zero up to the limit of thedesigned capacity.

30

/

9

2) Second, if this restricted sponsor inpUt, such as a school building, isused to simultaneoutily produce both a sponsor output and an NYCoutput, the marginal cost of using that input fot the NYC Oroject is zero

.,,up to the limit of the designed capacity.3) Finally, even when there are no joint inputs or excess capacity,many of the inputs to the NYC program do nothave market pricesso- that the prices of theseinputs must be dstiniated 6r "shadow- priced."

The combined result of these three facttors is likely to be'an overstatementof true total costs to the combined government units (sponsor plus. federal) "aswell as an overstatement of total social cnsts. ShadOw pricing or price estimationand the joint cost problem are disCussed below.

. Federal Reimbursement for Sponsor Inputs. An issue separate from the 10percent sionsor share concerns the federal reimbursement of the iponsor for useof certain sponsor inputs, such as building space. Again, the three issues ofpossible excess capacity, joint putputs,. and shadow pricing arise.

' The problem is made more complex became cost to the federalgovernment is not necessarily the same as cost tuthe sponsorA rental paymentto a sponsor can be an overestimate of the true cost to the sponsor even thoughit might cost the federal government more to rent the same facilities 'on the openmarket. For instance, if a school system has excess classroom 'capacity, themarginal or extra ..cost f using' that excess capacity hi zero up to the limit ofdesigned capacity, as indicated above. If the federal govermnent doenot haveaccess ,to .that exceta capacity, it must pay 0 rent izi the market for comparablespace. Thul, the alternative cost to the federel government justifies the payrnentof a rent to the school system, even though the true marginal cost to the scheolsystem may be less than that rent. As long as the federal government paysschool system less or no more than it would have to pay in the market, then thispayment is ratiOnal from the standpoint of the federal. government. To theextent that the school system has excess capacity, it receives awindfall gain. Infact, since the federal government has not rented in the market but has rentedfrom the school dist rict, then, if excess capacity exisfs in the school district, someor part of the rental payment is a transfer payment and not a social comi Thus,it is reasoneole to assume that total federal costs may also overitate this portionof the social cost of the program. The same result would arise if 'the federalgovernment reimbursed a sponsor for the use of a joint input which was being

-employed to .produce sponsor output not associated with the educationalprograM in question as well ai to produce the program output itself.

Shadow hieing Even though the sponsor may be required by law tocontribute a certain percem uf the total cost of the program, the sponsor's-sharecan often be in the form of goods in. kind, whose market prices are thenestimated or "shadow-priced" in- negotiations between the local 'sponsor andfederal government officials. (See McKun in Chase, .1968, I.) The federalregulations are not very explicit about procedures for this shadow pricing. (SeeFederal Procurement Regulations, 1968, pp. 1501-1520, 1.) Thus,, considerable

1A trinsfeipayment is defined as a payment for which no compensating service hasbeen rendered. Its effect is to redistribute income.

49 31

.arbitrarintis can creep into the estiMate of the sponger's shi:e. And, file not atall inconeeivable tiat different shadOw prices could .be'attached le the same setof real resources being used in diffulmt projects across the nation aven thoughthe opportunity Cost in each location could, conceivably, be the same.

Table. 2 indiCafes the range of price estimates on iiiagarbom space whichcciared in" the establishment of the rusource value Of Ilse Sponsor's share of

Neighborhood Youih Corps (NYC) project operation in the greater LOB Angeles'area. The estimates ramp from $1.60 per day per classroCom tó S40 Per day perclassroom. The General Accounting :Office felt that a figure of $5,25 per day.per . classroom would be Most reuonable; based On a 2fiday month: (See ..COmpiroller General's Report, 1968, pp. 3941, 11E)

TABLE 2

DIFFERENTIAL SHADOW FMCS ESTIMATES OF fit VALUE OF. CLASSROOM SFACE, GREATER LOS ANGELES AREA

° . EdscatioralOrepnizatku

Rote Pee DayPee arum

Los Angeles Unified School District $10, $34, and $40Los Angeles County SchoOl Districts*:

Willowbrook School pistricts $6 and $9° Compton City School Districts $5 .

Compton Union High School District.

$1.60 -.

Arehdiocese of Los Angeles . $3.60 and $6U.S. General Accounting Office. $5.50.

Source: Comptroller General's Report to the Conte's, ReWevi Of the Connomnity ActionFromm In the Los Angeles Ares Under the'Econonk Opportunq Act,. OffiCe of

. Economia OpportUnity, II-162865. March 11; 1968, p. 40.

, Because of these differences in estimates of shadow/ prices, the readtingdifferences. in estimates of total attributed costs can be large. For instance, for'two NYC projects. in the Los Angeles alba, the Government Accounting OffiCe'sestiriaate. of total value of contributed Classroom space was $318,309 While the

. estimate of the Los Angelei Unified School District was $1,048,500, a differenceof $730,191. (See Comptroller General's Report, 1968, p. 41,

: It is not clear what ..the reMlutioi .of this incomistency .misItt be, sincethese school inputs have no comparable market inPuti upon which to get a more

. valid economic.MenUre Of COIL

Three possible treatments for Valuing this capital exist First, one Can-arpse that once the capital stock exists, especially the physical plant andbuildinp, it becoines specific to the educational . process and thus has noalternative use. In this case,' soeial Capital costs would be 'zero in the that mn,since no opportunity cost is iMolved.in their use for a cohort a/students whouse the capital after the decision wai made to create the school. This is k tenuous

32:

e

assumption, though, for it is easy to discOver alternative uses for such capital;"'stock. Thui, the value of the edecational physical plant is not zero in competinguses, but since it is hot a perfect su5stitute for these coMpeting uses,, the markethie of the competing.uses does not exactly reflect the opportunity cost ofusing the non-renovated physical. plant for educational purposes. Hone went tothe market io price the vahie of the non-renovated educational plant in terms ofits potential value as a hospital simply by observing what the value of a hospitalttas, the value would be overstated. Thus, the value .is not zero, bet itie less thanthe apparent value of alternatives since, without renovation, it 'is not a perfectsubstitute. And, even with renovation, such faCtors asolocation, Which cannot bechanged, continue tO exist and affect the degree of substitutability, thus forcingone to further adjust the implied opportunity costs.

ond, historical costs of be4d1ng construction and site acquisitioM canbe. use but these historical costs are essentially irrelevant since they have neinecessa bearing on the present opportunity costs involved in using the cspitalstock in estion. They °do not reveal the'current, economic value of-the capitalresource. urrent economic value could be less than, equal to, or greater thanhistorical ost.

z. Third, the use of replacement costs is a possibility in the attempt tomeasure capital ostr,. However, it is obiious that in many .cases it Would costmore to exactly replace a building than the building is currently .worth ineconomic terms. The use of replacement costs would over-value the capitalresource, given a rising price level and assuMinsg-no compensating technologicalchanyes in construction technique. .

In short, it is not Obvious what 'price* resulting among these three choicesshould be attached to the capital inputs to gee a momure of the Opportunitycosts. None of the above is correct in a pure theoretica1sense.

The &pita! Recovery 'Factor. Even if. the true economic value* of thecapital resources in use has been measured, the problem still remains as to theMeasurement of the rate at which the given capitil stock is used up over thecourse of the investment process 'when more than one 'cohort of studenisemploys the capitalstock. Two courses of action have, been suggested for tese.One is to attempt tO measure an imputed rent ind depreciation to. the capitalstock by making analogies with respeci to what arnount of rent (i.e., return onthe 'capital investment) the capital item would yield if it were being employed inthe private sector Of the economy. SOMe estimate Of depreciation is added tothis. BM such's. technique is subject to a great deal of arbitrariness anduncertainty.(See Corazzini, 1966, 11.) Legal rules for depreciation alloWances donot reflect economic reilities.:-

In order tO get a measure of the rental oppertunity cost, it is necessary togo, to the market place and attempt to identify capital resOurces which representalternatives to the resources employedin the educational process. This will allowone to determine thevalue of foregone Oternatives. But, again, .any impined rentbased. on market observations inn Most Rely. overstate the value of the capitalresOurees which are. already cOmmitted tn education. Thlts,..a great deal ofjudynent is involved in adjtisting the observed markee prices so that theV moreclosely reflect the true opportunity costs.

4233

it An alternative technique for estimating the rate of capital use lies inemploying the "capital recovery factor" (CRF). The application of thistechnique automatically accounts for both rent (interest) and depreciation.

The capital recovery factor is thai factor which " .. when multiplied bythe present value of capital costs, is the level (average) end-of-year annualamount over the life of the project necessary to pay interest mend recover the.capital costs in full" (Hirshleifer, et aL, 1960, I. Chapter VII).

The formals is as follows:Coi (l + 0 n

c e(1+011-.1

where c is the capital. recovery factor (annual level capital cost); Co is thepresent value of capital in use; i is the social opportunity cost rate Of capitdinvestMent fundt;. and n is the number of years Over which benefits (of thecapital in question) are -returned, that is, the, project life. Insome respects, thistechnique is \no less arbitrary than that which imputes rent, ad depreciation.Apart from the problem of establishing the present value of the capita in use,essentially arbitrary judgments must be mode with respect to the values of n andL In addition, the rite of capital use is projected as a conitant annual amount,whereis the true rate .of capital use is quite likely to vary over time. This,. Ofcourse; can create a bias in one's estimate of present value or rate of return.

. Joint Costs. In atidltion to the thadow pricing problem, it is clear thatmuch of the sponsor input into an educational program is really of the nature of*a joint cost or joint input. The school physical plant is a case in point. In suchsituations, the input is being used to produce simultaneously two or moreseparate outputs. For instance, space lila currently operating school may becontributed to house the staff of a newly :established,- federally . supportedprogram. The total cost Of o pe ra ting . the phydcal. plant of the school it Olenprorated among the varioul outputs, including the new program; yet, it may catno more to operate. the physical plant after the presence of the new.programthan it did before. .,

Two types of overestimation of costs can' enter' the .analysii. First, apositive price may be put on in-kind resources contributed by the sponsor as itsshare Of project costs when, in fact, the marlinsl cost of this resource use maybe zero. This results in an upward bia in the estimate of sponsor share cost.

..Second, when the federal govenThiinburses a sponsor for indirect costs,the reiource input in question may be a joint input, thus resulting in an upwardbias in the measure of economic costs of the program in question as distinctfrom accounting or financial costs of the federal -government. This lattersituation is not unlikely.

-The problem of joint costs affects the Walk-cost analyisis in twO ways.First; as is discussed below, there is no non-edgy-Ay measure of total cost andaway cost. Since we often will not know what judgments may have been madewhen the sponsors prorated joint costs, one has to.accept whatever upward biasis present in the total costs reported for the. Sponsor share a well as in thefederally reimbursed sponsor ,costi. This situation exists for the measure Ofmarginal cost also; however, the conceptual, Problem of proration is handleddifferently.

.34

Issues in Prorating Joint Costs. There are two points of view with respectto the problem of proration when marginal benefitcost cOmparisonsare being'made. The first advises against prorating. The second argues that proration ispossible. The first . point of view is supporied by such persons as Hitch andMcKean (1965, 1) and Enthoven (in Hitch and McKean, 1965, 1). They arguethat the existence of joint costs does not effect the determination of marginalcosts, and, since efficient investment decisions among two or More alternative .

programs are made on the basis of marginal costs, the 'pretence of joint cosispresents no balk problems for benefitcost analysis. Not only is joint costallocation necessarily arbitrary in nature, it is not needed, given the emphasis on.marginal costs. True marginal costs are zero. When joint costs occur and involvetwo or more programs. or 'outputs, the -total 'cost of the set of programs or

r, outputs can be measured. Then the combined total discounted benefits of the setof programs or Outputs should equal or exceed their combined total discountedcosts. BUt total average costs of each of the twO programs simply cannot bemeasured in any non-arbitrary economic sense. This is no real loss; though, sinceto repeat, investment decisions among tio or more coiniseting proyams are cor-reedy made on the baiis of marginal and not average cost and benefit comparisons.

Within very broad limits joint inputs are similar to what is, known ineconomic analysis as public good. Just as the benefits from a public good, suchal national defense, ae pervasive and need not be rationed Of allocated. on anindividual basis amoArg comumers (since one person's consuinption does nodiminish the cosmn4tlon of that' sime good by other consumers), so, tcio, ajoint input need no be allocated among the. outputs stemming from it becauseeach output can u the jointinpot without limiting the use of the input by allother outputs.. major problem here is that, except for such services asnational defense, is very difficult, to identify a pure public good. A seCondaryproblem ip that the production 'process should be operating below capacity kir :

. the statement abtve to hold.

The argument for proration .hal been' advance4 recently by K. L. Weil(1968, pp% 1342.1345, 1; also, Judy in 'Somers and pod,1969,111). Given ajoint input, X, such as die physical plant of a achodistrlct which, along.withgeneral outputs, produces the output of a federI1y supported program, theargument for proration goes 'as Estims the 'total demand and themarginal revenues for each of the outputs in question. The marginal rankles ofeach' of the outputs in question are then used to silicate the joint casts. The sumof the marginal revenues for the outputs in question must equal the price of thejoint input. Thus, the cost of the joint input is allocated to each output according toits relative share of marginal revenue. The allocation of costs in this example willdepend to a large' extent on "the conditions of demand for each of the outputs of

. the school district in -question. Thus, for an identical production techniqueoccurring in two markets withl1fferen1 demands for the outputs in question,different allocations of joint costs could occur.

The major 'problem with implementing 'this technique is that. it isextremely difficult to estimate demand Curves for goods and services especiallyquasi-public goods like educationand it is even more diffiCult to identify

4.1 35

Vo

specific points on these curves. Thus, the operational practicality of thetechnique is questionable, given the current state of the art.

The controversy over allocating joint costs has not yet been resolved, butthe author of the present cost-effectiveness study tends to agree that joint costsshould not be prorated, even though a pure joint input, like a pure public good,is difficult to find in-actual practice.

Finally,- to the extent that previously existing physical facilities are beingused,. these can be treated as "sunk" costs from society's standpoint. As such,their cost in use for the new program is zero if they have no alternative use. Inshort, in terns of clarity of the cost concept, the federal share is less ambiguOusof the two major cost componentsfederal and sponsor. And, the federal sharemay be closer representation of true social economic costsI 1:' e an sponsor re combined.

Cost Issues with Wage Payments in Manpower Programs. The NYCprogram, the Job Corps and cooperative vocational education are Of specialinterest to . this analysis due to the special problens created by theWate that isreceived by the program participant. Total costs Ihould be increaied to theextent that the-time of the program participant is undervalued by the war ratehe receives. That is, if, on the average, a student could earn more at 'some jobother than his job with the NYC or Job Corps,then the difference between the.two earnings would need to be added to total social costi to get a true measurepf total foregone opportunities. Likewise, if this petiole would earn leu on a job

. other than the job on the manpower program, the difference between the two isa transfer payment in favor of the varticipant. and should be subtracted from thetotal social cost measure.

In this reprd, transfer payments, which simply redistribute income amongpoups, ire not considered social costi. It is hi- the nature of a transfer paymentthat what is given up by one individual or social group is, in turn, gained by adifferent individual or social group, b that, ignoring the problem ofinterpersonal comparison* Utility or .the capacity to enjoy .economic goodsand services, there is no net cmi of welfare within society at a whole.

The use of the total war payment to the prograra participant as a costprobably overstatei truesocial .cost..lf-a program is desired, to provide incometo young persons who otherwise would be in:the libov force,butiould remaintotally or :partially- unemployet4 then some of the payment to them is a transferpayment. Indeed, if could be assumed that' the NYC, or the Job Corps program isnot fulfilling its function unless the typical participant would hats been earning.less without the manpower job. This difference oiler and above what theparticipant could have earned is not an opportunity cost to him; .

On the other hand, the snimporr program participant is making somecontribution to social benefits; since it is unlikely that his productivity is zero.Since he is contributing to social outiiit, this benefit should be added to theother benefiti of the program,'in order to balance the benefitaost ledger.

Problens similar' to the above miist in treating the war payment in amanpower program as a private opportunity cost.2 First, ec o n om ic theory

.

2i am indebted to Thoiess Ribich for einificetion of the issues discussed in thas ar4 .the previous section.

36

would argue that the costs of participating in . the program are the cosn offoregone leisure. The earnings of the participant represent his cost ofparticipeting in the program. However, the wage payment, in turn, is a benefitand must also be added to the benefit side of the ledger. Thus, if the participantincurred no other cost or benefit, his private benefit-cost ratio would be equal toone.

Anothe? problem arises if the program is providing earnings which theparticipant otherwise would not bye earned due to involuntary unemploymentor the receipt of a lower wage rate in the market. In this case, if the programearnings are equal to or greater than the earnings one could receive in themarket, then the foregone earnings resulting from participation in the programare zero or negative. Negative foregone earnings are a benefit which must beadded to the benefit side of the benefit.cost ledger.

Finally, there is the possibiliiy that some of the participants may earn lessin the program than they coUld have in !. the market. In- such a caie, privateopportunity costs are understated. However, the overall presumptiOn is thatprivate costs are overstated or, what amounts to the same thing, private benefitsare understated.

To determine -if the manpower program wage is an over- or under- estimateof the foregone earnings of the participant,, one could Aiwal to earningsmeasures for this age group reported in the census. However, at least two pointsought to be Made. First, these participants are different from those reported inthe census, since, apparently, some proportion of the program participantswouid either have been chronically unemployed Or not in the labor force in theabsence of the program. Second, coneeptually, a relatively large influx of youngpersons into the labor mrket should lower the earnings of this group relatiie tothe average eimings reported in the census. Thug 'use of Census data wouldresult in an upward. bias. Of Course, this problem is even more serious when oneintends to measure foregone earninp to- the primary and secondary :schoolpopulation in general. Nor does it help to indicate that child labor laws prohibitthe employment of much of this group, for such laws, having been passed, canbe repealed. Experiments hi cooperative vocational education of 14 and .15 yearolds are in progress even at this moment under an experimental programoperated by the Bureau of Labor Standatdi, entitled the Work Experience andCareer ExploratiOn .Progrim. Should it prove to .be a success, one couldanticipate increasing numbers of 14 and 15 year olds in the labor market.

. The Extrepolation Of Benefits. A major problem in benefit-cost analysis isthe determination of the length of thne which benefits extend into the fature aswell as the shape of this benefit stream. Average benefit -streams for varioustypes of educational benefits simply. are not known with- any .precision. Mostbenefit-cost Studies of nunpower prcigrams have .only i few, months to one ortwo years as a followup period after training.'The benefits to, vcicational andtechnical education t have been variously estimated as continuing for six to 10years before vanishing (Hu, et aL, 1969,11; and Eninger, 1965;11). The masersfor this are unclear. One possibility is that general and college preparatorygraduates acquire more .on-the-job training after leaving high school than dovocational graduates, though this has 'not yet been verified. Another possibflitV is

4E--

37

that the more general flexible nature of the general and college preparatoryeducation allows the Sampling of a group of jobs which, on the average, have agieiter earnings growth progression. Vocational graduates may enter their jobs at .wage rates closer to their peak lifetime earnings than do students-in competing.curriculuins. Finally, the option valuethe- degree to. which a given level ofeducation allows access to additional formal or on-the-job trainingmay behigher for the general and iollege preparatoricurriculum. All these are possibleanswers, but the reasons, for the converging earnings time profiles .5611 have notbeen fully investigated.

In the absence of any precision concerning earnings profdes, the bestcourse is to employ sensitivity analysis to estimate the range of effects underdifferent assumptions concerning_mningi profdes. Borus. and Tash (1970. 1)propose a usettur sensitivity matrix which allows for variations in the growth.ofthe. earnings profile at negative, zero and positive rates as well as benefit streamswhich last for a short,, medium and lifetime earning period. This ii ate best'solution to the problem at this point., But, 'it leaves one with a variety of,estimates, no one of which is clearly a measure of the true value.

.

The Probleni of the Control Group. A final iasue in the measurement ofcosts and benefiti deals with, the use of control groups. Ideally, the' controlgroup should code from the samepopulation as the experimental group. Datashould be collected for both groups .on such things as' sociodemographiccharacteristics, program inputi, and program outputs, both before, during, andafter the program 'treatment. As a practical matter, though, %hi is almost neverdone. The study of the in-school Neighborhood Youth Corps in Cincinnati byGerald Robin (1969,111) is an exception to this statement.

Moat studies of 'educational ,and manpoler programe are retrospective innature and hence -must generate a control group after the fact. 'Two generalapproaches hare been used. The first is simply to compare the experiencesprogram participants had before the program with experiences they. hid aftertheprogram. The second method-Is to attempt to develoP a comparable group of,persons who have never had the treatment to serve ai a basis for comparison.

With the before/after comparison, one is troubled by .the fact that changesother -than the treatment occur over time which' can affect the:measure ofprogram outcomes. By their very nature, it is difficult to control for -thesefactors. For instance, given that earnings aid employment ire a measure ofoutcome, one will get biased results If the pre-, during, and post-programmeasurement periods extend over a business cycle..On what bads .do you adjustwages and 'employment, up or down to reflect a full employnieht level ofemploymentand earnings for theixperimental group over the study period?

Before/afteiCOmpirisons can distort one's measures of costs and benefitsin other ways. Figure 5 shows thelefore/after earnings prOfile Of a person whowas structurally unemployed but who then took retraining. Ideally, what onewishes to measure as a benefit is the area Under:the curves bounded by P1, P3,P2, Ps.. This cannot be done, since, ante the penion takes training; the linesegnent PIP2 is no longer observable. Thus, a possible stiategy to rt a measure'.of earnings change is to compare earningt et the time of entrance to .theprogram, . t1, with* the 'profile' of earnings after the program'. As can be seen, .

.38

* FIGURE 5HYPOTHETICAL BEFORE/AFTER EARNINGS PROFILE OF STRUCIVRALLY

\ UNEMPLOYED MANPOWER TRAINEE

Earnings Profile,with Training

1_t2

t3

P4

Earningi'Profilewithout Training

-

Time

however, the result will be to measure negative benefits to the trainees. This isnot an unlikely result when you are dealing with workers in high wage industries,such ai West Virginia ccial miners who become technologically displaced andstructurally unemployed rather abruptly. Another, stratea would be to estimatethe slope of the line bap1, extrapolate it, to P3 and subtract this earningsprojection froM the line segment P1P3. This will result in positive benefits, but aconsiderable understatement. A third alternative would be to estimate the slopeof the curve b2P1t2 and subtraCt this difference from the curve P1P3P4. This willresult in an overstatement of benefits, since the earnings profile has an inflectionpoint (it changes direction of slope) at Point Ps. Thus, none of these alternativesis very satisfactory.,

Hardin and Bonn (1969, III) exp.imented with their Michigan retrainingdata and found the pins from retraining were $1,524 using a before/after

'method; when using a control group, the pins Were only $216 in the 365-dayperiod after traininga difference . by i factor of 'more than seven. Thus,depending where one begins his before/after estimation on the time .profile ofincome, serious under- Of over-estimates of benefiti can occur..

However, serious problems also exist in the absence of a true exPerimentalstudy model where the experimental and Control groups are selected beforetreatment from a similar population of subjects. Manpower training benefit-cost

. 4839

studies have variously used program dropouts, unemployed or underemployedregistrants at employment security offices, and eligible, persons who wereaccepted Into the program but who did not partkipate. Although mostsociodemographic characteristics can be controlled for, the persistent problem ofself-ielection into the program remains to bias results. No technique thul far hasbeen too successful in controlling for Such bias,. thou* the estimation of a

'..discriminani *function is a help. A discriminant function permits an estimate ofthe' probibbity*f a person who is included in the, control group being a member

--ef the experimental 'group. One general statistical estiMatieon technique for thediscriminant . function. is :known as probit analysis. (Sa Laumann, 1965,Some rs,*.$tronwdorfer, 1970,111.)

V&. tional and technical education presents a particularly difficultproblem when one seek.; to develop a meaningful control group. Generally,participants-kr-vocational or technical programs' are compared against those inthe general oi callege preparatory cunicUlum. However, there exists afundamental problem in that all these groups do not come from the samepopulation of students. It can be expected that each of these persons will place adifferent weight on earnings, job status, the value of additional collegeeducation, and other factors associated% with 'the multiple outcomes ofeducation. Generally,, these relative weight!. are nOt known. Thu's, foi,instance, if

' wage' rates or earnings: are used is i simpls index of program benefits, a bias canresult. If vocational 'graduates place lesi emphasis on job status and moreemphasis on earnings than from the standpoint of,' say, college preparatorystudents whose emphasis may be the reverse, benefits to vocational educationmay be overestimated, Due to the fact that the different types of students areattempting to maximize different sets of satisfactions (or utility functions),there is, as yet, an unresolvable problem with the Use of these types ofcurriculums as control or comparisoh groups with vocational or technicalgraduates.

. Sammy; All costs, regardless of their institutional, form or *oblong; ofmeasurement, are opportwaitY att.:. As such, they repreient the value of thenut best alternative to which funds could be put were they not expended onvocational or manpower training.

Benefits. are the opposite of costs, and rePreaent opportunities gained as aresult of undertaking a partkular activity. .

A variety of conceptual end measurement problems faces the -analysis ofeducational investment in human beings. These are the following:

1) Earnings maximization versus utgity meximization;Complementarity in production and consumption;

3) Joint costs of production, consumption and investment;4) Non-market production and consumption; .

5) .Impact of education on values and preferences;:Risk; . . .

7) Complementarity, .subs6tutability, and inseparability of huminskills;

40 .

7"--

49

1

,

8) External effects;9) Income redisizibution effects;10) The influence of unemployment on the determination of costs and

benefits; and11) The problem of the control group.

Briefly, these. problemeaffect the analysis as follOws:Benefit-cost analylis slould measure the increase in welfare or utility due

to an educational program. No direct Measures of utility are poesible. Earnings. .

and other measures Of program effect therefore become indices of utility.bitalways IlleallUre it imperfectly. . .

. Education and training create simultaneous production-and consumptionbenefits. The latter are extiemely 'difficult to measure and have not been as yet.The result is to underestimate the benefits to education.

Many educational inpUts, such as a school building, simultaneously Createmore than one output. Such inputs are known es joint inputs": An unresolveddebate, haMpered by severe measurement problems, exists over the advisabilityof proration of these input costs. Hitch, McKean and Enthoven argue that thetrue igrginal cost in a joint input situation is zero for each of the outputsproduced. Each output uses the input without detracting from the ability ofother outputs to employ it, at least up to the &signed capacity of the input..Thus, the Marginal coet is Zero. Weil argues that joint costs should be prorated inaccordance with the relative degree to which the outputs produced by them addto total. benefits. But this requires the estimation of demand cUrves for thevarious outputs in question. Statistical estimation of, demand is very difficult todo, especially for quasi-public goods like education. ,

The failure to measure non-market production and consumption, whichmay be a very large component of benefit to education and training, mayseriously understate total benefits.

Changes in values, tastes' or preferences ducto the act of education ortraining alter the structure of relative pricei and hence the measure of costs andbenefits to any educational investment. Costs may no longer be measured on thesame basis at benefits, since the structure of relative prices before and during theiniestim.t will be different after the investment..This is a crucial point, since.. one, of the purposes of investment in education is to change noneconomicbehavior.

Risk and the finiteness Of human life; linked with the fact that usuallyonly one particular type of investment can be undertaken by a human at.a time(that is, training as a doctor precludes simultaneous training as a butcher), causedifficult policy' problems concerning.: the, type. of .educatioa to beprovided-specific or reneraland the realtive lengths of' time this educationshould continue. Partial solutionto the high ochool dropout problem hinges ofiappropriate analysis of these factors.

!

Investments in human beIngs Ire complementary and cannothe separatedfrom the human agent hi- whirl:they are 'embodied. Since a person undertaket aseries of investments over, time, it is difficult to empirically 'sort out the net.effect of any given educational immstment on human performance and welfare.Apparently, for instance, the more -education one gets, the greater is his

41

_

opportunity to pin on-the-job training. Any measUrement of benefits to theoriginal education over time must adjust for the existence of this on-the-jobtraining. Also, to what extent should the original training and the on-the-jobtraining which it enables be considered the same or separate investments?

External effects are either costs or benefits created by an economic actwhich affect the EC0110111iC behavior of second or, third parties but which are nottaken into economic consideration by the person or persons originally causingthem. Such effects are difficult to account for since they often do not occur inan institutional setting *here prices can be sniped directly to them.

Education is also asserted to provide considerable external benefits, but, atleast at the post-secondary level, this has recently been strongly challenged byWeisbrod and Hansen, among others. To the extent that such external effects dooccur, hoWever, they should be measured. Few studies have attempted even apartial measure to date. What is the exact value to a given perion dueto the factthat all 'other persons in the 'society are literate? Ifow can this effect bemeasured? .

Changes in the distribution of income cause problems similar to *thosedisciiiiiiiiinder changes in values,- tastes and prefezences. In addition; it is notnecessarily the case that a dollar of income taken from one person and given toanother luts no net effect on total social welfare. It is .quite possible that eachvalues the utility of the dollar gained or lost differently. This is especially aserious problem in evaluating the benefits of manpower training Whezt, given the .

-existence of cyclical unemployment, the placement Of a trainee in a job may,mean a non-trainee has been displaced from it. The pin of the one man iscancelled in whole or part by the !Ms of the other. Also, educational programs .

financed by taies imply the existence of transfer payments from the non-schoolto the school population. How does one evaluate the impact of these transfers onsocial welfare and, hence, on social costs? No benefit-cost study in secondary'orpost-secondary vocatiOnal *cation or manpower training has attempted to 'quantify these effects.

Unemployment is defined as either cyclical or fractional. Cyclicalunemployment is due to a lack of demand hi the economy. Structuralunemployment is simply long-duration fractional unemployment, where, for anynumber of . reasons, the available jobs in a labor market do not . match theavailable skills.

The existence of cyclical unemployMent creates problems of measurementin both Costs and benefrts. These are described as "vacuum" and "displacement"effects by Borus and represent cost of benefit biases due to changes in theincome' distribution brought about thiough the dinamics of Manpower trainingor similar' programs.lhe existence of structural unemployment does not causethese income transfers to occur.

The calculation of social and governmental benefits under conditionswhere federal monies are matched by local goveniment contributidna causessevere problems of measurenent of the value of social costs. Often the sponsorinputs are donated in-kind :ind their prices must be estimated or shidow-prieed.

. The local inputs may repres-lit the use. of existing excess capacity. Which, for avariety of institutional reasons, may have no noneducational purpose. Thus, thetrue social cost of the use of these inputs may be very low or even zero. Finally,

42 _ 51

.

. 0

the lotai sponsor, school or district,may be employing its resources to producejointly two programs, one oPiahich is partially lupported by federal monies' andone which is not. With 'such &joint input, there is no unambiguous theoreticalguidance on how thii joint input should be prorated. While the argument isunresolved, it is almost certain that:prorating the costs of a school's commonlyused physical plant on the basis of square feet ri student or some otherarbitrary rule lends bias to the estimate of average and total cost.

Before/after comparisons If program effects are inferior te those based onthe use of a properly selected, control voup. The choice of a control group touse depends On the, purpose of the analysis. Different control group comparisonstell different thinv about a program. For some purposeilt is desirable to useonly academic -or general students as a control for yOcational students. For othercomparison, one May wish to *use the student body of' a comprehensive high.school. One will get different results for a manpower trainfng program if hedesignates dropouts as the control group as diitinct from a random sample of theunemployed or those eligible who did not enter a program.

It is not commonly understood that observations on variables for thecontrol and experimental groups should be taken both before, .during and after.the training process. It would be too harsh to suagest that the use of a controlgroup is not fully appreciated, but, certainly .once the: control group is.designated, the sample units froin.,it should be selected randomly. This fact isOften not appreciated, thOugh departures from random selection are, of course,necessary if one wishes to pick a judgment sample for a very specific purpose.Under such conditiOns, however, the narrow purpose's of such a methodologshould clearly be recognized.

e

.

A REVIEW AND SYNTHESIS OF BENEFIT-COSTAND COSTEFFEC11VENESS STUDIES

7..

There are hundreds of studies, papers, reports and monographs 'whichattempt to evaluate ,;vocitional, technical., and manpower 'training- prograni.Some insights, can be pined from most of these: However,telatively few studiesexist which deal...ace:shrilly with the investMent aspects of such educationalactivity...This chapter provides a'. summary of- nost-effectiveness analysis ofseCOndary . and postiecondary. vocational-techriicil, education as . well as an'analysis 'Of two-year junior college education. Where similarities among the'populations served wasraht it, the aboye three .. types of erfircation- will be -compared :in economic inveitment . tem-. with selected Federal government .

manpower props= such as the MDTA; JOBS, Job Corpi or the NYC. Theobjectives of the various maripoWer programs,. at well as the poilulation each iii4Pintended to serve, ire displayed in. Table 3. Characteristias .of the 'actual -.

. populations served are displayedin Table.4. . .

. When. treating 'these various programs as 'substitutes for each other, one,. should note that4l4uperficial sinilarltles amOna the ponplaiions they serve

obscure some' vei§ significant dissimilarities. For instance; those served by theMDTA and those served. by vocptional education 'will differ in tern' of age,family life cycle, the opportunity costs (4regone war) they bear While being

. trained, quality and quantity of prior edUation, and o significantcharacteristics. These two propams 'carort be thought of bstitutes,for each ether. For another contrast NYC is beim ref en..16 and 17'year-old high school dropouts: The groupie clearly diff t fromihe.personsgenerally served .by the. institUtionikMDTA. The,MDTA mirror maynot be high school droPouts, and Over 85 percentof them 19 years d or.older. Finally; the in-school NYC is a rather imperfect mbetitute rative,vocational education-some would argue it is no substitute at all. .

Table 4 shows how the clientele of the varioir programs diffei in broadterm. Vocational education is a program which largely ..serves-white youths,both Males and females. The MDTA programa serve mainly whites and are adult

. .

programs.-The JOB Corps,. and JOBS program servemainly black Males with theJOB Corps concentrating on youths while JOBS concentrates on adults. TheConCentrated Employment .Program (cEr) mainly concentrates on black adults,while Operation Mainstream serves mainly white adults': The NYC Programa areevenly divided between whites and blacks and concMtrite on youths.

Thus, one should take care in awning that all 'them-programa are closesubstitutes for each other: Data on other socioeconomic of sociedemographic

odds/ 1111-

TABLE 3OBJECTIVES AND POPULATION TO BE SERVED FOR

, SELECTED FEDERAL MANPOWER PROGRAMS

rProgram andDate Started

Objectives! and. Services

. PopulationServed2

Concentrated Employ- Coordinated program ofment Program (CEP) manpower and suppor7May, 1967 tive services

Hardcore, unemployedyouths and adults in se- .

lected areas where they areconcentrated

Job Corps Resideatial program of- Low income, disadvan-January, 1965 intensive education, skill taged youth 16 to 21 years

training and related of ageservices

Job opportunities in the Uses private industry to Hardcore unemployed 18Business Sector (JOBS) hire, train, retain and up- years of age and overMarch, 1968 grade the program popu-, -

*. lation .

MDTA Institutional and Provides occupational Unemployed and underem-On-the-Job Training, training or retraining employed persona 16 yearsAugust, 1962 a claaroom setting or of age and over, two-thirds .

instruction combined of which must be disad-With supervised work at vantagedthe lob site under con-tracts with private andpublic employers

Neighborhood Youth Job preparation through DiSadvantaged youth ofCorps (NYC) Out-of- paid work experierre high school age (14-21).School, January, 1965 with remedial services Nevidesign of out-of-school

NYC limits population to16-17-Year-old dropout&

Vocational Education Full-or part-time voca- Youth or adults, in or1917 tiaital training, primatily out of public schools.

m a classroom setting to New emphasis on poorreduce the flow of un- .and disadvantaged*skilled or ill-preparedyouth into the labormarket

Notes:1 Fot 1 of these programs, the major objective is to upgrade or provide occupational skills

tha ill be of value in the labor market. Each program cites additional objectives, someof ch are economic in nature and some of which are psychological or social in nature;

0 but the major goal of each of these programs is to enhance the earnings and employmentof the group served. .

2 "Disadvantaged" means poor, not having suitible employnient and either (a) a scholdropout (b) a member of a minority,'(c) under 22 years of age, (d) 45 years of ase 0Eover, or (e) handicapped.

54

Source: . . Department of Labor, Menpower Repat of the President, karch,1970,Appendix A: "Guide to Federally Assisted Manpwer Training f:it.d SupportPrograms."

characteristics would likely show even greater dissimilarities. Thus, thepopulations served do differ. a great deal. For this reason it is not likely that thevograms are close substitutes since, although teaching technology may besini1ar that is, the same concepts, etc. are needed to teach machine shop inMDTA institutional as in secondary voc;tional jeducation, other institutionalfaCtors surrounding the teaChing process and .Aich affect the process, maydiffer and may also be needed to serve tW: particular groups.

Likewise, one should not strongly generalize that manpower training servesas a '"remedial" program for the "mistakes" of secondary vocationalcomprehensive education. The great disparity in the age/ distributions as well asrace indicates that significant cultural, social and econoriric factors can intervenebetween the time one leaves high sChool and the five, 10 or 15 years or so laterwhen he may take manpower training. Such changes may occur and do occurthat no educational planner in high school it even likely to be able to anticipate.

Next, one should be aware that the benefit-cost studies summarized herere por t only monetary economic costs and benefits or reductions inunemPloyment. They do not account for nonmonetary economic costs Ofbenefits, or benefits such as job satisfaction. However, for a fairly narrowinvestment analysis the monetary measures are considered to subsume the majorportion of all costs and benefits.

It is also important to note that each of theSe studies, 'uses .differentmethOdologies; for instance, 'each contiolsfor ,different sociodemographic

°variables and some studies use different control groups. Several studies use nocontrol group at all but rely on before/after comparisons. The tables reveal someof these d Terences in methodologies. While different concepts of cost or benefitcan be and were adjusted for, the basic methodologies underlying the studiescannot be changed. Hence, this analysis is a summary statement of what hasbeen done but, to some extent, the studies are not comparable, even whenpopulations served are the same.

The remainder of the chapter is organized as follows. Fint, the majorbenefit-cost studies will be analyzed in broad terms.. Secondaryvocational-technical education will be compared in investment terns withselec4sd secondary education alternatives, such as graduation from acomprehensive high school. Then, post-secondary vocational technicat educationand-two-year junior collee:c will be analyzed. Finally, the manpower programswill be analyzed in the same terms, starting with MDTA, then Job Corps, NYC,JOBS, and CEP.

Second, the impact of the programs on various sock lemographic groupsare considered. Only limited benefit-cost analysis is pre4nted he;e , though someof the data are previously unpublished.

Third, the earnings and employment benefits of program areas and skillsare presented. Most of the analysis in these three sections relys heavily onmultiple regression analysis to estimate net program effects.

47

.tn:r

wro

rsrq

tycv

l.cr-

A.,.

..

TA

BL

E 4

STU

DE

NT

CH

AR

AC

TE

RIS

I1C

S B

Y R

AC

E, S

EX

, AG

E A

ND

PR

OG

RA

MIN

PE

RC

EN

T, 1

969*

**

,

-

_

Voc

atio

naE

duca

tion

MD

TA

Inst

itutio

nal

MD

TA

On-

the-

Jolt

Job

Cor

psJO

BS

CE

PW

IN

..

Ope

ratio

nM

ains

trea

m:

: NY

C-

IA 0

/S

,

Whi

te79

.5**

.59

.266

.8:

26.2

12.7

.28

.056

.061

.646

.3 4

8.2

Bla

ck.

13.6

36.0

30.3

60.8

77_5

65.0

,40

.024

.047

.4 4

7.5

Mal

e45

.055

.6-

65.1

74.0

70.8

58.0

40.0

'

70.8

53.4

46.

0

Fem

ale

55.0

444

34.9

26.0

29.2

., 4

2.0

60.0

29.2

46.6

54.

0

Und

er 1

9-

Yea

rs o

f A

ge94

.31

12.5

-11

.191

.117

.338

.4°

16.0

-2.

096

.1 8

0.6

Not

es:

Und

er 2

2 ye

ars-

*" C

onta

in; b

oth

whi

tes

and

Ori

enta

ls.'

"* S

ome

of th

ese

data

ref

er to

fin

al y

ear

whi

le o

ther

' ref

ei to

par

ticul

ar ti

me

peri

ods

in 1

969.

Sour

ce:

Voc

atio

nal E

doC

atio

n: C

hene

cten

stks

of

Tet

er&

r a

nd S

tude

nt:,

1969

0E-1

0073

, Nat

iona

l Cen

ter

for

Edu

catio

nal S

tatis

tics,

U.S

.O

ffic

e of

Edu

catio

n, W

ashi

ngto

n, D

.C.:

U.S

. Gov

erni

nest

Pri

ntia

g O

ffic

e, D

ecem

ber,

197

0; a

nd-

Man

pom

er R

epor

t of

the

heri

dent

, U.S

. Dep

artm

ent o

f L

abor

, Was

hing

ton,

D.C

.: U

.S.G

Ove

rnm

ent P

rint

ing

Off

ice,

197

1.

7

Summery of Hewitt Broad.program Effects

Secondary VocationalTechnical Education. Among the secondarycurricula, it is most reasonable to evaluate the secondary vocational-technical'curriculum in economic terms. However, there is a problem in choosing theapproprieo control or coMparison group nainst which 'to judge the neteconomic performance of 'this curriculum. Table 5 displays the benefit-costanalysis of vocational-technical education with respect to two different controlgroups. The first control group is the combined curricula of the comprehensivehigh school. This control group would include the- academic or collegepreparatoy curriculum 'as well as the general And vocational-comprehensivecurriculum. The latter is 'essentially a general curriculum wedded with a group of .'vocational courses, none of which is intensive enough to give the student anyspecific highly marketable skills. The other comparison is against the academicItor college preparatory curriculum alone. Usine either of these brpail controlgroups creates problem of comparison, since the objectives of the variouscurricula are somewhat different, 'Additionally; the sociodemopaphiccharacteristics and personal objectives and goals of the groups partaking of thevarious curricula are different.; For instance as noted earlier, thevocational-technical graduate, as compared to the academic graduate, May pUt aheavier weight on earnings than on the nonmonetary gratification to be hadfrom a job. Thus, when the two types :4 graduates are compared; the benefits tothe vocational-technical graduate vis4-vis the academic graduate may beoverstated since having immediate money inCome insy be less important to the Tacademic graduate than to the vocational-technical graduate. The fact that thereis a higher dropout rate from vocational than from academic proems may bean expression of the fact thatiWenta who choose vocational courses may makejudgments over 'a shorter time horizon and weighi immediate financial rewardmore heavily than do academic curriculum students. This issue will beinvestigated at greater length in subsequent discussion.

The data in Table 5 display both internal rates Of return-the profit, rateof 'the program-and net present capital'values-the present value of 'benefitsminus the present value of costs. Three of the seven studies listed are nationwidein scope, while three relate to spciifid locales. Major reliance for policy decisionsshould be placed on the three studies which are nationwide in the scope of theirsample.

Average Costs and Benefits. The Fernbach and Somers study (1970, II)indicates that _vocational-technical graduates earn an average of S667 more peryear than do secondary academic graduates-Total social costs, including directoperating costs, capital costs, and foregone earnhip, amount to an average ofp.bout $720 per year. Thus, the average rate of return to vocational-technicaleducation for the Fembach-Somers sample of vocational-technical graduates isapproximately 21.4 percent fer the investment over a projected 10-year period.tsi 49.

TA

BL

E S

ctt

=.0

CO

MP

AR

AT

IVE

AN

ALY

SIS

OF

Ctii

T A

ND

BE

NE

FIT

ES

TIM

AT

ION

SO

F S

ELE

CT

ED

ST

UD

IES

OF

SE

CO

ND

AR

Y V

OC

AT

ION

AL:

TE

CH

NIC

AL

ED

UC

AT

ION

Nam

e of

Stu

dylim

eP

erio

dof

Sto

r I

LOC

Oof

Seed

y.E

xpes

inie

ntti

Gro

upC

ante

dG

rum

p

cidi

yar

Sse

meg

kpho

f 6

Dor

ados

of

Tra

iling

imY

ears

Dur

atio

n of

*unf

its it

sW

art.

Rat

e of

Ret

orts

*sce

nt)

Pres

ent V

alue

In D

olla

rs.

Mer

ging

Ave

rage

8M

orgi

nel

Ave

nge.

5%10

%

1. H

u, L

ei, a

ndSt

rom

sdor

fer

.

a1

n -1

b:21

. 2.

1959

-66

Bal

timor

e,Ph

ilade

lphi

aD

etro

itP

hila

delp

hia

Det

roit

Phila

delp

hia

Det

roit

Voc

atio

nal-

Tec

hnic

al?

Com

pre-

hens

ive7

464

386

485

403

343

643

343

643

3 3 3 3

6 6 6 6

93 33.6 8.2

31.8

240

1776 183

1772

Neg

1152

Neg

1102

2. f

ernb

ach

and

Som

ers

b.

1964

-69

Nat

ionw

ide

Voc

atio

nal-

Tec

hnic

alSe

cond

ary

Aca

dem

ic.

5925

96;

6153

711;

715;

738

.

667

667 .

3 3

10 10

. 25-

9

21.4

2811

2484

.

1583

.

1283

7). P

roje

ct T

AL

EN

T(M

ales

onl

y)

a. b.

1958

-65

Nat

ionw

ide

Voc

atio

nal-

Tec

hnic

alC

olle

gePr

epar

e-to

ry..

465;

483

;48

5356

0; 5

74;

585

c

.37

5

.375

3 3

10 19

17.7

13.8

1200

943

542

307

4. C

oraz

zini

(Mal

es o

nly)

3- I

- 2.

b. 1

.. 2.

'196

3-64

,

Wor

ches

ter,

Mas

sach

u-se

tts _

Voc

atio

nal-

Tec

hnic

al .

Com

m-

hens

ive

512

618

964

1129

312

3 12

312

312

2 ,

2 2 2

_ 10.

10 10 10

23.1 7.4

17.9

4.1

1059

219

862

Neg

596

Neg 41

2N

eg

5. &

tinge

s (M

ales

0111

019

56-6

5;

1950

-64

Nat

ionw

ide

Voc

atio

nal-

Tec

hnic

alA

cade

mic

.

.

a.'

.43

5;44

7;41

241

23

1021

.216

2478

3

4653

'

b..

.52

2;54

7;41

241

23

1018

.313

5863

1-

569

,19

60.6

4.

.

a.48

5;49

157

73

1027

.425

12'1

443

4963

b.58

2; 5

89;

.57

73

1022

.8 -

2245

1199

595

.

6.K

aufm

an a

ndL

ewis

1958

-65

Thr

ee c

ities

in P

enns

yl-

vani

a

Voc

atio

nal.-

Tec

hnic

alC

ombi

ned

Aca

dem

ican

dG

ener

al

1960

548;

553;

837

310

34.5

4278

2735

5623

.919

62`

-56

2567

;61

13

1025

.226

5515

4957

63. 9

'7.T

auss

ig19

62-6

5N

ew Y

ork

Voc

atio

nal-

Com

bine

d.

-

'C

ityT

echn

ical

Aca

dem

ican

d.

Gen

eral

a.38

9;41

2;24

0id

310

6.8

255

Neg

4171

201

13

10N

esN

es

b.48

4509

;24

010

310

.

4.6

Neg

Neg

5191

201

,3

10N

esN

es

Not

es:

1 T

ime

peri

od o

f st

udy

inch

sdes

the

trai

ning

per

iod

as w

ell a

s th

e av

aila

bk ti

me

for

follo

w-u

p at

the

time

the

data

wer

e ga

ther

ed.

2(a

) Si

gnif

ies

curr

ent o

pera

ting

cost

s: (

b) s

igni

fies

tota

l res

ourc

e co

sts

to s

ocie

ty in

clud

ing

curr

ent o

pera

ting

cost

s, c

apita

lcos

ts a

ndop

port

unity

cos

ts w

here

app

licab

le. T

his

is th

e ca

se f

or a

ll th

e fo

llow

ing

sim

ilar

tabl

es.

3 E

ach

cost

fig

ure

appl

ies

M a

dif

fere

nt y

ear

in th

e re

leva

nt th

ree-

year

trai

ning

Per

iod.

I

'

"-

The

six

-yea

r be

nefi

t for

Hu,

et a

l., is

bas

edon

the

estim

ated

leng

thlo

f tim

e be

nefi

tspe

rsis

ted

in th

e th

ree-

city

stu

dy. T

he 1

0-ye

ar

.

4'

bene

fit P

erio

d it

bate

d nn

Eni

nger

, Pro

cess

and

Pro

duct

.. .

: T

he P

rodu

ct, o

p. c

it.

.5C

osts

per

yea

r ar

e re

lativ

ely

law

sinc

e op

port

unity

cos

ts (

fore

gone

,w

ages

) ar

e as

sum

ed to

be.

the

sam

e be

twee

nse

cond

ary

voca

tiona

l stu

dent

s an

d co

mpr

ehen

sive

stud

ents

. Als

o, to

the

exte

nt th

at th

ey e

xist

, the

seop

port

unity

Cos

ts to

soc

iety

are

assu

med

to b

e qu

ite lo

w s

ince

the

infl

uxof

all

high

sch

ool s

tude

nts

into

the

labo

r m

arke

tat

onc

e w

ould

dep

ress

con

side

rabl

yan

alre

ady

low

leve

l of

earn

ings

for

this

age

grou

p..

6A

ll be

nefi

ts a

re b

efor

e-ta

X e

arni

ngs

and

repr

esen

t a s

ocia

l ben

efit,

that

is, a

n in

crea

se in

adde

d to

the

gros

s na

tiona

l pro

duct

.7

- N

eith

er o

fth

e3e

two

grou

ps h

ad a

nypo

st-s

econ

dary

or

*dor

col

lege

or

colle

ge e

duca

tion

in th

esi

x-ye

ar f

ollo

wup

per

iod_

8 T

he c

osts

are

true

ave

rage

cos

ts. T

he b

enef

its a

redi

ffer

ence

s be

twee

n av

erag

es o

f th

e tw

o co

mpa

riso

ngr

oups

. Thu

s, th

e es

timat

edra

tes

show

n he

re u

nder

stat

e th

e ac

tual

aver

age

rate

of

retu

rn.

I-

.

.9

Exa

ctco

mpo

nent

s of

thes

e co

st f

ugur

es a

re n

ot k

now

n. T

hey

incl

ude

at le

ast c

urre

nt o

pera

ting

cost

s. N

ote

that

they

are

gen

eral

lyco

nsis

tent

with

the

othe

r fi

gure

s on

the

tabl

e. T

hese

fig

ures

are

def

late

d, a

ssim

ing

cost

s w

ere

8553

/ann

um in

195

9, th

eba

seye

ar.

10 F

or m

ales

only

, in

trai

ning

rel

ated

jobs

, ass

umin

ga

12 c

ent p

ert h

our

gain

and

a 2

,000

hou

rw

orki

ng y

ear.

Una

djus

ted

for

any

soci

odem

ogra

phic

dif

fere

nces

exc

ept s

ex.

-11

Fem

ales

only

, in

trai

ning

rel

ated

jobs

, una

djus

ted

for an

y so

ciod

nogr

aphi

c di

ffer

ence

s ex

cept

sex

..

12 E

ach

cost

fig

ure

is f

or a

sep

irat

eye

ar, d

efla

ted

from

the

1964

-65

IXiie

yea

r. C

osts

are

for

mal

esan

d fe

mal

es c

ombi

ned.

'Coi

ts a

redi

ffer

ence

s be

twee

n tw

oav

erag

es.

'

I'

Sour

ces:

L T

eh-w

ei H

u, e

t at,

A C

ost

Eff

ectiv

enes

s St

udy

of V

ocat

iona

l Edu

catio

n: A

Com

pari

son

of V

ocat

iona

l and

Non

roca

tiona

lE

duca

tion

in S

eam

day

Scho

ols,

Uni

vers

ity P

ark,

Pen

nsyl

vani

a, M

arch

, 196

9.°.

2. B

enef

it D

ata

are

from

: Sus

an F

ernb

ach

and

Ger

ald

Gm

ers,

An

Ana

lysi

s of

the

Eco

nom

ic B

enef

its o

f V

ocat

iona

l Edu

catio

n at

.th

e Se

cond

ary,

Pos

t Sec

onda

ry, e

nd J

wO

or11

,hr-

limita

ry R

epor

t, M

adis

on, W

isC

onsi

n. M

ay, 1

9711

Cos

t dat

a ar

e fr

om: -

Am

eric

an I

nstit

utes

for

Res

earc

h, A

n A

nal

. of

Cos

t and

Per

form

atcw

Fac

tors

for

the

Ope

ratio

i and

Adm

inis

trat

ion

of

prt._

,

Vor

atio

nal S

choo

ls f

or S

econ

dary

Prog

ram

,tts

burg

h, P

enn

ania

, may

, 196

7.-

3. B

enzi

n da

ta a

re f

rom

: U.S

. Off

ice

of E

duka

n, O

ffic

e of

.... P

lann

ing

and

Eva

luat

ion,

unp

ublis

hed

Proj

ect T

AL

EN

Tda

ta,

5-ye

ar f

ollo

wup

info

rmat

ion

on h

igh

scho

olsr

tes

of 1

960

led

in H

owar

d V

ince

nt, "

An

Ana

lysi

s of

Voc

atio

nal E

duca

tion

inO

ur S

econ

dary

Sch

ools

," J

uly,

IS

9(r

evis

ed),

. Cos

t dlia

are

fro

m: A

mer

ican

Ins

titut

es f

or R

esea

rch,

An

Ana

lysi

s of

Cos

t. .

op. c

it.. M

ay, 1

967

t.

4. A

rthu

r I.

Cor

azzi

ni, "

l'he

Dec

isio

nto

Inv

est i

ntio

nal E

duca

tion:

An

Ana

lysi

sB

enef

its,"

in T

he J

ourn

al o

f H

uman

Res

ourc

es,S

uppl

emen

t: V

ocat

iona

l Edu

catio

n,V

olJl

l19

68.1

5. B

enef

it da

ta a

re f

rom

: Max

U.

Etti

nger

, The

Prt

4ess

and

Prf

teac

to

f T

and

1 H

igh

Scho

ol L

evel

Voc

atio

nal E

dict

:Wm

i n

the

Uni

ted

Stat

es T

he P

rodu

ct, A

mer

ican

Inst

ituR

esea

rch,

Pitt

sbur

gh, P

enns

ylva

nia,

Sep

tem

ber,

196

5. C

osts

data

are

fro

m:

Am

eric

an I

nstit

ute

for

Res

earc

h, A

nA

naly

sis

of C

osts

. .

-, o

p. F

iLi M

ay, 1

967.

6. C

oat a

nd b

enef

it da

ta a

re f

rom

Jac

obJ.

Kau

fman

and

Mor

gan

V. L

ewis

, The

Pote

ntia

l of

Voc

atio

nal E

deun

rion

: Obs

erva

tions

and

Con

clus

ions

, Uni

vers

ity P

ark,

Pen

nsyl

vani

a,M

ay, 1

968-

i ii

7. C

ost a

nd b

enef

it da

taar

e fr

om: M

icha

el I

C. T

ausr

ig, "

An

Eco

n m

icA

naly

sis

of V

ocat

iona

l Edu

catio

n in

New

Yor

kC

ity,"

Jou

rnal

of H

uman

Rea

moc

es, S

uppk

mem

t,V

ocat

iona

l Edu

catio

n, V

ol I

II, 1

968.

-

,

If the average social cost rate of capital is 10 percent, then vocational technicaleducation yields a relatively high rate of return .3

Mnginal Costs and Benefits. Given their obvious qualifications the abovestudies indicate that the average costs of vocational-technical education are morethan covered by the average benefits of the program. Thus, in absolute terns,the,program is operating in the black. However, a second question involves theeconomic returns to vocational-technical education relative to alternative Uses ofsocial capital. For example, should aJditional funds be spent onvocational-technical education relative t9 competing secondary curricular Theanswer to this question requires an estimation of the additional or extra benefitsyielded by vocational-technical e.clucation for each additional dollar spent. Ineconomic parlance, marginal (or extra) benefits must be compared to marginal(or extra) costs

. To repeat, the distinction between average and marginal is as follows.Average costs (or benefits) equal total costs (or benefits) divided by totalpersons in the program. Marginal costs (or benefits) are the additional costs (orbenefits) due to adding an extra person to the program. Marginal costs in thisanalysis are usually estimated with a Itatistical cost, function by relating totalcosts to total enrollments in a program to see how total costs change as totalenrollment changes by one unit. However, in some cases marginal cost in thesestudies is shown as the difference between two average costs-that of theexperimental group and that of the control. Marginal benefits in this analysis areestimated by comparing the difference in average performance of theexperimentaLgroup and the control group. Strictly speaking, all the marginalbenefits in this survey analysis arediffilinat-between-two averages-But, Toneaccepts the assumption that shifting a person from one group to the otherincreases the average benefit by the amount of the differences in the twoaverages, then this difference can be assumed to approximate a marginaldifference. A similar assumption must be made when differences in average costsare treated as marginal costs.

The studies of specific cities by Hu, et al., (1969, II) Kaufman and Lewis,(1968, 11) and that by Corazzini (1968, 11) and Taussig (1968, H), indicatelhatthe marginal rate of return to vocational-technical education only falls below thelower bound of five percent for the social rate of return to capital for New YorkCity: The marginal rate of return to vocational-technical (comPared to thecurricula of the comprehensive high school) is 31.8 percent in Detroit, 8.2percent in Philadelphia, and 17.9 percent in Worchester, Massachusetts. It is 4.6percent for males in training related jobs in Ne* York City but zero for females.

One qualification should be noted at this point. The analysis sugests thatsecondary vocational-technical graduates as a vp do better (earn more) than

3 Ten percent is the laktel upper limit placed on the social opportunity cost rate of capital.Under current conditims of high interest rates, one might argue for a higher upper limit.However, to the extent that this higher rate dimmest is due to inflation, it should be de-flated. The social rate of interest or the social opportunity cost rate is usually defined asthe Hales, deflated interest cost rate. The term "riskkss" implies no risk of defarlt onpaynient of interest or principal. It does not imply ilea or risk that the investment mayyield no reel benefit.

TA

BL

E 6

CO

MPA

RA

TIV

E A

NA

LY

SIS

OF

CO

S A

ND

BE

NE

FIT

EST

IMA

TE

S O

F SE

LE

CT

ED

ST

UD

IES

OF

POST

-SE

CO

ND

AR

Y V

OC

AT

ION

AL

-TE

CIC

AL

ED

UeA

TIO

N A

ND

JU

NIO

R C

OL

LE

GE

ED

UC

AT

ION

.

Nom

e of

Stu

dy

Tim

ePe

riod

of S

rody

lL

OC

IIII

of

Stud

yE

xpei

imen

tal

Gro

upC

astr

oG

roup

Cos

triti

ega

freg

iaiy

eai6

%M

ilos

ofT

rain

ing

in.

Yea

rs .

Dom

io o

f&

wri

ts in

Yea

rs4

Rat

e of

Rea

mfp

erce

nt)

Pres

ent V

alve

in D

olla

rsM

.i.jI

Ave

rage

Mar

ion;

Ave

nge

5%10

%

I. F

ernb

ach

and

Som

ers

a. 2

b.

1964

-69

.

Nat

ionw

ide

Post

-sec

ond-

ary

voca

tion-

al e

duca

tion

.

Seco

nd,

aty5

AM

-de

mic

2494

,25

197

. 308

4,31

32

996

996

2 2

10.

10

8.7

6.8

1198 64

2

Neg

Nes

2. S

omer

s, e

t al.

a. ..

1964

-69

".-

Nat

ionw

ide

Post

-sec

ond-

ary

voca

tion-

al e

duca

tion

Seco

ndar

yvo

catio

n-al

edu

ca-

tion

i

n

I

2814

2839

7

3504

,35

52 '

329

329

2 2

10

,10

Neg

Neg

Neg

Nes

Neg

Nes

3. S

omer

s, e

t aL

a. b.

1964

-69

"

Nat

ionw

ide

Juni

or c

olle

gePo

st-

Sela

ryvo

caed

ucat

ion

..

nal

2598

,26

167

3110

,31

44

1642

1642

.

210 10

24.9

20.0

-

,.

.

6986

5971

4196

3204

4. C

arro

ll an

dlh

nen .

1959

-64

.

,

Nor

thC

arol

ina

"n

Post

-sec

ond-

ary

voca

tion-

al e

duca

tion

Seco

ndar

iA

cad

"c

3551

3874

855

53.

243

16.5

1552

351

57

o

a

I

5: S

omer

s, e

t al.

b..

1964

-69

Nat

ionw

ide -

Juni

or c

olle

pSe

cond

ary

Voc

atio

nal

educ

atio

ni

3474

,-34

74i

1656

210

'N\

\'6N

.,.7.

5400

2617

Not

es:

1T

ime

peri

od o

f st

udy

incl

udes

the

trai

ning

peri

od a

s w

ell a

s th

eav

aila

ble

time

for

follo

wup

at t

he ti

me

the

data

wer

e ga

ther

ed.

2 (a

) in

clud

es c

urre

ntop

erat

ing

cost

san

d op

peA

tuni

ty c

osts

; (b)

incl

udes

cur

rent

ope

ratin

gco

sts,

cap

ital c

osts

and

oppo

rtun

ity

cost

s.'3

Ben

efits

are

est

imat

ed to

incr

ease

at a

rat

e of

:161

per

yea

r an

dre

ach

$1,0

38 in

the

four

th y

ear

afte

rgra

duat

ion.

In

gene

ral,

a

two

perc

ent g

row

thra

te w

as a

pplie

dto

lthe

earn

ings

dif

fer-

Atti

ebe

twee

n th

e po

st-s

econ

dary

tech

nica

l gra

duat

es a

ndth

eir

cont

rol

grou

p, c

ompr

ehen

sive

high

sch

oolg

radu

ates

.t

4. E

xcep

t in

the

case

of

the

Car

roll

and

stud

y, b

enef

it du

ratio

nis

bas

ed o

n th

ees

timat

e of

Max

U.E

ning

er, T

he P

roce

ss a

nd

Prod

uct o

f T

and

I H

igh

Scho

ol L

evel

catio

nal E

duca

tion

in th

eU

nite

d St

ates

: Tne

Prod

uct,

Pitts

burg

lf,Pe

nnsy

lvan

ia, 1

965.

5 O

ppor

tuni

ty c

osts

are

base

d on

the

earn

ings

expe

rien

ce o

f se

cond

ary

voca

tiona

l rat

her

than

seco

ndar

y ac

adem

ic. T

his

resu

lts in

an

upw

ard

bias

in th

e co

stes

timat

e am

l a c

orre

spon

ding

redu

ctio

n in

net

ben

efits

.

6 A

ll be

nefi

ts a

rebe

fore

-tax

ear

ning

s an

die

pres

ent a

soc

ial b

enef

it,th

at is

, an

incr

ease

in v

alue

-add

ed in

the

gros

s na

tiona

lpro

duct

.

7 C

osts

are

exp

ress

edfo

r ea

ch o

f th

e tw

osu

cces

sive

yea

rs o

f tr

aini

ng.

W8

Cos

ts a

re J

otal

soci

al c

osts

, whi

chin

clua

ie c

urre

nt o

pera

ting

cost

s, c

apita

lcost

s an

d op

port

unity

cost

s.

!

Sour

ces:

1, B

enef

itda

ta f

rom

: Sus

anFe

rnba

char

Li G

eral

d G

._ S

omer

s, A

n A

naly

sis'

ofth

e E

cono

mic

Ben

efits

of V

ocat

iona

l Edu

catio

nat

the

}

.

L--

---

Seco

ndar

y, P

bst-

Seco

ndar

yan

d Ju

nior

rol

lege

Lev

els,

Pre

limin

ary

Rep

ort,

Mad

ison

,Wis

cons

in, M

ay, 1

970.

Cos

t dat

a ar

e fr

om:

Fern

bach

and

Sbm

ers,

Ana

lysi

s .,.

. op

. cit.

,M

ay, 1

970,

and

Will

iam

C. M

orsc

h, S

tudy

of C

omm

unity

Col

lege

s an

d V

ocat

iona

lT

tain

ing

Cen

ters

: Cos

tAna

frsi

s, W

ashi

ngto

n,D

.C.,

1970

.,

2. B

enef

it da

ta a

refr

om G

eral

d G

. Som

ers,

et a

l. T

heE

ffec

tiven

ess

of V

ocat

iona

land

Tec

hnic

alPr

ogra

ms:

A N

atio

nalF

ollo

wup

Surv

ey. M

adis

on, W

isco

nsin

,19

70.

1

.

-Cos

t dat

a ar

e fr

omFe

rnba

ch a

nd C

omar

aA

naly

sis

.. .

op.

cit.,

May

, 070

and

Moo

ch, S

tudy

, 197

0;

3. B

enef

it da

taam

_fr

om S

omer

s, e

t,al.

Sirv

ey, .

. .

op. c

it., 1

971.

..

Cos

t dat

a ar

efr

ont-

Feni

bach

and

Som

i s, A

naly

sis

. . .

op. c

it.,M

ay, 1

970

and

Man

ch,S

tudy

, 197

0.

4. A

dger

B. C

arro

ll;-

nd L

o. l

hnen

"Cos

ts a

nd R

etur

nsfo

r T

wo

Yea

rs o

fPos

t Sec

onda

ry T

echn

ical

Scho

olin

g:A

Pilo

t Stu

dy,"

Ictu

rnal

of

POlit

ical

Eco

nom

y, V

oL-7

5 N

o. 6

,Dec

embe

r, 1

967.

5. B

enef

it an

d co

stda

ta a

re f

rom

Som

er, e

tal.,

Sur

vey

....

op.

cit.,

197

1.

academic or comprehensive high school graduates. But not all ocCupationalspec ialties . in vocational-technical education pay. off equally well. Thebenefit-cost analysis, thus far, onlY answers the question of which broadcurriculunfirea society should inveit its additional social capital in;, it.does dotindicate which occupational skill or specialty one should chaise within thatbroad curdenlum, given the qualificatios to the analyses. It does indicate thatthe average rnix 'of *ills in vocational education gains A higher rate of returnthan the averap nix of skill:in the comparisons area. -- .

. Tic) additional conunents &add be made. First, the muffinsl rate ofreturn is higher in the Hu, et aL and Kaufman-LeWis studies, where there hasbeen multivariate control for various sociodemographic variables (see Table II)than in the Corazzini and Taussig studies which control only for sex. Next, theHu, et aL and Kaufman-Lewis studies depend for their data on more .elabOratelabor market questionnaires than do the. Corazzini and Taussig studies. Forexample, a dx year employment and earnings-history exists for the Hu, et aLstudy while torazzini uses starting _wage differentials. Thus, based on thesedifferences in methodology more confidence should be placed in the relativemagnitUdes of the results in 11u, et al. and Kaufman-Lewis than in the other twosingle city studies. By this, we mean that the Corazzini and Taussig results mayunderstate somewhat. the money ,:.Aue of voCational-technical education.

'Nevertheless, these are all case studies and, in the final analysis, do not present abase of results broad enough on which to base national expenditure decisions.

The studies of Fernbach 'and Somers, Project TALENT and Eninger are allbised on national samples. The rates of return here more closely approximateaverage rates of return. As can be seen, 'the 'rates are relatively high 'andconsistent even given the differing methodologies, nonresponse rates, etc. Therates appear to he welt aixiVe this study's presumed upper bound of 10 percentfor the social opportunity cost rate of capital. The study in process by theNatiOnal Planning Association should dispel any remaining ambiguities as to theactual money value of vocational technical education (See Chapter V).

Post-Secondary Voattional,Technical Education and Junior College Thesecond major context for decisions regarding the training of the U.S. labor force

lies in- the area offost4econdary vocational-technical-education. The study uponwhich the bulk of this analysitis based-aircrourat the-Univenity_of Wisconsin.__and. the Bureau of Social Science Research (See Table 6). The analysis pertainstO a nationwide, sample of secondary academic and vocational-technicalgraduates and post-secondaiy vocazional.technical and junior, college graduates.As 'can be seen in Table 6, the marginal rate of return for post-secondaryvocational-technical education is 6.8 percent with respect to secondary academiceducation (see Fernbach and Somers (1970, 11) study 1 in Table 6). While the 6.8percent rate is less than' the assumed 10 percent social cost of capital, it is stillhigher than the usual lower bound of the social capital cast estimate, which Isfive percent. Based .on these results', it is econamically efficient for society toinvest in post-secondary voeational-technical -education for i person who is anacademic curriculum high school graduate. This judgment is borne out by theCarrollind lhnen study (1967;11) which shows a marginal rate of return of 16.5Percent for post-secondary vocational education relative to aeademic high school

56

graduation in North Carolina. Howl,/er, Somers, et d. (1971, IV), also shows thata person who is a graduate of a secondary vOcational-technkid curriculum suffersa net economic loss if he undertakes two additional years of pbstsecondary

"...Ns' vocational-technical education.' It nay be that on economic efficiency groundstherefore, aocielY shossld disCourage this educational sequence, units or untiladditional 'empirkal evidence shows a more favorable rate of return. Of course,this judgment can be tempered by nonefficiency considerations. Namely; it may..not be politically possible or socially desiraVe to prohibit this educationalsequence in a free society. Also,some economic and all noneconomic benefitsare unaccOunted for. . .

Somers, etal, (1971, IV) also provide evidence on the economic returns to. junior cojege training. The marginal rate of return to junior college relative topost irconEjary vcicational.technical ethication is 20 percent. Thus, it may beeconomical y more rational for society to invest in two years of junior cone,than for i . to invest in two years of postsecondary vocationaechnicaleduCation. finally, unlike post-secondary vocational-technical education', twoyears of junior college relative to secondary vocational trait:11g yield a marginalrate of return of 17.6 percent.

. There remains the problem, noted at the outset, that the populations .

served by the two types of school may differ. To the extent that this is so,benefit-cost comparisons between the two- types of postiecondary education arenot strictly vslid. One may object th2t the pagpeption of each tYpe ofpostseCondary education into its skiil or Course components will reveal thateach has some skills that pay off well -in -economic tenm and others that are oflow economic value. The point to be made here, however, is that the average mixof skills and courses taught in the junior college yields a higher rate of return .

t han .the average mix of skills and courses* taueltt in postaecondaryvocational-technicil institutions.

Institutional and On-the-Job Manpower Trainint It is often asserted thatvocational education should be training the labor, force for jobs in the future.the"jobs of tomorrow." However, reflection on do hands of econotnic planningand economic projection in general, given a. technologically dynamic economy,should bring one to the realization that this is a counsel 'of perfection. Manypeople change occupations several times during theii lives, and three who do notnormally do so, such as profenional persons, find the requirements of their jobsto_be a.tonstantly evolving process which only aosiduous on-the-job training ctnkeep one abreast Of.Even if vocational education were training appropriateiy for.the "jobs of today," the 'short run in ihich most of us live, it 'would not beunreisonable to expect- the national manpower retraining effort to fipgrade the .

labor force at the tichnological .requirements ifoi human cspital dunge. Also,training for the jobs of tomorrow implies benefits which will not arrive until ;

. tomorrow; 'either, with the attendant fact 'that these benefits are discounted athigher geometrically compounded discount factors. Thus, manpower training is .

.certainly complementary to and not necesearilycoMpetitive with or a substitute .

for vocational training. However, as will be indicated below, vocationaleducation might learn _some lesecins from'manpower training, which 'Concentratesthe educational effort in a relatively short calendar elite period.

c

TA

BL

E 7

-

CO

MPA

RA

TIV

E A

NA

LY

SIS

OF

CO

ST A

ND

BE

NE

FIT

EST

IMA

TE

S O

F SE

LE

CT

ED

ST

UD

IES

OF

MD

TA

IN

STIT

UT

ION

AL

-AN

D O

N-T

HE

NI0

11 R

ET

RA

ININ

G A

ND

AR

A R

ET

RA

ININ

G

- N

ame

of S

tudy

8T

ime

Peri

odof

San

dy 1

Loc

os o

fE

h*E

rlam

eSal

lC

amp

Con

trol

Gro

up

cosi

fyie

r2is

meg

kiyi

er3

Dun

a= o

fIn

ansg

in Y

ems4

Dai

ntie

s et

Iles

efits

isl'a

usS

Rat

e et

Rea

mfg

esee

nt)

' Pre

sent

%M

e ita

Odi

nM

OW

Ase

lnp

MO

Wkr

este

r.3%

---1

1-6.

1. M

ath -

.

/19

65-6

6

'

Nat

ionw

ide

I MD

TA

gra

d-ua

tes

and

drop

outs

-

1une

mpl

oy.

ad r

ela-

tives

or

neig

hbor

s

1983

1983

.

409

409

1 1

Id 35°

15.9

20.2

1110

4490

482

1783

2. H

ardi

n an

d B

orns

,19

62-6

5...

,M

ichi

gan

MD

TA

en-

roU

ees

MD

TA

aPP8

-ca

nts

1272

.127

225

125

11

10 3514

.719

.763

427

0324

610

44

3. M

uir,

et a

l.

a. I

nstit

utio

nal

..

b. O

n-th

e-Jo

b.

4

1963

-65

1963

-65

Nat

ionw

ide

Judg

men

tSa

nti,1

4

Nat

ionw

ide

Judg

men

tSa

mpl

e

MD

TA

gra

-

duat

zr

MD

TA

grid

tiate

s

IBef

ore.

afte

rC

ompr

i-sO

fl

IBef

o-Ia

ftr ompa

ri-

son

2444

2444

2132

2132

1338

1338 .

1208

1208

.

I 1

10 33 10 32

'

54.0

547

.

56.0

56.7

7512

1806

4

.64

5316

150

5252

9418

4800

8524

,4. S

trom

sdor

fer

1959

-63

Wes

t

V;u

gini

a IA

RA

and

stat

egr

adua

tes

632

632

874

874

1 1

. 10 33

138.

0 .

138.

358

2613

172

4308

7114

.

5. B

onn

'19

62-6

3

, Con

nect

icut

.

AR

A a

nd-.

...t..

..- te

pro

gram

grad

uate

s

1314

t345

911

591 1

,10 39

3012

--32

.520

29§0

93 .

1279

2766

6. P

age.

.19

58-6

1.

Mas

sach

u. .

'ens

.

MD

TA

grad

uate

s1-

693

.169

3'

pa 874

1 1

' -

10.

3550

.851

.648

1512

017

3343

6124

ass.

so'

.

p

.-

7. S

ewel

l-a.

Ins

titut

iona

l

..

.b.

On-

The

-Job

1965

-67

.

.

Nor

th.

Car

olin

a'

-

MD

TA

gra

-du

ates

:dis

-ad

lrat

ased

rura

lw

orke

rs

Non

-tr

aine

eap

plic

ant

2138

2138 '

1206

1206

' '..

296

296

603

603

..

1 1

.

1 1

10 31-

... 10 33

6.3)

13.5

49_0

493

148

2477

.

- 3450

8444

Nes

668

2499

4564

:Not

es: L

Tim

e pe

riod

of,

stgc

ly in

clud

es b

oth

the

trai

ning

per

iod

and

the

peri

od a

vaila

ble

for

mig

asur

ing

bene

fits

at t

he ti

me

of d

ata

colle

ctio

n_2.

All

cost

s ar

e to

tal s

ocia

l cos

ta a

nd th

us in

clud

e cu

rren

t ope

ratin

g co

sts,

cap

ital c

osts

and

opp

ortu

nity

cos

ts.

'.

1. A

B b

enef

its a

re b

efor

e-ta

x ea

rnin

gs a

nd r

epre

sent

a s

ocia

l ben

efit,

lhat

is, a

n in

crea

se in

val

ue-a

dded

in th

e gr

oss

najo

nal p

rodu

ct.

4. '

Mos

t MD

TA

and

AR

A tr

aini

ng la

sted

less

than

one

yea

r. T

he o

ne,y

ear

-dur

atio

n as

sum

ptio

n, th

eref

ore,

Und

ersK

tes

the

pres

ent

valu

e of

cos

ts a

nd th

e ra

te o

f re

turn

a s

mal

l am

ount

in th

e di

scou

nt p

roce

ss; h

owev

er, i

t was

use

d fo

r pu

rpos

es o

f si

mpl

ific

atio

n.S.

The

.10-

year

per

iod

of b

enef

its is

in a

rbitr

ary

peri

od c

hose

n fo

r co

mpa

riio

n w

ith th

e-es

timat

ed d

urat

ion

of b

enef

its a

ccru

ing

tose

cond

ary

voca

tiona

l-te

chni

cal e

duca

tion.

The

long

er p

erio

ds a

re th

e ne

t est

imat

ed w

orki

ng li

fe o

f th

e tr

aine

et a

fter

the

end

of \

-tr

aini

ng.

.-

..

..

-

6. T

he s

tudy

of

Am

es L

Ste

rn, "

Con

sequ

ence

s of

Pla

nt C

losu

re,"

Jou

rnal

' of

Hum

an R

esou

rces

, Win

gr, 1

972,

for

thco

min

g, is

not

repo

rted

her

e. I

n a

stud

y us

ing

Soci

al S

ecur

ity-A

dmin

istr

atio

n ea

rnin

gs d

ata

-for

MD

TA

trai

nees

inith

e K

ansa

s C

ity la

bor

mar

ket

for

the

year

s 19

6447

, Ste

rn f

ound

that

wor

kers

ret

rain

ed a

fter

a p

lant

shu

tdow

n ea

rned

$57

2 le

ss in

196

7 th

an d

id d

ispl

aced

,- -

.w

orke

rs f

rom

the

sam

e pl

ant s

hutd

own

who

sou

ght w

ork

on th

eir

own

in th

e m

arke

t. T

hus,

the

pres

ent v

alue

of

the

trai

nees

'ea

rnin

gs it

neg

ativ

e as

wel

l as

the

rate

of

retu

rn.

..

,7.

Wor

kers

une

mpl

oyed

dur

ing

Prog

ram

.g.

Une

Voy

ed w

orke

rs w

ho r

efus

ed r

etra

inin

g-

.,

-'

.,

-Sou

rces

: 1: '

Ben

efits

are

take

n fr

cfm

: Ear

l D. M

ain,

"A

Nat

ionw

ide

Eva

luat

ion

of M

DT

AIn

stitu

tiona

l Job

Tra

inin

g,"

Jour

nal o

f ku

man

-. .

---

..-:',

'R

esou

rces

, VoL

[W

M?.

2, S

prin

g.19

68. C

osts

ate

take

n fr

om: R

alpl

t E. S

mith

," o

fego

ne E

arni

ngs

Dur

ing

Man

pow

er T

rain

ing,

"W

orki

ng P

aper

350

-11,

The

Urb

an I

nstit

ute,

28

Janu

ary

1970

_ pu

blis

hed

in'

-Joi

ntE

cono

mic

Com

mitt

ee, S

ubco

mm

ittee

on E

cono

my,

in G

over

nmen

t, N

atio

nal P

rior

ities

, 148

Jun

e 19

70, a

nd G

arth

DT

A: F

ound

atio

n of

Fed

eral

Mak

pow

er -

Polic

y, th

e Jo

hnt H

opki

ns P

ress

, 196

8..

,.

%0

..'

-

2. E

inar

Har

din

and

MiC

hael

E. B

onn,

Eco

nom

ic B

eret

s an

d C

osts

of

Rep

airi

ng C

owes

inM

ichi

gan,

Eas

t Lan

sing

, Mic

hn,.

Dec

embe

r, 1

969-

e.n

3. B

enef

its a

re ta

ken

from

: Alla

n H

. Mui

r, e

t at,

Cos

i/Eff

ectiv

eues

s A

nabi

sis

of C

M-t

he-J

ob a

ndIn

stitu

tiona

l. C

ours

es, W

aihi

ngto

n,.

.i

.

-42

D. C

.,4un

e, 1

967.

Cos

ts a

n ta

ken

from

: Ral

ph E

. Sm

ith, "

Ear

ning

s. ..

," o

p. c

it., 1

970.

. Ask

A

'

a

4. E

rnst

W. S

trom

ador

fer.

"D

eter

min

ant's

of

Eco

nom

ic S

ucce

ss in

Ret

rain

ing

the

Une

mpl

oyed

," J

ourn

al o

f H

uman

Res

ourc

es, V

oLH

I, N

o. 2

, Spr

ing,

196

8.I

°.

,.

5. B

enef

its a

re ta

ken

fron

t Mic

hael

E. B

orus

, "A

Ben

efit-

Cos

t.; A

naly

sis

of th

e E

cono

mic

Eff

ectiv

enes

s of

Ret

rain

ing

the

Une

mpl

oyed

"Y

ale

Eco

nom

ic E

ssay

s. V

ol. 4

, No.

;, F

all,

1964

. Cos

ts a

re ta

ken

fiom

: Ral

ph F

.. Sm

ith, "

Ear

ning

s ...

"op

. cit.

,19

70.

^

6. B

enef

its a

re ta

ken

from

: Dav

id A

. Pag

e, "

Ret

rain

ing

Und

er th

e^14

Anp

ower

Dev

elop

men

t Act

: A c

ost-

Ben

efit

Ana

lysi

s,"

in J

ohn

D. M

ontg

omer

y an

d A

rthu

r Sm

ithie

s, P

ublic

Pol

icy

tII

'Vol

. 13

Har

vard

Uni

vers

ity P

ress

, 196

4. C

Ost

s ar

e' ta

ken

from

: Ral

ph E

.Sm

ith, "

Ear

ning

s . .

:,"

op. c

it., 1

970.

.e.

7. B

enef

its a

nd c

osts

are

take

n fr

om: D

avid

0. S

ewel

l, na

min

g th

e Po

or, K

ings

ton,

Ont

ario

, 197

1.., ..

b

^

`lre

.

TA

BL

E S

CO

MPA

RA

TIV

E A

NA

LY

SIS

OF

CO

ST A

ND

BE

NE

FIT

EST

IMA

TE

S O

F SE

LE

CT

ED

ST

UD

IES

OF

ME

JO

BS

AN

D C

EP

PRO

GR

AM

S.

I

Nam

e of

Bod

y

Tim

peri

odof

Sha

dyla

m a

Stud

yE

apd1

IG

i.up

Cea

trel

Gra

sp

Mar

sem

egity

ar__

_sof

Avo

w

Dum

dum

Tra

sia

Yen

s

Dan

dies

of B

emer

ksia

Yea

n

I R

ote

ofR

eno

.(ps

reed

)

/WN

Wla

Dam

"Ilr

-711

1V

ain

Ann

ie.

fari

pill

JOB

S:1:

Gre

enle

igh

Aza

ocia

tesi

Inc

..

1969

-70

10 S

MSA

s. Pr

ogra

men

rolle

esB

efor

e-af

ter

com

port

.ao

n

3488

134

88

,

,227

4212

7413 1

.

104

40

7

64.8

65.2

0

1423

733

836

2966

1340

1

2. D

epar

tmen

t of

Lab

or19

66-6

8N

atio

nwid

ePr

ovam

enro

llees

Bef

ore-

afte

rco

mpo

rt-

son

3239

132

3910

152

1015

-13 1

. .,

,

.104

4028

.931

.3.

,

4380

1350

327

2560

79

CE

P:3.

Ope

ratio

nsR

esea

tch,

Inc

.,

1967

-68

7 ce

ntra

lci

ties

Enr

olle

espl

aced

injo

bs

Bef

ore-

afte

rco

mpa

ri-

ion

3470

3470

2228

522

28c

1 1

10 37 .,,

63.7

64.2

1308

031

255

9291

1650

4

Not

es:

1. C

osts

are

tota

l soc

ial ,

cosi

s an

d in

clud

e al

l pay

men

ts to

the

firn

i plu

s oP

port

umty

cos

ts r

epre

sent

ed b

y th

e pa

rtic

ipan

ts' s

tart

ing

wag

e an

d fi

nal w

age

drie

rent

is:..

The

se c

osts

may

ove

rsta

te tr

ue s

ocia

l ;co

sts.

san

ce *

som

e of

the

cost

s in

pute

d by

fir

ms

may

be

paym

ents

for

idle

cap

acity

, the

use

of

whi

ch d

oes

not r

epre

sent

a s

ocia

l cos

t; th

at is

, in

the

abse

nce

of th

e JO

BS

prog

ram

, the

capa

city

wou

ld n

et h

ave

had

any

alte

rnat

ive

uw.

2. B

enef

it di

ffer

entia

ls e

re d

efla

ted

hy th

e C

onsu

mer

Pri

ce I

ndex

as

a pa

rtia

l adj

ustm

ent f

or e

h/m

acs

in la

bor

mar

ket c

ondi

tions

betw

een

the

"bef

ore"

and

"af

ter"

mea

lure

men

t per

iods

.3.

The

one

-yea

r on

-the

-job

trai

ning

per

iod

is a

n ob

viou

s si

mpl

ific

atio

n \ f

or c

ompu

tatio

nal p

urpo

ses.

The

Sub

wm

mitt

ee o

nE

mpl

oym

ent,

Man

pow

er, a

nd P

over

ty, 1

970,

sug

gest

s th

at th

e ow

the-

jok

trai

ning

per

iod

is a

bout

nin

e m

onth

s. S

ee p

. 169

.4.

The

10-

year

dis

coun

t per

iod

is a

n ar

bitr

ary

peri

od c

hose

n fo

r pu

rpor

s of

sta

ndar

d co

mpa

riso

n. T

he 4

11ye

ar p

erio

d re

pres

ents

the

aver

age

wor

king

life

rem

aini

ng to

the

JOB

S en

role

e af

ter

his

on-t

he-j

ob tr

aini

ng is

com

plet

ed.

5. T

hese

are

gro

ss w

eigh

ted

wag

e ra

te d

iffe

rent

ials

bas

ed u

pon

the

wag

e ra

te e

xper

ienc

e of

min

Ori

ty d

isad

vant

aged

mal

es, a

ge 1

8-21

and

age

22-4

4, a

nd m

inor

ity d

isad

vant

aged

fem

ales

, age

18-

21 a

nd 2

2-44

.

.

urce

s:1.

Ben

efit

data

are

fro

m: G

reen

leig

h A

ssoc

iate

s, I

nc.,

The

Job

Opp

mun

ities

in th

e, B

usin

ess

iect

or P

rogr

am: A

n E

mlu

atio

n of

Impa

ct in

Ten

.Sta

ndar

d M

etro

polit

an S

tark

:Mel

Are

as, W

ashi

ngto

n, D

. C.;

1970

. Cos

t dat

a ar

e fr

om: S

yste

m D

evel

opm

ent

Cor

pora

tion,

Eva

luat

ion

of th

e JO

BS

Prog

ram

in N

ine

Citk

s, F

alls

Chu

rch,

Vir

gini

a, 1

969.

2. B

enef

it da

ta a

re f

rom

: U.S

: Sen

ate,

Com

mitt

ee o

n L

abor

and

Pub

lic W

elfa

re, S

ubco

mm

ittee

on

Em

ploy

men

t, M

anpo

wer

and

Pove

rty,

The

JO

BS

Prog

ram

(Jo

b O

ppor

tuni

ties

in th

e B

usin

ess

Sect

or),

Bac

kgro

und

Info

rmat

ion,

91s

t Con

gres

s, 2

nd S

essi

on,.

Apr

a, 1

970.

Cos

t dat

a ar

e fr

om: S

yste

m D

evel

opm

ent C

orpo

ratio

n, E

wth

istio

n. o

p. c

it., 1

969.

3. B

enef

it da

ta a

re f

rom

; Ope

ratio

ns R

esea

rch,

Inc

., C

EP

Eva

luat

bn M

etho

dolo

gy, P

hase

I R

epor

t, 'T

echn

ical

Rep

ort 6

17, D

raft

,Si

lver

Spr

ing,

Mar

ylan

d, 1

3 Ju

ly 1

970.

Cos

t dat

a se

e fr

om: L

easc

o Sy

stem

s an

d R

esea

rch

Cor

pora

tion,

Qua

ntite

tive

Ana

lysi

s of

the

Con

cent

rate

d E

mpl

oym

v:r

Prog

ram

, VoL

11,

Silv

er S

prin

g, M

aryl

end,

196

9.

Seven .different studies of manpower training exist that present abenefit-cost analysis. (See Table 7.) The Main study (1968, Ill.) is judged to givean accurate assessment of the net returns to institutional manpower training,since its methodology involves the use of an appropriate control group for anationwide sample and adjustments were made for major sociodemographic,motivational, and economic variables. If the benefits to such training areassumed to last only 10 years, then the marginal rate of return is 15.9percent11---the benefits are assumed to last 'the remaining working-fife Of-the trainee, 35years, then the marginal rate of n-is 20:2 percent. Note that because thebeneft increments_ar igh, the extra 25 years of benefit stream add little

jo-the--rate---6f-Teturn. The high benefits relative to costs imply a high discountrate which makes the extra years of benefit relatively unimportant. The othernationwide study, by Muir et al. (1967, Ill) Shows much higher Marginal rates ofreturn for institutional MDTA training, but there is an upward bias in theseestimates due to the use of a "before/after" labor market comparison for thetrainee rather than a control group comparison. In short, manpower training is anecessary complement to vocational training in a technologically evolvingeconomy. The fact that different populations may be served by the twoprograms also reinforces their complementarity,,The marginal rates of return tomanpower training at least equal and are probably higher than those tosecondary vocational education. But, to repeat, one should not necessarily drawthe conclusion that manpower training can be substituted for vocationaleducation, however, since the two program do serve different social groups.

Next, it should be noted that for the MDTA program, the marginal rates ofreturn to the institutional and the on-the-job cOmponents of the program areSimilar, based on the benefits by the Muir et al. study. Thus, given present data,there is no economic efficiency basis for expanding one of .these programcomponents at the expense of the other.

One troublesome aspect of these studies which casts some doubt on the.. empirical reasonableness of these high rates of return is the consistent failure of

the market to provide funds for what appears to be a return of liberalproportions. The response to this nuy be in the institutional constraintssurrounding the capital market-mainly the quasi-illegalilty of indenturingoneself plu3 the fact that the created capital is inseparable from the human agentand hence, not separately .capitalizable and marketable. However, with such highreturns one suspects that the market would eventually respond with institutionalarrangements to make the funds available. That it hasn't makes one suspect thatmajor risk factors which would reduce these'rates remain unaccounted for.

Finally, the study by Stem should be noted (forthcoming 1972, Ill) (SeeTable 7, notes). While his is not a benefit-cost study, Stem reports that displacedworkers who were retrained did considerably less well than -displaced personshon: the same plants who either transferred to a new job on the firm's invitation(not too surprising a result) or who, more importantly, simply sought new jobsin the market on their own initiative. Since Stem 'controlled for a variety ofsociodemographic variables and used social security data for.his wage measures,there should be little reporting error in it, though the method of extrapolatingthe data does impart some bias. In short, for all locales and all populations ofworkers, the final word is not in yet.

62

JOBS and CiP. The JOBS and CEP training programs are close substitutesfor MDTA training. The JOBS program is similar to the MDTA on-the-jobtraining program except insofar as the initiative of the JOBS program may be,more with private employers. The CEP contains elements of the institutional and.-on-the-job MDTA training. The study results appear in Table 8.

Unfortunately, before/after comparisons of enrollee experience must berelied upon for evaluating J011S and CEP, rather than the use of more suitablecontrol groups. The best analysis of the costs and benefits of the JOBS progranisis that by the U.S. Department of Labor. This analysis is based on a nationalrandom sample taken froni social security recqrds. A before/after comparison isused to Measure benefits. If a 10-year benefit period is.assumed, the marginalrate of return is 28.9 percent, while the rate Is 31.3 percent when benefits areassumed to last the remainder of an enrollee's wotking life. The MDTA studythat is most similar tothe JOBS analysis is the nationwide evaluation by Muir, etaL, which also uses a before/after comparison: Comparable rates of return toon-the-job MDTA trgning are 561) percent for a 10year benefit period and 56.7'percent for .. remainini. 4inking life after training. Thus, under, currentarrangements, Mink on-the-job training is yielding a marginal rate of return

which is almost twice that of the JOBS program. Other things being equal, then,_it may be desirable to devote additional social capital to MDTA on-the-jobtraining rather than to the JOBS program. Other things may not bi equal,however,iince it may be desirable to maintain or expand the present level of theJOBS prograM to continue private. involvement and intitiative in manpowertraining. In any case, the recent economic downturns have seen a drastic cutbackin the JOBS program, additional evidence that manpower policies in the absenceof general high demand will have little effectiveness.

The only study of CEP which allows a benefit-cost comparison relates todata gathered from seven central cities. The marginal rates of return based on abefore/after comparison . are quite high and fall in the mid-range of ratesestimated for institutional MDTA training: Thus, there is little basis at this timefor making a distinction between the two types of programs on efficiencygrounds. A nationwide evaluation of CEP based on an appropriate randomsample with an appropriate control group would be useful though the similarityof this program to the various MDTA components may make this superfluous.

Finally, there is no economic evaluation of the Work Incentive (WIN)program. Evaluations which do exist focus; on the administrative efficiency ofthe program, maialy within a sociological context. . . .

The Job Corim and NYC. The Job Corps and the out-of-school NYC areapproximate substitutes for each other. As Table 9 shows, the costs of the JobCorps are considerably higher than those of the out-of-school NYC. This is notnecessarily a criticism of the Job Corps since. it is patently wrong to make

1)efficiency judgments solely'on the basis o cOst comparisons without knowledgeof relative program benefits. However, f m the limited information available,the benefits of the Job Corps and the outi3f-school NYC appear to be of similarmagnitude. . .

The Cain (1967, III) study and the Resource Management Corporationstudy (1969, III) of the Job Corps are based on the same set of data, namely,

11' i63

TA

BL

E 9

CO

MPA

RA

TIV

E A

NA

LY

SIS

OF

CO

ST A

ND

BE

NE

FIT

EST

IMA

TE

S O

F SE

LE

CT

ED

&T

UD

IES

_OF

ME

JO

B C

OR

OS

AN

D N

EIG

HB

OR

HO

OD

YO

UT

H C

OR

PS (

NY

C)

Nam

e id

See

dy

Ilse

reel

edof

Sea

*W

es o

fSt

udy

Exp

lain

edG

rasp

Can

ted

Gro

up

Cos

t/Yen

rih

isei

wye

ar-

!D

amie

net

Tra

bieg

.in

Yam

Dor

atio

nO

f B

enef

itsin

Yea

rs

Ral

e of

Ret

urn

(pew

ee.)

Iles

est Y

ds,'

be D

ans

I lio

sind

Am

oyIl

lajir

l1

Am

ass

10%

JOB

CO

MPS

:I

.'

Cai

n '

1966

-67

Nat

ionw

ide

'Jud

gmen

t .Sa

mpl

ePr

ogra

mPa

rtic

ipan

tsN

o-dt

ows

3756

3756

.193

193

1 1

10,45

Nes

4.4

Nes

Neg

Nes

Nag

2.0E

019

66-6

8 '

Nat

ionw

ide

Prog

ram

Part

icip

anti

`

.

Bef

ore-

afte

rco

snpa

ri-

Ion

3613

3613

774

774

14.1 1

10 45.

,16

.921

.422

5196

6110

3936

55

3.R

esou

rce

Man

agem

ent

Cor

pora

tion

a. M

ales

.

b. F

emal

es

1965

-68

Nat

ionl

ide

Judg

men

tSa

mpl

e

Prog

ram

Part

icip

ants

No-

show

s

. 3756

3756

3756

3756

87 87 135

135

:1. 1 1 1

10 46 10 46

'

Nes

Nes

Meg 2.4

Net

Neg

Neg

Nes

Nes

Neg

Meg

Nes

NY

C O

sit-

of-

Scho

ol:

4.'B

orns

, et a

l.19

67In

dian

aPr

ogra

mPa

rtic

ipan

ts.

No-

show

s,E

ligib

leN

on-p

w-

561

561

172

172

I

1 1

10 4728

.130

.7, 7

3124

1145

110

36

NY

C in

Scbo

alan

d B

anos

erS

5.So

mer

s an

dSt

rom

sdor

fer

1965

-69

Nat

ionw

ide

,

..

Prog

ram

Part

icip

ants

,tici

pant

s

Elig

ible

Non

-Pa

rtic

i-pa

nts

,--

-

0

a. I

n-Sc

hool

b. S

umm

er.

c. C

ombi

ned

In-

Scho

ol a

ndSu

mne

r'1

I

46 41

502

2e 44

7310

80.4

429

382

421

128.

033

230

510

1860

.033

128

5

Not

es:

I. T

his

cate

gory

is f

or N

YC

par

ticip

ants

who

per

ticip

ated

in b

oth

a su

mm

er a

nd a

n in

-sch

ool

2. T

hese

are

mon

thly

cos

ts.

3. D

urat

ion

is e

xpre

ssed

in m

onth

s.T

his

mon

thly

ben

efit

is n

ot s

tatis

tical

ly d

iffe

rent

fro

m z

ero.

5. G

eral

d D

. Rob

in in

his

stu

dy o

f th

e in

-sch

ool N

eigh

borh

ood

You

th C

orps

fou

nd n

o ef

fect

rate

for

NY

C p

artic

ipan

ts. (

See

Rob

in, I

969

,1II

.) S

ince

this

stu

dy w

aren

act

ual c

ontr

olle

dproj

ect.

for

the

prog

ram

in r

educ

ing

the

drop

out

expe

rim

en1

mod

el, t

he r

esul

ts a

re o

fso

me

inte

rest

Som

ers

and

Stro

msd

orfe

r (1

970,

111

) fo

und

only

a m

inim

al e

ffec

t of

the

the

drop

out r

ate.

Nam

ely,

an

addi

tiona

l mon

th in

the

prog

ram

incr

ease

d sc

hool

atte

ndan

cein

-sch

ool

and

sum

mer

NY

C o

n re

duci

ngy

one

addi

tiona

l day

.

SoU

rces

:1.

Gle

n G

. Cai

n, "

Ben

efit/

Cos

t Ana

lyse

s fo

r Jo

b C

orps

," M

adis

on, W

isco

nsin

, 196

7.2.

Off

ice

of E

cono

mic

Opp

ortu

nity

, A a

nd X

Rep

orts

Na

II, "

Job

Cor

ps B

enef

it/C

ost S

Plan

s m

id E

valu

atio

n D

ivis

ion,

Pla

ns a

nd P

rogr

am D

irec

tota

te, J

ob C

orps

, Was

hing

ton,

DC

.y,

" E

valu

atio

n an

d R

esea

rch

Bra

nch,

3. B

enef

its a

re ta

ken

from

: Res

ourc

e M

anag

emen

t Cor

pora

tion,

-Ere

issa

tions

of

the

Now

int

Ana

lysi

s- f

or M

enpo

ner

botto

ms,

IN

C R

epor

t 131

-054

, Bet

hesd

a, M

aryl

and,

196

9.'O

lene

fit/C

ost

.," o

p. c

it., 1

967.

eery

: The

Fea

sibi

lity

of B

enef

it-C

ost

osts

are

take

n fr

om: G

len

G. C

ain,

.

4. B

onn,

Mic

hael

E.,

et e

t, "A

Ben

efit

Cos

t Ana

lyis

of

the

Nei

ghbo

rhoo

d Y

outh

Cor

ps:

elc-

4orw

reef

of

flur

rene

Res

t:duc

es, 5

, 139

-159

, Spr

ing,

197

0.So

mer

s, G

eral

d G

. and

Str

omsd

orfe

r, E

rnst

W.,

A C

ost-

Eff

ectir

enes

s St

udy

of th

e In

-Sch

Out

-of-

Scho

ol P

rogr

am in

Ind

irin

a,"

end

Sum

mer

Nei

ghbo

rhoo

d Y

outh

Cor

ps, M

adis

on, W

isco

nsin

, 197

0.

'.

-.

.

.

.

,.

.

iT

AB

LE

10

--

CO

MP

AR

AT

IVE

AN

ALY

SIS

OF

UN

EM

PLO

YM

EN

T R

ED

UC

MD

TA

AN

D A

RA

INS

TIT

UT

ION

AL

RE

TR

AIN

ING

.

,

a

.

,'

";*

'

c,

.

OF

Nam

e of

Stu

dy

.

Tim

e Pe

riod

of S

tudy

..

LO

CO

S of

&na

yE

xper

imen

tal

Gro

up

1,...

7.

.,-

-"--

- _

Com

a& G

roup

.

Var

iabl

es C

ontr

olle

d

1 .

Met

hod

Con

t,of

Ited

ectin

inU

lm 1

110Y

-m

eat R

ate

ter

cotta

ge r

ims)

Peri

odM

aim

edN

umbe

r of

Olm

emat

ions

I. B

orus

;

1962

-63

Con

nect

icut

.

AR

A a

nd s

tate

prog

ram

!

grad

uate

s

Une

mpl

oyed

wor

kers

who

refu

scd

trai

ning

Age

; Edu

catio

n; M

arita

lst

atus

; No.

old

epen

dent

s;co

urse

are

a; w

eeks

fro

men

d of

trai

ning

to in

terv

iew

;tr

aini

ng s

tatu

s.

1.,

Reg

ress

itinh

I

Com

plet

e-D

idno

t use

Tra

in-

ing:

-95

-

Dro

pout

with

out e

m7

ploy

men

t: -

12.5

(ind

of

trai

ning

to in

ter-

view

285

2. G

ibba

rd &

SWIM

S19

59-6

3.

Wes

t Vir

gini

aA

RA

and

sta

tepr

ogra

mgr

adua

tes

.

Une

mpl

oyed

and

unde

r-em

ploy

edw

orke

rstr

aini

ng

Age

; Edu

tion;

Mar

ital

stat

us; s

ex; r

aft;

labo

rm

arke

t are

a; p

rior

labo

rfo

rce

expe

rien

ce; r

egul

aroc

cupa

tion;

pre

viou

sm

obili

ty; y

ear

and

quar

ter

of tr

aini

ng e

nd; t

rain

ing

stat

us.

Reg

ress

ion

1 6

,

Dro

pout

s: n

odi

ffer

ence

.. N

ot a

ccep

ted:

-27.

7D

id n

otre

port

: -20

.8N

ontr

aine

es:

;28.

8

1 ye

araf

ter

trai

ning

1065 -

3. M

ain

0

1965

-66

Nat

ionw

ide

MD

TA

gra

du-

ates

and

drop

outs

-c

Une

mpl

oyed

rela

tives

and

.,ne

jObo

rs

Ay;

Edu

catio

n., M

arita

l!

stat

us; s

ex; r

ace;

reg

ion;

I

mon

ths

unem

ploy

ed 1

yea

rbe

fore

trai

ning

; mai

n w

age

earn

er; c

hild

ren

unde

r i 8

;tr

aini

ng s

tatu

s.

Reg

ress

ion2

',

_

Con

trol

s:-2

0.0

-

End

of

trai

ning

toin

terv

iew

2260

,

4. S

olie

(m

ales

only

)V

*

1962

-64

Ten

ness

eeV

.

AR

A g

radu

ates

.

Rek

ts-

Dro

pout

sN

ontr

aine

es

Age

; edu

catio

n; m

arita

l'

stat

us; l

abor

mar

ket a

rea;

prio

r la

bor

forc

e ex

peri

-en

ce;-

occu

patio

n; m

obili

ty;

prev

ious

trai

ning

; tra

inin

gst

atus

.

Reg

reni

on.2

5

V

Rej

ects

: No

diff

eren

ceD

ropo

uts:

-12.

5N

ontr

aine

es:

-13.

0

24 m

onth

saf

ter

trai

ning

217

5. S

trom

sdor

fer

1959

-63

Wes

t Vir

gini

aA

RA

gra

duat

es' D

ropo

uts

Age

; edu

catio

n; m

arita

lR

egr4

sion

Dro

pout

s: -

8.11

8-m

onth

879

Rej

ects

stat

us; s

ex; r

ace;

labo

r7

Rej

ects

: -26

.0pe

riod

-N

on-r

epor

ters

mar

ket a

rea;

pri

or la

bor

Non

-rep

orte

rs:a

fter

Non

tram

ees

forc

e ex

peri

ence

;reg

ular

-26.

7tr

aini

ng1

occu

patic

in; m

obili

ty;

Non

trai

nees

:tr

aini

ng s

tatu

s.-1

93'

Not

es:

1. T

he 'v

alue

in th

e co

lum

n Is

the

coef

fici

ent o

f th

e m

ultip

le d

eter

min

atio

n.T

hus,

in th

e G

ibba

rd a

nd S

omer

s st

udy,

the

vari

able

sin

clud

ed in

the

regr

essi

on a

ceou

nted

for

16

perc

ent o

f th

e va

rian

ce in

the r

ate

of e

mpl

oym

ent i

n th

e 12

-mon

tli p

eriO

d af

ter

the

end

of tr

aini

ng.

2. C

oeff

icie

nt o

f m

ultip

le,d

eter

min

atio

n w

as n

ot r

epor

ted.

.4C

111

Sour

ces:

For

Mai

n, B

orus

and

Str

omsd

orfe

r, ie

e T

able

7-

2. H

arol

d A

. Gib

bard

and

Ger

ald

G. S

omer

s, "

Gov

ernm

ent R

etra

inin

gof

the

Une

mpl

oyed

rri

Wes

t Vir

gini

a,"

in G

eral

d G

. Som

ers,

(ed.

), R

efra

inin

g th

e U

nem

ploy

ed M

adis

on, W

isco

nsin

: Uni

vers

ityof

Wis

cons

in P

ress

, 196

8.R

icha

id J

. Sol

ie, "

An

Eva

luat

ion

of th

eE

ffec

ts o

f R

etra

inin

g in

Ten

ness

ee,"

in G

ikal

d G

: Som

ers,

(ed

.). R

etra

inin

gth

e

,

Une

mpl

oyed

Mad

ison

, Wis

cons

in: U

nive

rsity

of

Wis

cons

in P

ress

,19

68.

Naa

aeau

glig

mg

*TA

BL

E 1

1C

OM

PAR

AT

IVE

AN

AL

YSI

S O

F PE

RC

EN

T O

FT

IME

- E

MIL

OY

ED

FOR

VO

CA

TIO

NA

LSECONDARY EDUCATION

t

Nam

e of

Stu

dy

Tun

e ,

Peri

od o

fSt

udy.

.L

ocus

of

Stud

y.

aped

-m

enta

lG

roup

Con

trol

Gro

upV

aria

bks

CO

ntro

lkd

Met

hod

ofC

outr

oll

Dif

fere

ntia

lR

educ

tion

inU

nem

ploy

-.m

ent r

ate

(per

-ce

nt*

Pant

s)

' Peri

odM

easu

red

Num

ber

ofO

bsei

va-

tions

I .

liu, L

ee,

Stro

msd

orfz

r19

59-6

6

.

Bal

timor

e,Ph

ilade

lphi

a,D

etro

it

Voc

atio

nal-

Tec

hnic

alA

cade

mic

Cur

ricu

lum

;la

bor

mar

ket;

sex;

IQ

; rac

e;m

arita

l sta

tus;

V

fath

er's

edu

catio

n.

V

Reg

ress

iols

t yea

r:il

16t

h ye

ar:

.28

6-ye

arav

erag

e:.1

8

1St y

r.: 1

4.2

6th

yr.:

10.1

6-54

. ave

rage

:7.

5

.

1st y

ear,

6th

year

,av

erag

e

1255

2.. K

aufm

anan

d L

ewis

--

1958

-65

Thr

ee c

ities

in P

enns

yl-

vani

a

Voc

atio

nal-

Tec

hnic

alan

d,

Gen

eral

Bus

ines

s

'Aca

dem

ic

...

City

siz

e; I

Q;.

V

curr

icul

um.,"

sei-

race

;oc

cupa

tion

onfi

rst j

ob; p

ost-

high

scho

ol tr

aini

ng;

trai

ning

rel

ated

ness

of o

ccU

patio

n:

Reg

ress

ion

.12

Voc

atio

nal-

Tec

hnic

al:

5.7

Gen

eral

Bus

ines

s: 6

.1

..-Peri

odsi

nce

grad

uatio

n

.

1226

Not

es:

I T

he v

alue

in th

e co

luni

n is

the

cOef

f de

nt o

f ni

ultip

le d

eter

min

atio

n:T

hus,

in th

e H

u, it

a/.

stud

y, th

e va

non

ce in

the

perc

ent o

f tim

e em

ploy

ed in

the

firs

t yea

r af

ter

grad

uatin

g fr

om v

ocat

iona

l hig

h sc

hool

.

Sour

ces:

See

Tab

le S

..

.

les

cont

rolle

d ac

coun

for

I I

per

cent

of

the

vari

-

^

:+1

follow-Up data cellected by. Louis Harris Associates. Althongh.theassumptionsthese studies make concerning relative benefits differ somewhat, the differencesare not greet enough to cause major divergences in the estimated marginal rate 'ofreturn. In both studies-, .the rate of return Is less than the five percent ilowerbound assumed for the social cost rate of capital. The 0E0 study (study 42 inTable 9) of the lob Corps bases benefits on a before/after tomParison. Such acomparison has a built-in upward bias to It, although the magnitude of the bias isnot known. If foreed to make a judgment, we would judge that the NYC is amore efficient social investment than the lob Corps, since the itudy of Borns et .aL .(1970, 111) has had to make fewer compromises with optima! social sciencemethodology.

Finally, the infchool. NYC has some of theattributes of a cooperativevocational educatioror a work study program. That is, while the mein focus ofthe_imschool_NYCAs_to-reducethe-oppodunity-costsof- going-to school, thework experience is also intended to impart some skills, though there is nonecessary tie between the skills one uses on the NYC job and the courses onetakes in school. HoweVer, the in-schopl NYC is a substitute for cooPerativevOcational education-insofar as it imPtoves work discipline* and reduces theopportunity costs of high school education. While the Somers-Stroirsdorferstudy (1970; 111) shovis almost no effect of the in-school and summer NYC inreducing the high school diopout rate, as the table shows, the earnings benefitsattributable to the program are very higji, especially' in light of fne shod 18month benefit period.

Cost-Effectiveness Aspects of Education and Training Those studieswhich used regression analysis to estimate benefits also provided estimates of thenet gain .in employment\ due to their particular program. Tables 10 and 11display the results as well as show the basic structure of the regressionmodels.

The similarity of the results is noteworthy, given the differences inmethodologies. The Gibbard and Somers (1968, III) Main (1968, III) and Solie(1968, 111) studies all sugest a gain in employment over the nontraineecomparison iioup of greater than 10 percentage pe'dits but less than 30. Perhapsthe Main estimate of a 20 percentage point pin is the best estimate. TheStronsdorfer data is essentially the .same as the GibbardSomers data. The onlyconflict in the data is with respect to the experience of *opouts. Gibbard andSomers report no difference for dropouts vis-a-vis trainess, while Solie,Stromsdorfer (1968,111) and Borns (1964, III) report trainee employment gainsover dropouts in the area of 10 percentage points.

With respect to secondaty vocational-technical education,. the employmentpins over the study periods measured appear tà fall in a range frOna five to 10percent. (See Table 11.)

Impicts of VOcztional and Manpower Training on SdectedSociodemographic Groups

Tables 12 through 20 display the effects of vocational, technical andmanpower training on selected sociodemographic groups. Generally, this analysissu ffers from the fact that small cell sizes exist for many criticalsociodemographic subgrOups, thus resulting in ambiguous results.

a

'TABLE 12

NET EFFECTS ON EARNINGS (IN DOLLARS) AND EMPLOYMENT (INPERCENTAGE POINTS), VOCATIONAL VERSUS COMPREHENSIVE

GRADUATES FOR SEPARATE REGRESSIONS BY RACE AND SEX

Sarapk Gr oups

First Year AfterGraduation

Sixth Year AfterGraduation

'''t Average in1 Six Years

E2 N2 E N E N

White MaleComprehensive .

Vocational 43**3 9.0** 30 2.0 44** 5.7**n=854' ,(14) .(3.0) (16) (1.8) (14) (2.l)

Nonwhite maleComprehensiveVocational 21 9.0 61 7.1 49 4.7n=98 (27) (8.7) (38) (5.5) (29) (6.1)

White femaleComprehensive 1Vocational 65** 19.5** 9 4.4 46** 12.7**n=1522 ( 7) (2.1) (11) (2.4) (7) (1.6).

Nonwhite femaleComprehensive!Vocational 41** 10.8* 32 . 54 43** 9.3*n=293' . (13) (4.7) (21) (4.5) (13) (3.6)

Source:School Graduates," Journal of Human Resources; VI (1), Winter, 1971.

1 This regressot of the variable enters into the intercept term. The other regreslors of thevariable areinterpieted as deviations from this mgressor. The variablo of labor market,IQ, marital status, and father's education are included in the separate equations, but the tcoefficients are deleted here.

2 E denotes average before tax monthly earnings, and N denotes percent of timeemployed. .

3 These itatistics are the partial regression coefficient and (in parentheses) the standarderror of the coefficient. The statistic indicates that white male vocational graduatesearned $43 more per month than did white male comprehensive graduates in the firstyear after graduation.

. * significant at the .0! level of significance, two-tailed test." significant at the .01 level Ofsignificance, two-tailed test.

Teh-wei Hu, et al., "Economic Returns te Vocational and Comprehensive High

One of the major questions plaguing the analysis of the effects ofvocational education is the issue concerning the length of time that benefitspersist. Table 12 shows that for one study for white and nonwhite males andfemales, benefits to vocational education tend to disappear after about six years.This contrasts' with the results in Eninger's study (1965, 11) of T and 1 educationfor males which estimated that benefits tend to disappear after 10 years. Thus, it

70

seems apparent that the high advantage vocational graduates eAperience doestend to disappear. The question is why, and this has not been adequatelyanalyzed. The reason may be that comprehensive students gain more on-the-jobtraining than do vocational students. Hu, et a/. (1969,11) found that high schoolgraduates completing some post-high school education hadhigher earnings ($396per year) in the sixth year after leaving high school and were employed twomore weeks during that year. But they did not investigate the exact reasons forthe,sonvergence of earnings from the standpoint of on-the-job training.

It may also be possible that comprehensive graduates have access to-jobladders which allow a more rapid earnings progression (which also implies greateropportunity for on-the-job training). This hypothesis, too, remains to be,investigated.

Vocational Education and Blacks. Table 13 displays evidence of the .premium which blacks pay to participate in the labor market.

The socially disruptive effects of racial discrimination extend throughoutthe labor market and result id generally lower earnings and eMployment ofnonwhites, both for those who elect a comprehensive high school curriculumand those who study within the vocational-technical curriculum.

Labor market discrithination based on race may be considered as of twokinds. The first can be termed "historical discrimination," attributable to

-practices and institutions that result in a generally lower leVel of health,education, and training for the nonwhite population coMpared to the whitepopulation. The second can be termed "current labor market discrimination,"the result of discrimMation between nonwhites and whites who have equalproductivity in the labor market.

. The analysis which follows is one of current labor market discrimination.The sample of observations are derived from nonwhite and white high schoolgraduates from the 1959-60 graduating classes in Detroit, Philadelphia, andBaltimore. None of these graduates had any post-secondary or two year or fouryear college education at the time they were interviewed, six years after theirgraduation. In addition, the white and nonwhite samples were futtherstandardized on the basis. of IQ, father's education, sex, marital status, type ofcurriculum followed in high School; and condition of the labor market at thetime of graduation. To our knowledge, this is the only major study of racialdiscriminatiMfrin the labsar market that controls for all the above influencessimultaneously and, hence,. isolates in its present form the effects of currentlabor market discrimination on nonwhites. Table 13 shows the .results of theanalysis of a random sample of white and nonwhite high school graduates inthree northern cities..

In the first Year after graduation white vocational-kademic graduates earn$124 more per month than do nonwhite graduates from that curriculum. Also,the whites are employed 30.63 percentage points more than nonwhite graduates.By the sixth year after high school graduation there is no statistically significantdifference in the earnings of white and nonwhite vocational-academic graduates.Howeier, in order to achieve earning equality, nonwhites must be employed8.44 percentage points more than whites.,

71

-

5

1.

TABLE 13

EPFECTS OF RACIAL DISCRIMINATION ON EARNINGSAND EMPLOYMENT FOR GRADUATES FROM SELECTED SECONDARY

CURRICULUMS. THREE NORTHERN CITIES, 19S 9-60.19661

- .Average Mint My More

Tax brainyPercent of Time

Employed

Sixth6-YewAverap First Sixth

6-YearAverage

Vocational-Academic and Vocational-Technical Secondary Graduates,Total Samplen = 1080 '

Maion = 322

Females i

n= 758 1

124**(13)

144**(33)

120**(14)

-6(19)

106**,(35)

-20(20)

81**(14)

145**(31)

71**(14)

30.61**(3.91)

22.11**10.4)

32.910!(3.6)

-84* __12.0**(3.5)

2.0(4.4)

-7.6(4.5)

(2.6)

9:7*(4.5)

15.8**(3.0)

Comprehensive Secondary Graduates- .

Total Samplen = 1687

Malesn = 630 ,Femalesn = 1057

100** 24(12) (17)

99** 145**i (27) (32)

: 93** -11(11) .(18)

76**(11)

123**(26)

59**(11)

23.6**' (3.1)

17.1**(6.0)

24.0**(3.6)

-6.0*(2.9)

.

7.0(3.9)

-8.7*(3.7)

1.1**(2.3)

10.9*(4.3)

11.1**(2.7)

Source: Unpublished data from Teh-wei Hu, et aL, A Cost Effectiveness Analysis ofVocational Edswatkat, Institute for Research on Human Resources, ThePennsylvania State University, University Park, Pennsylvania, March, 1969.

1. This analysis standardizes for the effects of month and Year of high school graduation,\ labor market at the time of graduation, IQ, post-high school training other than 2-year or

4-year college, education, marital status, sex and father's education. Thus, the differencesbetween whites and blacks in this analyds are a very close measure of currept racialdiscrimination prictkes as they existed rz three northern cities from 1959-60 through1966.

There statistics are the partial regression coefficients and their stizdard errors inparentheres. The partial regression coefficient Is interpreted as follows: It measures thedifference between average earnings or employment of white graduates and blackgraduates. Thus, white vocational-technical graduates earned $124 more per Month thanblr ck vocational graduater in the first year after they graduated.

2. First = first year after graduation; Sixth = sixth year after 'graduation; 6-year average =average experience during the 6-year period after graduation.

** significant at the .01 level of significance, two-taled test.significant at the .05 level of significance, two-tailed test.

72 so

0

The' data show, that discrimination as meatured by monthly earnings is ..

more serious in in :absolute sense for nonwhite' male vOCaticinal-technicalgraduates than , it is for their nonwhite fenule counterparts: White Malevocational-technical graduates earn $144 more per.month than their nonwhitecOunterparts in the first year after graduation; white femakiocational-technicalgraduates earn $120 more per month in the first year after graduation than theirnonwhite counterparts.

. . .

.

In the sixth year after graduation white males' earl $106 more than theirnonwhite male Vocational-technical counterparts. Although there' i nostatistically significant difference between the earnings of the two female thnicgroups,.the nonwhite. females are financially better off in that they are ea ian average: of $20 per month more and winking an average 7.56 percentagepoints less time than their white female cOunterpart4 The nonwhite malevoCationakechnical graduates, in addition, to earning $106 per Month less thanthe white graduates, are shown as wosking More '(2.0 percentage points).

The picture Pi slightly. different for COMprehaiiive high school r.aduates. .

In the first year after graduation, Whites earn $100 'a month more than ..

nonwhites and Whites are employed 23.57 percentage 'points more. In the sixthyear after. graduation there is no difference in monthly earnings between these .

racial groupings, butto achieve thia, nonwhites must work six perCentage pointsmore than their white comprehensive program counterparts. Whereas for .

'nonwhite male vocational-technical graduates the absolute . earnings cost ofdistriminatiem improved somewhat over The six-year post-graduation period, fornonwhite 'mile comprehensiVe graduates the impact of discrimination increased.In the first year nonwhites earned only $99 per month less than their *tatecounterparts; the difference increased to $145 per month in the sixth year aftergraduation. For nonwhite female comprehensive graduates the eaminipsituationUnproved. There was a $93 pet mpnth difference in the first year. In the sixth

year there was no statistically significant difference in eariiings, but this waSachieved only because nonwhite female comprehensive graduates worked 8.69percentage points more in the sli!th year than did their white femalecounterparts. , '

Thus, for a black male oi othetnonwhite 'male, it isbetter in the.lonirunto be vocational-technical graduate,than a comprehensive high ichool graduate.It will cost about,$39.á month less in-current labor market. discrimination.Among nonwhite females, however, the absolute iiiktimination costis lower forcomprehensive gradnates'than, for vocational-technical graduatiii, *- \

Education and Dropout Behavior. High schooi dropouts are inimportant sociemograPhic group whosemeeds are apparently not being Met by '

. the edOcational system. Thete-hasimn much criticism of vocational educationV because its dropont rate ha been highirthan that of .the academic curriculum.ty.

,Table 14 .ffives the relative dropout rates for Project TALENT Males. Thiscriticism oir vdcatiOnal-technical education is warranted, but is more coMplexthan a simple condemnation of the. program; as the discussion following thetable suggests.

As suggested previously, the, student who, takes the vocaVonaltechnicalcurliculum is someWhat . different froM the acadernie student in terms of

73

C) I

TABLE 14

DROPOUT RATES OF HIGH SCHOOL STUDENTSBY PROGRAM AND ABILITY QUARTIL!.!

0

( iAbility Quartde

High Si.hool Program Total ,

Low II IH High,

General 16.2 25.6 16.7 9.1 5.7 .

College Prep 3.9 22.5. 6.8 1.9 1.4COmmercial 12.5 .18.3 10.9 r 9 3 5.7Vocational 22.4 29.8 .. 18.7 8.7 .-- 2

, Agriculture 27.3 39.3 9.9 1_:___2 ___2

0

Dropput rates are based on information collected on 10th graders in 1940 and*.11ow-up analysis in 1963;.both miles and females are included.

2 The size of the population within the cell did notówarrint the calculation ofdmpout rates. -

Source: Unpublished data from Project -TALENT,--Reported in -Howard Vincent, "AnAnalysis of Vocational EducatiOn in Our Secondary Schools," Office of Program.Planning and Evaluation, U.S. Office of Education, July 21, 1967.

sOciodemographic backfround as well as in reasons' for choOsing a specific kind. ofedicalion. He may hakie a shorter time litUriztintitat is, he may value present

0economic gain more highly thin future econninic gain. This is an4mderstandablebehavior pattern for the disadvantaged person who is enrolled in vocational;courses. Likewise, the yocattnal student may value the money incomecof a jobmore highly and the' statin of rjob less higIdy thini would an academic graduate.Also, he is training for a jots in which he is to be employed immediately uponleavinghigh school.Wherilabor markets are tig& he does 'not niceisarily need tobe a high school graduate to get this job as long as he has thi skills necessary iomeet the occupational minimum. (There is /a direct relation between.; theunemPloyment rate and the high school retenlion 'rate which backs upthi;hypothesis). Thus, he miy, in Fleet, be able to fulfill thi fnirpose 'Of the

.

programplacement in a job dropping out of school. , In contras the.academic student Must have a hiei school diplopia in order to proCeed t ahenext step in hisnectipational.cireercollege attendance. Also; the ultimate I;the college-bound. student strives for is different from that for which thvocational graduate 'strives. One May be able to get such a job if he is a c011ege

. diopout, but not if he is a high school dropout. In short', the comnbination ofdifferent 'sOciodemographic ackground, different .'weights on incoMe vertils

:status, different time .horixonsand different constraints op job' entry (e.g., highschool diploma,pecessary/not necessary) Make it inevitable .tkat therewill be ahigher gioss drimout rate for vocational-technical educatioii: The phenomenon0 P

ia fact; built into the program.What the data in Table 14 suggest; therefore, is that the calendar time

, spent in a vocational-technical program may be too long. This pOssible excessivetime is actually the'reiult of a curriculum mix that is forced upon the students

TABLE 15 .

PRESENT VALUE OP EARNING STREAMS FOR MALES AGE 17BY OCCUPATION, YEARVIF SCHOOL COMPLETED, AND ETHNIC

LA li CAMP.' , Ir 'IJR. II nr. ILAN! IDLY 0 11 Pt I ZO, 1177U .

Whiteommmo

Nonwhite

. Experienced Civilian Labor ForceHS 4 .$38,384 $26,329HS 1-3 35 960 23 645

2,424 2,684.' Experienced Civilian Labor Force -

HS 4 . 39,0182HS 1-3 / 36 6012

2,417

. Professional, Technical and KindredHS 4 44,428 33,0373HS 1-3 42 448

1,980

. Designers and DraftsmenHS 4 ,HS 1-3

46,98046 41 1

.

5695. Farmers and Farm Managers . .

HS 4 ... 22,762HS 1-3 21507

1,255

. Managers, Officials and Proprietors'HS 4 45,941HS 1-3 0 45 869

72

. . Buyers and Department Store HeadsHS 4 46,049HS 1-3 - 46 891 ,

-842. Clerical and Kindred

HS 4 37,066 31,7 13HS 1,3 35 770 30 952

1,296 761. Bookkeep!rs

HS 4 35,902HS 1-3 . 35 065

.. 837

10. Shipping and Receiving Clerks ., HS 4 34,730 31,9882HS 1-3 35 689 36,2 13

-959-- -4,22511. A11 Other Clerical . ,

HS 4 _ -- -

HS 1-3 '

36,525 __,35 699 \

-28,767328 012

.. - 1,826 755

White Nonwhite12. Sales Workers

HS 4 $38,067 $28,2812HS 1-3 32 178

: 5,889_4642----0/

1 . Insurance, Brokers and UnderwritersHS 4 44,430HS 1-3 45 464

1,0341 . Craftsmen, Foremen and Kindred

HS 4 42,548 30,956

\HS 1-3 42 155 28 727

393 2,22915. Brickmasons, Stonemasons and Tile

HS 4 45,081HS 1-3 - 42 539

2,54216. Carpenters

HS 4 3&,449HS 1-3 38,624

---1773-

17. Compositors and TypesettersHS 4

, HS 1-3

,42,85944 979

.

2,12018. Electricians

HS 4 . 46,103HS 1- 3' 48 3582,255 ..

19. Lineman and Service. HS 4 46,889

HS 1-3 48,9922,033

20. Machinists ._HS 4 - . 43,707HS 1-3 44 187

2 . Mechanics and Repairmen- HS 4 . 38,816

HS 1-3 38 769'' ,/.1," 47

22. Airplane Mechanics and Repair, HS-4 45,0493

HS '1-3.

- . 45 149

23. Auto Mechanics and RepairHS 4 . 35,962

.

HS 1-3 36 428..

-76

t.a

White Non White2 . Painters, Construction and Maintenance

$35,51 1HS 1-3 32,925

-1,37g25. Plumbers and Ilpefitters

HS 4 46,446HS 1-3 45 427

1,10926. Toolmakers and Diemakers, Setters

HS 4HS1-3

52,8475321 1,

-38427. Operadves and Kindred

.. HS 4 37,576 27,167HS 1-3 36 821 26 5 19

765 64828. Truck and Tractor Drivers

H$ 4 37,502HS 1-3 3,8 997 .

--1,4-9529. Other Specified Operatives

tIS 4 . : 37,089 25,9073HS 1-3 35,256 25,729

1,833 -178Service Workers

.

HS 4.

HS 1-3 .

30,86027 431

21,24920 170 -

3,429 1,07931. Barbers

HS 4 33,622HS 1-3

--TA-2-3

32. Protective Service Workers11S4 41,895HS 1-3 40 453

1,44233. Other Service lilduding Households

HS 4 . 24,659 20,330HS 1-3. 22 720 19 754

1,939 1 57634. Farm Laborers & Foremen

18,693 1 1 ,6022HS 1-3 16,540 9 7842

2,153 1,818Notes:1 'The data presented were calculated as follows:

Return to 4 at age 17 = 'Y4 17'- + Y4 ig + Y4-64

(1+01

I

Return to 1-3 at age 17.= + y1-318

y1-364

IS

(1+01 0+02 (1+048

where .,

Y4 and Y1-3 = median earnings of those with 4 years of high school and0 1-3 years, respectively, subscripts refer to agg,

and

17= 0 by assumption.

AlsO

'Y18 = Y19 = = Y24

Y25 =. Y26 = Y34

and

. Y55 = Y56. 6 Y64 for Y4 and Y1-3,again, by issuMption,',..

= 10 percent

2 Age 18-24 and 2564 cohorts used.3 Age 55-64 cohort earnings estimated.

Source:

Stuart 0. Schweitzer, "Occupational Choice, High School Graduation, and Investment inHuman CapitaV Hearings of the Joint Economic Committee, Subwmmittee on Economyin Government, National Priorities,.1-18 June 1970.

bin which may not correspond to labor market realities or the needs andlong-term plans of students. The MDTA program purports to give a Man entrylevel skills after no more than 52 weeks or one calendar year of training. Whatpoint is there in dragging out the education for this equivalent goal to twocalendar years in high school? The dropout rate from vocational-technicaleducation might be reduced if the calendar time spent in high school were cid,by one year, from four to three years. As Table 15 shows, an additional calendaryear of high school is a detriment to preparation in a number Of occupations,such as electrician and machinist. One should note, also, that most of the

. occupations listed represent skill specialties offered in vocational-technicalschools.

Finally, further evidence to support 'this hypothesis is found 7n the studyon low achievers (defined as those potential inductees who failed to pass theArmed Forces Qualification Test) by -nansen, Weisbrod and Scankni (1970, H).The authors found that extra education, in the sense ofadditional, years ofschooling, was of much less value in improving earnings than were various typesof skill trair;ng learned outside Of school. in short, they. point up the crucial.difference between extended time in school, versus what one aztually learns inschool.

1

18.

i770

,

.4A

TABLE 16

IMPACT OF MANPOWER TRAINING .ON SELECTED. SOCIODEMOGRAPHIC GROUPS,

PERCENT OF TIME EMPLOYED IN 18-MONTH POST-TRAINING PERIOD

SampleSubgroup

.

DropoutDid Not.Report

Non-Trainee -

b

(s)

b(s) (s)

m

(s)

Total .

Sam . le .8.9* -25.2* -1 8.1* 60.3 -(n=854) (3.5) (5.6) (3.6) (30.9)lEduation .

D-8 -26.3* -47.6* -29.8* 46.3(n=207) (9.9) (12.5) (8.6) (38.6)9-11 -9.5 -164 --21.4* 59.9(n=254) (6.4) (10.1) (6.6) (35.8)12 -6.0 18.4* -10.9 68.0(n=320) (5.2) (8.7) (5.8) (35.1)13 and over -10.4 -4.0 -40.0* 675(n=73) (10.6) (28.8) (10.7) (32.5)

Alli30 or less -5.5 -14.7 -8.9 63.9(n=408) (4.7) (6.9) (5.0) (34.7)31-45 -9.6 , -302* -21-7e 59.0(n=321) (6.9) (10.1) (6.2) (38.3)46 and over , -42.0* .86.3* -36.9* 51.9(n=125) (123) (22.6) (1 13) (39.3)Sex .

Male -5,4 -25.0* -14.7* 63.8(n=618) (3.8) . (6.6) (4.5) (36.2)FeMale 0 -7.0 -21.7* -1 5.4* 50.9(n=236) (8.8) (10.2) (6.4) (37.4)

Age and Sex -

Mak -

30 or less -7.0 .19.1 12.0 68.8(n..-'3V8) (5.0) (8.1) (6.2) (33.4)31-45 :3.8 -30.6* -19.7* . 62.7(n=218) 1 7.3) (13.2) (8.3) (37.6)46 and over I -24.1 -68.5* -16.4 -.. 47.9(n=82) (21.3) (34.2) (21.9) (38.0)Female ii .

30 or less -0.6 -21.1 2.4 46.4(n=90) (14.2) (14.0) (9.2) (33.5) ,31-45 25.3 0 -17.4 -23.4*

, 51.3(n=103) (12.0) (15.2) 449.8) . '',.3.7)

a.

-

79

.45 and civet(n=43)

Prior LaborForte Experience

-97.5*(33.4)

-102,4(42.3)

-61.5*(25.1)

-

59.6(41.0)

NLF 47.5 -40.5* -23.3* 52.5(n=131) (9.9) (15.9) (9.0) (39.2)

UE 6 mos. or less -2.7 -23.0 -19.1* 67.9(n=173) (6.6) (13.9) (7.7) (33.1)UE over 6 months 14.6* -20.5* ;236* 46.8(n=242) (6.6) (11,8) (7.2) (38.4)EMI) 6 mos. or less 1.7 5.9 17.1 66.9(n=69) (i3.0) (22.2) (21.0) (34.4)EMI) over 6 months 0.9 -11,8 4.2 70.8(n=239) (7.8) (10,2) (7.6) (32.3)Eth* Origin .

White -9.2* -24.1* -18.1* 60.7(n=831) (3.5) (5.8) (3.6) (36.8)Marital StatusMarried -9.7 * -24.8* -17.0* 60.6(n=586) (4.2) , -----(7.1) (4.3) (38.0)Single -23.7* -16.3* 60.3(n=198) (7.6) (10.1) (7.6) (34.5)Widowed, Separatedor Divorced -6.2 -21.0 -1.6 57.2(n=70) (15.3) (27.6) (17.0) (35.3)

Notes: Significant at the .05 level or higher.b lathe partial regression coefficient.

(s) is the standard error of the partial regression coefficient or standard deviationof the mean.

m mean of the dependent variable.(n) is the number of observations in the interaction subsample.

-In addition to what is shown in the table for each int( raction equation:1. All equations contain a labor market variable in dummy form.2. All equrtions contain a job placement variable in f summy form.3. No equic.3n contains the ago-sex interaction sets4. Age and age squared are in equations containing an edue.ation variable..5. Education and education squared ate in the eqrations containing an age variable.6. All equations contain a mobility variable in chunmy form.7. The regreuor for the "Reject" training status is included in the model but omiited in the

table.

Thils, for example, for all respondents who have only 0-8 years of education, the followingindependent variables are used to explain earnings and employment: age and age squared;sex for male and female; labor market area for Charleston and Huntington, and McDowell,

c-Monongalia, and Harrison counties; prior labor force experience-NLF. 6 mos. or less, NLF°over 6 mos., UE 6 mos. or less, UE over 6 mos., EMP 6. mos. or less rind EMP over 6 mOs.;placement effort-metal working skills and 4raining for a specific company; maritalstatumarried, singia, widowed and separated or divorced; race-white and nonwhite;mobility-whether or not a person moved any distance at all; and, rmally, the training status_variable as shoWn. Thusa the same equation is run for the respondents who are 30 years oldor less, except that education ind education squared are substituted for the ago-variable. .

80

A

The values of the partial regression coefficients ;for the training status variable aredifferences from the trainee regressor which enters the intercept term. Thus, thosenontrainees who had only 0-8 years of education werf employed 29.9 percentage points lesstime than the trainees over the 18-month post-retraining period.

Source:

Unpublished data, West Virginia Retraining Research Project, Department of Economics,University of Wisconsin, Madison, Wisconsin.

Effects of Manpower Training on Sociodemographic Groups. Tables 16and 17 give a detailed analysis of the effects of Ares Redevelopment Training onseveral crucial sociodemographic groups in terms of earnings and employment..The pattern of net effects of retraining among the independent variables isdifferent in terms of size, sign, and statistical significance for the two dependentvariables. The independent variables generally explain more of the variation inthe earnings variable than they do for the employment variable. The coefficientsof determination as a group tend to be significant at a higher level of statisticalsignificance with respect to the earnings compared to the employment variable.

Employment. Major interest in this study lies in the size, sign, andstatistical significance of the partial regression coefficient of the nontrainee

, regressor of the training status variable. If retraining is effective, nontraineesshould have less favorable employment and earnings experience than treinees.Therefore, the sign of the partial regression coefficient of the nontraineeregressor should be negative since the nontrainee experience is-interpreted interms of its deviation from the trainee experience. Retraining should raise themarginal productivity of the trainee, other things equal, and hence, trahseesshould have higher employment or earnings than nontrainees.4 With respect toem loyment, this is generally the case. Exclusive of the results for the totalsample, 16 of the equations in Table 16 show that retraining has a positiveand statistically significant effect on post-training employment. Eight equationsshow no statistical difference, though the signs are negative but for two cases.

Of interest is the suggestion in the data that retraining has been of greaterhelp to those persons with greater labor mirket disabilities than to those personswho have higher labor market qualifications. Thus, for those with only 0-8 yearsèf education, trainees are employed 29.9 percentagi points more or 5.4 mooths(18 x 29.9) longer than nontrainees. For those with 9-11 years education thedifference is 21.4 percentage points, or 3.8 months. And for those with 12 yearseducation the, net employment difference between trainees and nontraineesdrops to 10.9 percentage points, or 1.9 months. This is not entirely unexpectedsince it is theoretically reasonable to assume that a given amount of retrainingwill have I% larger impact on marginal productivity for a person with a low stockof skill or acquired ability than it will for a person with a high stock of skill. The13 and over education group represents an .abberation from this trend. Apossible reason for this abberation could be that the trainees with 11 or moreyears of education are still attending ccIlege while the nontrainee counterparts

4 For a given level of wages, eMployment will increase if productivity increases since this willincrease the demand for labor. For a given level of employment, wages will increase if productivity increases employment because the rise in productvity, other tMnp equal, will in-crease the demand.for labor.

-Al

are not. College students may be using the training process itself as a means toacquire part-time employmen t. However, the trend holds up for the age and theprior labor force experience subgroups. For those who are employed prior to thetime when retraining began, there is no statistically significant post-trainiiigemployment difference between trainees and nontrainees. For males or feinalesunder 30 years of age there is no statistical ,difference between trainees andnotitrainees. These results suggest that, if a goal of retraining is to increaseemployment, retraining efforts should be concentrated on those persons withlabor market disabilities the older, the less educated and the unemployed orthe non-labor force participant. Theyounger, more educated persons who have ahistory of very recent or con§istent employment tend to benefit very little fromretraining in terms of net incieases in employment.

The pattern of statistical significance for dropouts is of interest since thosedropouts who suffer the more severe labor market disabilities such as loweducation, advanced age or long-term unemployment, fare significantly worse .than the trainees. However, for most .other sample subgroups, there is nostatistically significant difference in employment between trainees and dropouts.This phenomenon does not mean that completion of retraining was of littlebenefit to the trainees, since dropouts had, on the average, completed aboutone-third of their retraining before dropping out and often took jobs in relatedtraining areas. However, one implication might be that, on the average, certaingroups of the unemployed need less training to prepare them for employment ina given job *an was being offered to those groups in the retraining programs. Ofcourse, this it the theme underlying the discussion of dropout behavior above.

Effects on Earnings. The advantages of retraining are less dramatic withrespect to increasing earnings than they are with respect to increasingemployment, if one of the major objectives .of retraining is .to significantlyincrease earnings, then, for the total sample, this objective is achieved. For thetotaLsample, trainees earn $43 per month more over the 18-month period thando nontrainees. The difference is significant at the .01 level of significance.However, the picture differs among the 24 sample subgroups. For these 24subgroups, only nirie of the differences in earnings between trainees andnontrainees are statistically significant. Four of the coefficient's for thenontrainees have a positive sign, implying negative benefits to retraining; but thecoefficients are not statistically significant, and so there is, in effect, nodifference in tarnings between the trainees and the nontrainees for those groups.Also, the lack of statistical significance for other groups, even though theregression coefficient has the appropriate negative sign, indicates no differencebetween trainees and nontrainees.

Again, there is a tendency, though not nearly as clear cut as withemployment, for retiaining to have the greatest absolute impaCt on earnings forthose with the more severe labor market disabilities. Trainees with 0-8 and 9-11-years of education, eam significantly more than nontrainees in the sameeducationatsubgroups but there is no statistical difference in 'earnings betweentrainees and nontrainees for 12 and 13 years or over of education. For thosewho were not in the labor force or who were unemployed for over six monthsbefore retraining began, retraining has a significant effect. But for those who

82

<,

TABLE 17IMPACT OF TRAINING ON SELECTED

SOCIODEMOGRAPH1C GROUPSTOTAL 18-MONTH POST-TRAINING EARNINGS, IN DOLLARS

Sample

IlligreluP

Dropout Did notReport

Non-Trainee

b

(s)

b(s)

b(s)

..

m

(s)

TotalSample -651* -1297* -782* 3199n=(854) (226) (3 59) (229) (2628)Education0-8 -1 137 -2318* -1 1 37* 2509(n=207) (668) (8 2 5) (583) (2562)9-11 -514 .........009 -1290* 3005(n=254) (404) (639) ' (42 0) (2592)1 2 a -677* -1032* -99 3614 -(11=320) (298) (502) (330) (2532)1 3 and over -1 129 3406 -742 4010(n=73) (918) (2488) (91 9) (2837)

Age30 or less I -978* -970* -534 3402(n=408) . (283) (4 16) (299) 12452)31-45 -119 -1266 4081* 3272(n=321) (425) (706) (43 3) (2893)46 and over(n=125)

. ,l'

-1 529*(774) '

-3598*(1398)-

-595(71 3)

2348(200)

SexMale -612* -1582* -800* 3868(n=618) - (280) , (484) (329) (2695) s

Feinale -343 -746 -273 1446(n=236) (331) . (385) (242) (1319)

Age and SexMale30 Or less -1 122 -1394* -745' 3985(n=318) (326) (53 5) (410) (2404)31-45 110 -1458 -979 4908(n=218) (586) (1059) : (666) (3039)._45 and.over -864 -3582 ll. 162 2808(n=82) (I 520) (2436) (1560) (2572)Female30 or less _ -90 -972 - 1 16 .1344

(n=90). (5 1 8) (509) (334)31°45 -262 -353(n=103) (502) ;(638) (409)46 and over -3062* -1 324 °

(n=43) Mf9) (1293) (768)Prior Labor ForceExperienceNEP -757 -1584 -758*(n=.1 31) (485) (780) (442)UE 6 mos or less -81 6 -167 -805(n=1 73) - (440) (9 6) (5 1 8)UE over 6 mos -934* 90 -1348*(n=242) (399) 71 8) (434)EMP 6 mos or less -173 1494 149(n=69) (1 779) (171 8) (1 628)EMP over 6 mos 31 8 -541 -62(n=239) (555) (726) (541)RaceWhite -689* 1244* -8 1 6*'(n=831) (226) (375) (228)

Marital StatusMarried -337 / -1 110* -664*(n=586) (277) (468) (282)Single -1 364 -1443* -1 1 29*(n=1 98) (450) (597) (45 3)Widowed , Separatedor Divorced 175 -925' 1 064(n=70) (794) (1430) (879)

Note: The notes of Table 16 aU apply to Table 17.

(1192)1524

(14,38)

1472(1288)

2034(2026)4004

(2550)2433

(2534)3825

(3064)3844

(2504)

32201(261 7)

3418(2762)2942

(2332)

120911 (1 822)

were employed before 'retraining or who were unemployed only six months orless, there is no *statistical difference: in earnings between the trainees andnontrainees. In short, while 'retraining pays well foi the study sample in theaggregate, important subgrom)vithin the sample have tended not to benefit atall on the average. These retalfi; of course, apply only to the specific saMple inthis study ?slid refer to a very specific locale and time so that generalizationsbeyond this sample are risky if riot inappropriate.

These reiults also highlight an important policy issue concerning theprivate and social objectives of retraining. If the objectiie of retraining is simplytO get people back to wotk, regardless of the wage tate they receive, then, on theaverage, even for rather specific categories of people, retraining can, be judged asuccess. for this sample. Rut, if the fiat1 of retraining is to yield a net moneybenefit over money costs, then retraininglr not been very effective, fer, certain

Ind 12

TABLE 18

ES11MATED NET EXPECTED INTERNAL RATES OFRETURN TO SELECTED TRAINEE SUIltROUPS

Trainee SubgroupAverage!

CostsDifferencein Average

Benefits

Social

Rite of Retarn4

Education .

74255$515591

656548525

789

21

789789789

401401401

.

321696590628687

576

599688

334

.

.

758860662

4952

35627203972

5331822

497653

-108

.

.77z,..,

32328822

5055732898.992412

544

- 443752

-71:12

\

.

98.91529

4.8794

,

46.1129.3 .

70.3

61.4

,35.4

56.2 L782 ..

- a

3

50.3219.3

1

155.9 4----;'777

150.7

_-__3__

90.7

.

168.5101.4

J.

0-8

9-11.1213 and over . a

AK30 or less31-45

-, 46 and over

SexMaleFemale

Sex and Age

Male.30 or less31-45

46 and over. .

Female

30 or less -

.31-45

46 and over,

Prior Labor Force.Expeilence

NFLDE 6,awaths or lessDE over 6 monthsEMP 6 months or lessEMP over 6 months

Race .

White

Marital StatusMarried -Single

. Widowed, Separated,Divorced

1 These isminp diffeientials ate the 18 month benefit diffe4tids adjested to a12 month basis:*

ett,

1

,17's

, .- ..2 'Not statistically significant.3 Negative rates of return are implied.4 A 10 year post-training benefit period is assumed. .

Source: Costs: Glen G. Cain and Ernst W. Stromsdorfer, "An Economic Evaluation ofGovernment Retraining Programs in West Virginia." in Gerald G. Somers, (ed.),Retraining the Unemployed MadisOn, Wisconsin, 1968. Table 1X.2, p. 313.

Renefitt Unpublished date West Virginia Retraining project.

tsan1i4stgroups gain little or no net monetary benefit on the average. Twoimplicions arise. First, perhaps some of these subgroups simply did not requireretraim or required less of it or different kinds than that which was offered.Second, erale may fail arid general disenchantment may result among thosesubgroups who do not benefit in money terms. Such an effect could threaten. the CsoCial and political commitment' of society to the entire retraining program and/could deprive of retraining those subgroups which can receive some benefit fromthe programs.

Investment Effects. The earnings data in Table 17 can be combined withcost data .from Cain and Stromsdorfer (1969, Ill) to provide estimates-of thesocial rates of return for these various groups. Table 18 displays the results. Forthose uses where the benefits are statistically significant, the rates are generallyvery high. In addition, the subgroups with the greatest labor market disabilitiestend Jo benefit the most. However, no female subgroup has benefit differentials

whic Neire statisdcally significant. For those _benefit differentials which lackstatist al significance, the rate of return is effectively zero. Were it true that thenegative-benefits-for ,certain subgroups sucl, as single trainees were statisticallysignificant; then'the internal rates here would be negative. Since they are notstatistically significant, the rate of return for tfiese'subgroups is effectively zero,too. ' _

_ When benefits are statistically significant, they equal the cost outlays in ayear or two, well within the data estimation pgriod of this study. Thus, even ifconsiderable umierestimation of money costs exists, which is not the case,adjustMent. for diis cost underestimation would still leave the rates Vry higfi.The same holds true for potential overestimation of benefits due 'to the/act thatbenefits are assumed to accrue by a constant amount for 10 vats aftrr training.-

The results from the Sewell study generally corroborate the findings in theWest Virginia Study. As Table 19 shovtipr both on-the-job and institutionaltraining, blacks benefit more dian whites in terms oreamings; those .over 44

1 . years benefit more than those under 21 y9irs. And farMeris and sharecroppers1benefit even more than those pehons ,fineMployed over<ix months. For'on-the-job training, thoSe with 0-8 years of education earn 13,dol1ars more perweek than thfir control gilups, and surpasi those with 9-11 grades of education

.... by two dollars per week, while the situatiski is reverse with respect to.instifutional training.. ," .11.

flii:ally, Table 20 brings-additional evidence to bear on these issues basedon thel Hardin and Borus study (1969, III). In additiOn it shows° the effect ofdifferent lengths a training. WhereSs the West Ihrginia study shows trainees

- earning about $780 more than nontraMees in an 18 month period after training,the Hardin and Borus study, shows irainees 1 skorter (60-200 class hours)t

..... -

86

TABLE 19 ,,IMPACT OF MANPOWER TRAINING ON WEEKJ.Y EARNINGS AND

HOURS WORKED PER WEEK FOR SELECTED DISADVANTAGED GROUPS

'

Sample Subgroup- -

Weekly earnings (dglism) Hours worked per week

On-the-jobTraining1

Mai.tutional2

On-the-jobTraining

Insti-tutional

Labor Force experienceprior to trainingFarmer (96) 17.8** 14.7** 4.1* 8.1**Unemployed over

6 months (37) 10.1* 8.4 4.3 -0.5 ,

ColorNonwhite (416) 8.9 7.5** 2.1* .2.1

Age as of 1 Jan., 196644 years & over (95) 1 7** 7.8 6.2** 73*Under. 2 1 years (79) 0.0 -8.8 2.6 0.9

Education0-8 grades (224) I 3.0** 7.0* , 4.8* , 4.0*9-11 grades-(l40)T ---1-110*-11-- llitut 3.8* 0.2

Notes: * Significant at the .0.5 level:** Significant at the .01 level.1 Net regression copfilcient for on-the-job training.2 Net regression coefficient for institutional training.

Nurnberg in parenthesis are the sample sies.

Source: David 0. Sewell, Training the Poor, KingstorL Onturio, 1971, Table 8, p. 70.

-training progrnms earning about $980 in a 12 'month period, or about twice themonthly benefits. However, the patterns of teturn respec, increasinglevels Of education is. similar, more educated people ben tles:-..biltr mightexpectkitt those who were previously. on Welfari benefit more than thosewho were not prior to the time training began. Again, it is inter sting4to note thegenerally negative benefits gained by those groups other than nonwhites and -

those whites of 5-8 yeais education who stayed in the training program for over.200 hours of instruction.

. .

Effects of Vocational and Manpower Training by Program Area' and Skill

There are very few good studies of the differential earnings benefits ofvocational program areas or pccupational skills. There are no useful benefit-costanalyses by skills, mainly because of the difficulty of collecting costidata by

4° skill. It is difficult enough to collect on the basis of broad curriculurplateas or

87

1-

TABLE 20SOCIAL ECONOMIC BENEFITS OF TRAINING BY,

SOCIOECONOMIC GROUPS FOR VARYING:LENGTHS OF TRAINING, IN DOLLARS

.° Socioeconomic Group Clumoom noon Per Trainee

64200 201-1920

, . .Whole SanipliWhite Men ..

White WomenNonwhite MenNonwhite *Millen ,

White, 54 Years Education" White, 9-11 Years Education

White, 12 years & over EducationNoneihite,. 5-8 Years EduCation.Nonwhite, 9-11 Years Education _

Nonwhite, 12 Years & Over EdncationPrior Welfare:

IYes I, r

NoTraining Occupation:

FactoryHealth Care -.

OfficeAuto Repsir.Other

,

,

S 976670998

1,3851;3121,002

829750

1,330. 1,293

1,359

1,049959

8451,140

1

negative-3

-129-247

-60305171

-1512

2,0011,553

-44-61

-15-251

-9810

-123716

-1060

Notes;

Source:

1 No. 60-200 hour classes were intended for office occupations.2 there is only one obiervadon for this albgroup.-Adapted from Hardin, Einar and Bores, Michael E., Economic Benefits and Costsof Retraining. (Lexington, Mau.: (Lexington Books, D.C., Heath and Co. 1971),'Chapter 12, Tables 12-2, p. 144; 12-3, p. 146; 12-4, p. 147; 12-5, p. 149;1 2-6, p.151; and 12-7, p. 152.

7

even for a yocational school if this school is part of a larger school diMrict suchthat accounting records are aggregated. To be hue, there are separatebenefit-cost studies o( given skills, in a case study context; honever, the widediversity of methodologies, with the fact that the, studies do not contain all thenecessary data or information on their. methodologiei, , Makes- it extremely.difficult to. make valid comparisons.among skilhs

. .

Secondary Vocational Course,. The Hu, et al. study (1969,11) providesearninp and employment comparisons among 12 different vocational skillsoffered in vocational-technical programs in three 'northern cities. Usingregreuion analysis' to stadardize for the effects of such factors as sex, age, IQ,race, marital status sod father's .education, it was found that there- was nostatistically ignescent difference among the skills when compared with the

88 SG

TABLE 21 .

PERCENT OF TIME EMPLOYED IN PERCENTAGE poem ANDAVERAGE BEFORE TAX MONTHLY EARNINGS Fog NON-COLLEGE

ATTENDING VOCATIONAL-TECHNICAL HIGH SCHOOL GRADUATES '

WadableAvenge isSix Yesn

First Year MeeGraduation

Sixth Yoe AfinGreenling

(1). b (s) Ili (s)

Courses'n=565

Percent of TimeEmployid

Commercial2 I

Food Service . -5.8 (7.3) -13.6 (8.3) -4.4 (11 .8)Building Trades .

,

Occupations -3.3 (12.6) 15.5 (14.2) -1g.6 (20.4)Mechanical and Repair -2.2 (6.4) -8.4 (7.2) 2.2 (10.3)Tool Deiiin 1.4 (6.5) 0.3 (7.3) 8.2 (10.5)Wood WorkingOccupations 2.7 (8.6) -7.8 (9.7) 9.0 (I:: .9)Electrical and ,

Electronics 2.8 (6.6) -6.8 (7.5) 1.9 (10.7)Aghiculture & Horti-culture , 4.5 (17.2) -18.5 (19.3) -0.8 (27.7)Professional Occupations -2.4 (4.6) 4,6 , (5.2) . 5.9 ( 7.4)Distributive Education -6.0 (10.6) 16.2 (11.9) -22.1 (17.1)Personal Servias .10.8 (5.5) 4.4 (6.2) -0.2 ( 8.8) -aOthing and Fabrics -8.8 (7.4) 6.5 (8.3) -21.6 (11.9)

Average Before-Tax Monthly Earninp. ----.

Commercial2 ,

FoOd Service -21 (44) -70 (41) :-16 (59)Building TradesOccupations -87 (76) -36 (71) -166 (101)Mechanical and Repair 4 (38) -58 (36) 45 (51)Tool Design 98* (38) 94" (36) 100. (52)Wood Working Occupa-tions 51 (52) -21 (48) 69 (69)Electrical and Electronics 8 (40) -0 (37) 8 (53)Agriculture & .

Horticultuie 34 (103) -108 (96) 140 (138)Profesgonal OccuPations 26 (28) -30 (26). 79 (37)Distributive-Education - -49 (64) 3 (60) -71 (85)Personal Services . -85** (33) -75.- --(31)- -84. (44)Clothing and Fabrics -23 (44) 11 (41) -49 (59)

, Notes: Significant at the .05 level.** Significant at the .01 level.

. b is the partial regression coefficient.(s) is the standard error of the partial regression 'coefficient.

I The i,..--gression model controls for the influence of labor market at iime of graduation, sex, lQ,race '. marital stat us and father's education.

2 This regressor of the 'violable enters into the intercept term. The other regressors of the variableare interpreted as deviations from this regressor. Thus, over the six year period after.graduation,,persons trained in food service occupations were employed 5.8 percentage points less than thosewho were trained in commerical occupations. However, this difference is n^.1 statistically signi-ficant .

Source: Tebwei u. et at., A Cott-Effectivenets Study of Vocational Education.University Pak, Pennsylvania, 1969.

commercial courses, 'either in the first, or sixth year after high school graduationas well as over ihe six year .average; (See Table 21.) Tool design yieldsapproximately $100 a month snore than Commercial cburses during the first,.sixth, and six year average after high school graduation. Personal services coursessuch as beautician Courses, earn from $75 to $65 less, as shown in Table 21.

However, if sonie course other than commercial were taken as the standardof comparison, a different pattern of earnings and employment differentialswould emerge, so the evidence in this and similar tables in this piesentafion ismit always conclusive.' A slightly different portrayal of the data could give adifferent set of impressions. However, the general lack of statistical significancein Table. 21 is striking and might well persist if other skills were used as al-.comparison, especially with respect to the employment variable:: .

. Table 22 contrasts .the . earnings and employment of the six vocationalprogram areas against the academic curriculum. Thus, the average hourly startingwage rate for distributive education ,graduates was 21 Cents less than that ofstudents who graduated from the academic curriculum.

Likewise, graduates from the technical program earned 23 cents more ontheir first job than did the -academic graduates. Up to three years later, officegraduate's were earning 22 cents more per hour on their current or last job; tradeand indUstryt 19 cents more and technical 26 cents &ore. But there was nodifference betwom the hourly earnings of the academic graduates and the

- distributive, health ur ':iticulture graduates.. .Unlike the results shown in Table 21, there are marked differences in

employment level, with, for instance, office graduates experiencing 14.1percentage points more of employment than academic graduates. These resultswould be more interesting, however, had they been shown by sex and possibly,ethnic origin.

Junior College and the Vocational hymn Areas. Table 23 shows asomewhat similar comparison for junior college graduates, although hele thecontrast is between the technical program area and the five other . areas.Curiously3.-the health occupations group earned about 34 cents more per hout onits starting wage rite than did thelithiikal-graduatesThe-Hardin.and Borns

90

2

study (1969, 111) also shows health occupations achieving earnings greater thanthose of factory workers. (See Table 20.) In addition, those in the agricultureprogram earned 39 cents per how- less than technical graduates with respect totheir current wage rate, a direction of effect which is not unreasonable. There isno statistically significant difference. in current way rate betweez technicaleducation and . distributive, 'office or trade and industry. WIAe thh lack of'cliffereace between the current hourly wage rates of technical and trade and.industry is not unreasonable, it does not seem likely that there should be nodifferential between technical and office or distributive education where there isa high concentration of females: This is so even in light of the fact that Sharpand Myint (1970, II) control' for sex in in ',their regression .modeCsinceinteraction effects between sex and the various programs have not beenaccounted for in the model.

Tables 24, 25, and 26 show the interaction effecte between level ofeducation /mid vocational program area. The striking fact about these results isthe general lack of statistical significance between educatiOnal levels foreach of

a

:TABLE 21REGRESSION ANALYSIS OF STARTING WAGE; CURRENT WAGE;

AVERAGE MONTHLY EARNINGS AND PERCENT OF TIME EMPLOYED,BY PROGRAM AREA, FOR SECONDARY VOCATIONAL GRADUATES

harmStarting WayCisteist

wi1224Wage

n11224BON.,IPS 1 9

Percent of timeaa1028

b3 ()3 b (a) b (s) b :-(a)

Academic2 .

Office 1.17" .06 .22" .07 77** 19 14.1** 2.9Trade & Industry .05 .07 .19* .08. :26 21'. 9.4** 3.3Distributive .21" .08 .09 .09 .37 24 1 1.3** 39Health .17 .17 .26 :19 58 45 14.9 8.2Agriculture .04 .09 .002 .1 1 7 30 5.3 4.5Technical .2 3** .09. .2e .1 1 68* 31 6.5 4.5

*Minificant at the .05 level of sipificance.*Significant at the .01 k'vel of significance.

All significance tests are twotailed tests.Notes:

.1The wage rate regression equations also control for the effects of:Neon, father's education, sex,Jrsarital status, socioeconomic status, race and additional education after graduation. The earninpequation ciiitrols for all the above except .socioeconomic status and additional education afterpaduation. The employment equation controls for all of the ibove except socioeconomic status.

Amu variable against which the others are compared. Thus, students in office occupations had astarting wee 17 cents higher than students in the academic curriculum.

3b lithe partial regression coefficient. (s) is the standard error of ele partial regression coefficient.

Source:Susan Fernbach end Gerald G. Somei s, An Analysis of the Ecomotnk Benefits of Vocational

Studies in Vocational end Technksi Education, Madison, Wisconsin, 1970. Tables 5.3,5.6, and 5.9.

91

TABLE 23RELA1WE IIGURLY WAGE RATES, BY PROGRAM AREA, .

. FOR JUNIOR COLLEGE GRADUATES

Starting-Wage Rate Comet Way Rate

bi- Pr Orem. pre68

. .

hill ilReredosCoefficient

Significance Pad IIRemotionCoefficient

Siyif Icaoce,

Technical1HealthDistributive EducationAgricultureOfficeTrade ind Industry

. .344-.177-.075-.024.001

.0099 .

.1251

.4687

.8411

.9940

.

.347'.031

-.392-.186-.073

.0829,.8586.0123.3048 -.6568

.

Motes:. 1ln addition, this regrealion model controls for the effects of relatedness of work to training;

additional education, sex, repon, age, marital status, race, father's occupation, father's education;and socioeconomic status. Since technicel occuptions are contraeted 'apinst each of theotherskills, the table shows, for instance, that health occupations earned 34.4 cents more for theirstarting wage rate than did technical occupations.

Source:Lame M. Sharp and Thelma .Myint, Graduates of Vocatiorsal-Teduskal Programs

atikies,.Wathington, DC., September, 1970. Tables 5.33 and 5.34.

G.

the program areas. Thus, it is only in health occupations that the junior vollegelevel earns a higher last or current wage rate than vocational high school. There isrio difference in first, last DI current wage rate or earnings between vOcationalhigh school and post-vocational high school education. However, this result issupported also by' the study of Corazzini (1966, II) for limited occupations inthe trade and. industry program area. Clearly, since the sample sizes for theprogram areas are relatively large, this pattern of results is of great importance iffurther research 'substantiates the results. Since Morsch (1970, II) found that forgiven skill areas there was little difference in cost between post-vocational highschool and junior college, yet both !evels of education cost more than vocationalor technical high school education, then the'efficacy of these two extra years of .

education istcalled into question. ,

One study is insufficient evidence on which to make far-reaching policyconclusions of an educational institution such as the junior college, communitycollege or post-secondary school system. First, we are talking only in terms ofmoney returns. Next, the study, although nationwide in scope, has problemsfrom: the standpoint of non-response bias. Finally, when program areas areaggregated, junior college yields a significant return over vocational high school,thouet the benefit of post-secondary school over vocational or technicalseocndarY education is less clear cut and may be noti-existent.

Tables 25 and 26 also show the interution between vocational programareas and education levels, but in this case, office. education is compared against

92

TA

BL

E 2

E.

*RE

GR

ES

SIO

N A

NA

LYS

IS O

F W

AG

E R

AT

E O

N F

IRS

T M

k W

AG

E R

AT

EG

i4LA

ST

OR

CU

RR

EN

T J

OB

,A

ND

AV

ER

AG

E M

ON

TH

LY E

AR

NIN

GS

, BY

ED

UC

AT

ION

LE

VE

L, F

OR

SU

AR

AT

E R

EG

RE

SS

ION

S B

Y P

RO

GR

AM

iRE

A

I

'

,

Edi

catio

nalL

evel

i,

Tra

sk n

odW

intr

y'D

istr

ibut

ive

.-

"I.

--H

eal&

Api

cultu

re . .Tec

hnic

al

4)-

(s)

-:: (

s)(s

)b

(s)

b(s

)

Wag

e R

ite, F

irst

Job

2.Po

st-V

ocat

iona

l Hig

h Sc

hool

Juni

or C

olle

ge

Wag

e R

ate,

last

or

Cur

rent

Job

Post

-Voc

atio

nal H

igh

Scho

olJu

nior

Col

legi

i

..

Ave

rage

Mon

thly

Ear

ning

sPo

st-V

ocat

iona

l Hig

h Sc

hool

.

.

.n=

495°

.01

.08

-.10

.13

n=49

5.0

1.1

2-.

03.2

0n=

428

.

+3 +

n=30

0-.

04.1

3:1

8.2

1.

c,n=

300

.001

.19

.41

.30

n=25

6, +

-+

-

. n=

88.9

1* .3

9 .

.12

.51

-

n=88

.02

.45

.78

' .61

n=73 -

n=19

7.2

2 , .

18.6

7**

.16

n=19

7.0

7.2

9.8

5" .2

7n=

175

-

+ 191

.

.

n 82=

-.09

.39

-.28

36

n=82

-14

.56

;..60

.52

1161

3

+-

+

.c,,, \

n=36

2.0

7.1

2_

.09

.14

'n=

362

-.02

.16

-.0

7 -

.19

_n=

337

+G

+

Nili

n:Sn

itifi

cant

at t

he .0

5 le

vel o

f ilf

icen

ce, t

wo-

taile

d te

st.

."."

Sipi

fice

nt a

t the

.01

keel

of

sign

ific

ance

, tw

o-ta

ded

test

..

bis

the

pane

l rep

esei

on c

oeff

icie

nt.

.(s

)is

the

stan

dard

Met

of

the

part

ial r

egre

ssio

n co

effi

cien

t.-

1 T

he r

egre

ssio

n M

odel

s fi

n fi

rst,

Ian

or c

urre

nt w

ay r

ate

cont

rol f

or r

egio

n, r

elat

edne

ss o

f jo

b to

trai

ning

, soc

ioec

onom

ic s

tatu

s, f

athe

r's e

duca

tion,

sex

, age

,mar

-ita

l sta

tus,

rac

e, u

rbes

-rur

al s

ettin

g ai

d gr

ads

poin

t ave

rage

.The

mod

ei is

the

sam

e fo

r av

erag

e m

onth

ly e

amin

p ex

cept

that

rel

ated

ness

of

job

to tr

aini

ng is

ex-

-pr

esse

d fo

r bo

th f

irst

End

s:w

rest

or

last

job.

_

2 Se

cond

ary

tOca

tionM

Psdu

atss

we

the

grou

p ag

ains

t whi

ch th

ese

coef

fici

ents

are

to b

e co

mpa

red.

Thu

s, th

ose

offi

ce o

ccup

atio

n gr

adua

tes

from

pos

t-se

cond

ary

voca

tiosa

ticho

ol e

arne

d on

e ce

nt e

s ho

ur le

ss-i

ns !

fun_

fis

t job

s th

en d

id O

ffic

e oc

cups

tion

grad

uate

s fr

om m

cond

ary

high

who

& ;

-.

3 +

indi

cate

s th

at th

e pa

rtia

l reg

ress

ion

coef

fici

ent i

s po

ntiv

e bu

t not

sta

tistic

ally

sig

nifi

cant

; - in

dica

tes

that

the

part

ial r

egre

wio

n co

effi

cien

t is

nega

tiveb

ut n

ot

stat

istic

ally

sig

aiic

ant.

-.

.

Sour

ce: G

eral

d G

. Som

ers,

et e

l., T

he E

ffec

tiven

ess

of V

oost

iow

el e

nd T

ecim

icel

ItO

gren

u: A

Nat

iona

l Fol

low

-up

Surv

ey, M

adis

on, W

isco

nsin

, 197

1. T

able

V1.

13,

Tab

le V

1.18

, App

endi

x T

able

23.

..

.

...

TA

BL

E 2

5R

EG

RE

SSIO

N A

NA

LY

SIS

OF

WA

GE

RA

TE

ON

'FIR

ST J

OB

, WA

GE

RA

TE

ON

CU

RR

EN

T O

R L

AST

JO

B-

AN

D A

VE

RA

GE

MO

NT

HL

Y E

AR

NIN

GS;

SY

PR

OG

RA

M A

RE

A, B

Y E

DU

CA

TIO

N L

EV

EL

&lo

catio

n L

evel

l

Tra

de a

ndIn

dusb

y3D

istr

ibut

ive

_Hea

lthA

gric

ultu

reT

echn

ical

-

(s)

,.(s

)b

(s)

b(s

)b

(s)

..

.

Wag

e R

ate,

Fir

st J

ob.

Voc

atio

nal H

igh

Scho

ol.0

9.1

0-.

21.1

1-.

04.1

9,.

.01

.13

.13

. .10

n=57

2'

.

Post

-'oca

tiona

l Hig

hSc

hooi

.26*

*.1

0.2

9*.1

3.1

5.0

9-.

10.3

033

**.1

0

n=72

$-

..

Juni

or C

olle

gen=

321

.'

.13

.20

.23

:21

43"

.15

-.02

.17

.13

.15

Wag

e R

ate,

Cur

rent

or

Las

t Job

4V

ocat

iona

l Hig

h Sc

hool

.20

.15

-.20

16.2

1.2

8.0

6.1

8.1

6.1

5Po

st-V

ocat

iona

l Hig

h Sc

hool

,.4

2".1

3_4

3".1

7.1

4.2

1.

-.16

.40

.36*

*°.1

3Ju

nior

Col

lege

-.12

33.2

734

.27

,.24

..

-A 5

'.-

---.

28.0

2.2

5

Ave

rage

Mon

tbijf

Ear

ning

sV

ocat

iona

l Hig

h Sc

hool

n=43

5+

2.

-

__

Post

-Voc

atio

nal H

igh

Scho

oln=

647

--

69*

++

Juni

or C

oile

ge+

+--

+

Not

es:

Sign

ific

ant a

t the

.05

leve

l of

sign

ific

ance

, tw

o-ta

iled

test

.Si

gnif

fant

at t

he .0

1 le

vel o

f si

gnif

ican

ce, t

wo-

taile

d te

st.

b is

the

part

ial r

egre

ssio

n co

effi

cien

t.M

is th

e st

anda

rd e

rror

of

the

part

ial r

epes

sion

coe

ffic

ient

ou...

a

-,

IE

d*. r

egre

ssio

n m

odel

con

trol

s fo

r gr

adua

te o

r dr

opou

t sta

tus,

reg

ion,

soc

icie

dono

mic

sta

tus,

rel

ated

ness

of j

ob to

trai

ning

, fat

her's

edu

mtio

n, s

cx, a

ge, m

alita

l

MIL

O, r

ace,

aid

ed e

duca

tion,

urb

an-r

ural

set

ting,

and

gra

de p

oint

ave

rage

..

2 +

indi

cate

s th

e pa

rtia

l reg

ress

ion

coef

ficie

nt is

pos

itive

but

not

sty

listic

ally

sig

nifim

nt; -

indi

cate

s th

e pa

rtia

l reg

ress

ion

coef

ficie

nt is

neg

ativ

e bu

t not

sta

tisti-

cony

siig

nifie

nnt.

3 F

oe a

ny e

duca

tiona

l lev

el, e

ach

of th

e po

ps=

are

as is

con

tras

ted

with

Offi

ce E

duca

tion.

Thu

s, T

rade

and

Indu

stry

gra

duat

es in

voc

atio

nal h

igh

scho

ol e

arne

dni

ne c

enti

an h

our

mor

e on

thei

r ru

st jo

bs th

an d

id o

ffic

e oC

cupa

tion

grad

uate

s.

4 Sa

mpl

e du

es a

re s

ame

es f

or w

age

rate

, fir

stjo

b.

Sou

rce:

Ger

ald

G. S

omer

s, e

t aL,

The

Eff

ectiv

enes

s of

Voc

atio

nal a

nd T

echn

ical

Pro

gram

s: A

Nat

iona

l Fol

low

-up

Surv

ey.

Mad

ison

, Wis

cons

in, 1

971,

Tab

le V

I.10,

Tab

le V

I.17,

App

endi

x T

able

20.

.

TA

BL

E 2

6R

EG

RE

SSIO

N.A

NA

LY

SIS

ON

WA

GE

RA

TE

FO

R F

IRST

AN

D C

UR

RE

NT

OR

LA

ST J

OB

AN

DA

VE

HA

GE

MO

NM

LY

EA

RN

ING

S, B

Y P

RO

GR

AM

AR

EA

, FO

R V

OC

AT

ION

AL

HIG

HW

HO

. OL

PO

ST-S

EC

ON

DA

RY

VO

CA

TIO

NA

L S

CH

OO

L A

ND

JU

NIO

R C

OL

LE

GE

GR

AD

UA

TE

S, S

EPA

RA

TE

"

RE

GR

ESS

iON

S

.

,Se

m P

ieps

-

-T

rade

and

2

WIN

*.

Dis

trib

utiv

eH

eal*

Api

cultu

re.

Tec

hnic

al

b(s

)b

(s)

b(s

)b

(s)

b.

Wap

i Rat

e,.

.--

Firs

t Job

Tot

al S

ampl

e-.

004

.05

.03

.07

.19"

.07

-.06

°.0

9.1

4**

.06

n=15

24-

Mal

en=

823

.15

.10

..2

6*.1

3.1

8.4

0.1

4.1

2.3

1**

.09

Fem

ale

n=70

1-.

15.

.10 -

0

-.11

.08

.14*

.07

nore

spon

ses

-.08

.10

Wag

e R

ate,

'.

'

Cur

rent

or

Lut

Job

Tot

al S

ampl

e'

.

-.08

.09

.11

.10

.24*

.10

-.06

.13

.17*

.08

n=15

24m

ale

' n=

1323

-.3

4**.

.13

35*

- .

:17

.37

.54 .

.17

.17

.39*

*,.1

2'

Fem

ale

-.21

.15

.01

.13

.18

.10

'no

-.17

''

.16

n=70

1.

resp

onse

s

Mer

ge M

onth

ly -

,0.

Ear

ning

s".

lota

l Sam

ple'

-n=

1337

Mal

e

n=73

5Fe

mal

e'.6

0!*

n=60

2

+

,+ no

resp

onse

s

-35

**

60*

.

Not

es:

Sign

ific

ant a

t the

.05

Wel

Of

igni

fica

nce,

two-

tille

d te

st,

Sipi

tican

t at t

he .0

1 le

ml O

f si

gnif

kanc

e, tw

o-ta

iled

test

.b

is th

e pa

rtia

l reg

rem

ion

coef

fici

ent

OD

is th

e st

anda

rd e

rror

of

tlii p

artia

l reg

rem

ion

coef

fici

ent

,1

TY

res

tudi

s m

odel

s jil

t con

trol

for

leve

l 'of

edu

catio

n, r

egio

n so

cioe

cono

mic

sta

tus,

rela

tcdn

ees

of jo

b to

trai

ning

, fat

her's

airi

catio

n; N

s, m

arita

l Sta

tus,

MC

C,.

adde

d ed

ucat

ion,

nir

al-u

rbaa

get

ting

and

grad

e po

int a

vera

ge.

.. 2

Eac

h of

the

prop

-ant

are

as m

a co

ntra

sted

witl

icfr

.m e

duca

tion.

Thu

s, f

oi th

e w

age

rate

; fir

st jo

b, to

tal s

ampl

e, tr

ade

and

indu

strP

..

-m

eupa

tions

sen

sed

.4 c

ents

less

per

bor

than

did

off

ice

occu

patio

n *i

lls.

.

3 +

WIM

P po

sitiv

e ba

t sta

tistic

ally

insi

gnif

ican

t par

tialr

eres

sion

coe

ffic

ient

;in

dica

tes

a ne

gativ

e bu

t ita

tistic

ally

'mag

nifi

cent

.

repe

l= c

oeff

icie

nt.

.

Sour

ce:

Ger

ald

G. S

omer

s, e

t eL

. The

Eff

eetir

eses

i of

Voc

atio

nal a

ndT

echn

kal.P

rogr

ants

: A N

atio

nal F

ollo

w-u

p 'S

urve

y, W

adis

on,

Wis

cons

in, T

abks

V1.

9, y

i.1s,

App

endi

x T

able

18,

197

1..

.,

.I

%

7.

p.

?.

b.

.

0 '

. .

.o

each of the five remaining\ program areas. In. Table 25, some consistent resultsbegin to show. Tilde and industry, distributive, and technical education allyield Iilher. 'first and last or current wage rIftsc than do the office programs.Howtever, except for the health program; junior ollege shows mi net benefitbetween office educatiort and eath of the five other programs. Clearly, this is astartling result, if trtie, indbears further Investigation to substantiate it. Ananalysis of the added interaction effects of sex would have helped shed morelight On these issues, tut wheri one breaks the saMple into subgroups ceil sizes incritical areas dirninish rapidly. This of coUrse, means:a very large initial size forthe aggregate sample, ind, since there are not many economies of scale in thistype .of survey,-;esPecially settle critical data editing stage, it implies very high.costs for this kind of complex analysis.

Table 26, however, doe.? shed 'some light Ortinteraction iffectsibetweenmules and females. Note, for instance; that for the`lotal sample the positive male .and negative female earnings benefits of trade and industry compared to officeoccupations cancel each other out. One would wrorigly bt: led to conclude thatthere is no difference in ea(nings between trade and industry and offiee'. .

occupations. Vet the Significant faCt is that males earn $69,ntore 'and fetnale's$60 less.in trade and industry occupations in contrast to office- occupation*?Throughout the ,table one can see how the male and fenialeeffects dilute orcancel each other in the analysis of the total sample. It should be clear tat flo.study which purports to measure the impact of education or training on labormarket success :should neglect the analysis of sex interaction effects: Theclassical way to do this is to separate the sample into its male- and feinFlecomponents, though anotherway is to add interaction terms to one's regressionequation. . ,

In summary, Table 26 'displays reasônable :results for tirles fer last orcurrent wage rate and for average monthly earnings. 'A cautious analysisimplymust rennin skeptical concerning the results in the cells aetheremainder of the

4 table.

I.

Manpower Daihing and Skill Areas. Tables 27 and. 2/3 show the effects of'27Manpower training on improgerished ruraLand fqm laborers in North Carolina,

'most of wHom are black. Table 27 again brings out the'importance of controlling-..for sex interastions.,(See tables above for further analysis toUchini on this issue.)Thus, for malls-4tIT coMpleters carn aborit.one dam:a week more than doinstitutional completers when both- are compared with their reipective.controlgroups. The table suggests that male 03T'4opouti,.d0-mtschliorse than maleinstitutional training dropouts so_that'Afein dropOuts and completers arecombined,all institutional-tfam'"Ies earn about a dollar More per 'week than allOJT traineesrYet; If one. were forced to Make a judgment, for this sample and

--studilcale, OJT. training seems to have a slight edge Oven' institutipnaltrainingfor males. This is clearly the case for females where both OIT coMpleteri anddropouts earned, relativelY large amdunu is contrasted to institutionalcompleters and dropouts who gained no earnings benefits vis-s-V1s,4heirrespective control groups. ,

98

a

0.

TAKE 27IMPACT OF MANPOWER AMNING ON WEEKLY EARNINGS

AND HOURS WORKED PER WEEK, MALES AND FEMALES

,

.

.Weekly earnings

(dollars)Hours worked

per week

Male'. ,,,

..;,..., OJT completersOJT dropouts . -

Institutional completersInstitutional dropoutsAll OJT traineesAll institutional trainees

Females 1

OJT immpletersOJT dropouts '-..institutional completersInstitutional dropoutsAll OJT traineesAll instutional trainees

.

,

.

u,... ,

9.5**0.7 -.8.5*-1-8.17.4**8.3**

-17.0**11.4**0.2-3.214.5**°-0.3'

0

0.61.91.92 9. ...0.92.1

10.111f*

4.0-1.2-1.87.7**

Notes: Significant at the .01 level.1 The coefficients are to be interpreted as deviations from the average

experience of the control group members. Thus, male OJTcompletsrs earned $9.5 a week more than male centrol groupmembers and worked .6 hours more per wftk.

Source: David 0. Sewell, Training the Poor, Kingston, Ontario, 1.971,Table 7, p.71.

= Table 28 displays weekly earnilgs benefits by occupation for the same.sample. of rural poor. (See also Table 20.) 'Only seivice workers do notexperience a net gain in weekly earnings compared to the control group. In viewof the heavy concentration of blacks in this study sample, plus changingaspirations and racial attitudes; perhaps it is not too efficacious to train blacksfor service jobs. For a variety of reasons, they simply may not sucCeed at thisoccupational activity.

Finally, in his study E.. well (1969,1111) discusies The issue cOncerning wagerate improvement versus employment improvement as a measure .of the impactof training, on productivity. While, as. mentioned above, the general argumentthat the effects of education and training should not be confounded byemployment effects is a correct one; it is an argisnent tempered for an economyWhere no structural unemployment or other rigidities or' structural bOttlenecksconfuse the picture. Given the obvious structural unemployment_and ---underemployment of his sample, earnings is a better measire of program effectthan wage rates. Wage rates can be assumed inflexible downward ecially inthe short run, due .to suph things as the social minimUm wage. Thus, they wr

. \

TABLE TS

MULTIPLE REGRESSIONS OF WEEKLY EARNINGS AFTER TRAINING,A ND ME COMPONENTS OF WEEKLY EARNINGS AFTER TRAINING,

ON RESPONDENT CHARACTERISTICS: TRAINING STATUS CLASSIFIEDBY TRAINING OCCUPATION

Weeklyawnings(dollen)

Noon '

workedper vveek

Newlyminim.(cents)

lutitutional Trainees! .

Net regression coefficients:Building tradesmen 2.91* -4.0 36.51*Nurse aide - . -5.4* -8.7"Oicupational farm tractor" 0.3 3.9 5.3

, Miscellaneous . 14.4** 5.1* 26.31*

All Trainees! N,

Net regression coefficients: N,Clerical and sales workers 26.21. 8.2** \ 34.3**Building tradesmen 10.5** 0.5

. \Mechanics .6.7** 4.6** 40.0Plant operatives 6A** r.6 157**Servick workers -0.1 -1.5 -7.0 \

Notes: *Significant at the .05 level.1* Significant at the .01 level.1 The coefficients are to be interpreted as deviations from the average

experience of control group members; thus building tradesmenunder the institutional training componenj earned $12.90 MOM perweek than their control group counterparts. .

Source: David 0. Sewell, Trek* the !tor, Kingston, Ontario, 1971, Tigble 9,Pale 79.

. not adjust themselves downward to allow persons of low productivity ,to be -

employed in the market place. Hence, the strategy is to. raise worker'sproductivity up to that level required to justify paying him current money. wlevels. When this occurs, the program will show ,no wage effect ,bu anemployment and , earnings effect. Thus, Main was' inconect to concliide thatretraining hid no impact on the productivity of MDTA institutional trainees inhis national sample simply because trainee wage rates did not rise. In short, for .

MDTA trainees and other 'structurally unemployed workers, employment andeainings and 'hot wage rates are the preferred indices ,of labor marketperformance. *

Taken in this light, the residts reported bY Sewell are of some interest,especially since he shows clerical and ealea workers redeiiing greater earningsbenefits than building tradesmen, mechanics Of plant operatives, occupationalareas where, in Contrast to clerical and sales, there has been a longer hctoricalpenetration by blacks.

100 .0.1S,11113

IN

SUMMARY, IMPLICATIONS, RECOMMENDATIONS

Ilummsry and Implications

Methodology. .The methodology exista to perform benefit-cost stndies inan appropriate manner. Much of the current ambiguity that exists concerningthe implications of benefit-cost studies is the result of Using faulty methodology.The use of control groups is- an absolute necessity, for-instance, yet manystudies, especially the earlier ones, use.no control or comparison group.

Adequate sampling procedures and adjUstment for non-responn bias andself-selection bias are necessitiei. Random probability samples of the poptilationof interest 'seem Obviout, yet, judgment samples continue to be.-used. It costs

5 more to perform a random sample since the population frame muat be .

established before sampling is performed, but the extra cost is preferable to thepossibility that the entire study may have to be junked because no one it sure ofthe meaning or applicability of the results.

Control for non-response bias is costly, too: It often costs several hundreddollars to run down* a single observation. But,-in Many caies even the mostelementary information is lacking with respect to non-response groups. This can,usually be traced to the fact .that it costs more per unit o( information and isdistracting to deal with a group of persons on whom there, is often no concreteinformation available.

Self-selection bias, the bane of retrospective evaluations, has mit beendealt with. adequately in the past. The -use Of. probit analysis, a standardeconometric technique, to estimate a discriminant function can help a long way,especially if motivational or other psychological information is available.

Both tabular analysis tand regression analysis should be used. Tabulitionsare valuable in laying out . the main dimensions of the study, .but it is .orilyregression analysis which has the flexibility to handle large numbers of varksbkssimultaneously. In retrospective studies of population cmss-sections, even a basiclist of . variables to properly investigate labor market influences can, run to adozen; sex, age, race, education, marital status, socioeconinnic status, IQ,occupation, industry, labor market struCture and geographic region are likelycandidates. Tabular presentation , simply cannot simultaneously handle thisnumber of variables and their interactione efficiently. -

Finally, more attention .should be paid beforehand to the acceptable levelsof, statistical significance (tempered by cost constraints) which are desired. Themethodology for establishing desired levels of statistiftl significance in one'ssample is laid out in most elementary statistics texts Li the use of ,it is notapparent in the stUdies represented in this survey. An example of the use of sucha .methodology linked with cost constraints is with the Gary Income-Maintenance Experiment which, is being conducted by Indiana University at

1.09. 101

Gary. By establishing control over significance levels beforehand, it woUld be'possible tO avoid the calculation of complex results, all .or most of which are oflittle use since one has variances too high to provide definitive results.

From the standpoint of benefit-cost analysis, both average and marginalbenefit-tost ratios should be 'used. Thus far, most studies do not allowcomparisons on a marginal basis among program alternatives. .

Most of the studies do not compute true marginal benefits. Regression'coefficients based on . dumMy variablei yield differences between .. averages.Marginal benefits can be calculated by expressing prograM benefits as a functionof length of expOsure tolhe program. The study by Hardin and Borus (1969,

. III) does this rather well: If the vocational or technical students and their;controls are handled in the analiais by using dummy variables (or by Means ofcross:classification in tables) the result one gets is 'a difference. between the .

avers* experience of the vOcational student and that of the control group. A .

marginal, benefit cannot be estimated 'in this way unless the assumption is madethat tht difference between the averales is equal to the marginal difference, Not

.

-everYorie will' accept such an assumption.The diseussion over the true value of' the social opPortunity cost of capital

, remains-unresolved. For groups other than. society, the borrowing rite of funds,if thesducational investment is financed by borrowing, and the lending rate, ifthe irfvestment is financed from savings, are the appropriate rates fen individuals, .,firms, and governmentunits other than the federal governments

The proper Investment criterion to be used inmaximizing the'present valuesof neWbenefits depends -an the constraints which exist in the, progranesinstitutional structure. All of the three basic criteria auffer from shortcomingswhen humii capital innstments are considered.

Benefit-Cost: Secondary _Vocational Education. What can be said? Thefederal .government is still not 'confident since it has recently: funded a newnationwide Study of vocational education to be performed by the NationalPlanning Association. Yet the federal governMent appeari too. cautious: Exceptfor the. Taussig study (1968, II), all of the major studies here shOw a positiveeffect of secondary vocational education. A.Nrnajor question mark lies with thedata fromProject TALENT. These data have been yet to be effectively analyzed,by investigators not directly associated with the federal government. Since thisdata bank is a national sample, it is important 'that it be appropriately analyzed.in the very near future. The major problem with these data lies in a relativelyhigh non-response rate;

1n-short, based oil current evidence, it seems clear that die secondary voca-tional curriculum yields greater labor market benefits relative to the compiehen-sive curriculum. This judgment is qualified by the fact that the objective functionsof the two groups are not the same and the population served are not identical.

Little or no reliable knowledge of the relative investment gains of vocationalcourses of program areas exists. The estimates of benefitS of prOgram.areas ire-still inconclusive and sound cost measures are yet to be done. Benefit-costmeasures whichcomtiare sets of vocational skint on which one would be willing

/to. stake millioni of dollars simply don't exist. The benefit Measures areaMbiguous at best and are of no use in any case in the ,absence of measures of

,

102

_

realtive costs. It does no good,to provide tabulations of gross earnings or wagedifferentials by skill if information on costs is lacking. However, it is worthmentioning m this point that the analysis by Schweitzer (1970, II) suggests thatthe length of fOrITIThe schooling processsuggested by the Hansen-Wei

hooling for many occupational areas .may be too long.be longer than the learning process as 'furtherrod-Scanlon study. .

Benefit-Cost: Post-Secondary Vocational Education and Junior College.The studies are pointing to the likelihood that post-secondary vocational°education it a losing proposition for students who already gave a sound highschool vocational preparation. The junior college, however, does appear to yieldsubstantial benefits overt secondary vocational edutation, though again, theproblem ,exists that the objective functions of the .two different populationssimply may not be the saMe.

Morsch (1970, 11) reports that.the.sost for the same occupational trainingis about the sarne for post-se'condary vocational and junior college education. .Thus, if junior college pays on the, _average while post-secondary vocationaleducation does nOt, the, average mix Orikills at the two types of institutions maybe the critical point to investigate. Of course, as With secondary vocationaleducation, this analysis remains to be done effectively.

... Benefit-Cost: Manpower 7Yaining. That the present MDTA manpowertraining program is a worthwhile social investment seems indisputable. It iF stillnot clear, however, whether it is.the actual skills learned, the various servicesthat accompany retraining such as placeinent, or the sheepskin effect as aselective device which is mainly responsible for the high rates of return: Aneffort should be made to identify the net effects of these three possible causalfactors. Yet, one hesitates to recommend the additional investment of socialcapital in .this effort at this time unless it is accompanied bY a careful -study, ofthe absolute and relative costs and benefits of, differ:at occupational skillswithin the program plus an assessment of the reasons for the continuing"shortage" of persons in occupatiOns.such as nurse's aide even in the face ofcontinual retraining.

he studies of the remaining manpower programs'are a mixed iot. Rothdue to subtle and not so subtle differences in populations bens served as Well aswide variations in methodologies one cinnot make reasonable marginalinvestment adjustments among these programs based on these data. One canonly judge that thisiliogrami-afi_coyering their average costs. There is a basicproblem in attempting to use retiolpectivestudies 'for 'economic\decision-making. The usual. approach has been for the governrnennoexpressinterest in a target population, whereupon a program was designed to aid it. Thisprogram' Was then retrospectively evaluated. Two alternatives to this sequenceare available. First, planned experiments, as with the three ongoing incomemaintenance studies in Seattle -- Denver, Gary, Indiana, and New.Jersey-Pennsylvania should be attempted. Thus, consider whet might be thernproper training program for welfare mothers: The Work Incentive program(WIN), the MDTA approach, adult ofremedial education, and no program at, allimmediately piAsent themselves as possible altermtivet-Experiments with ainoper control group design could be devised to see which of these methods is

3

most effective in, alleviating the stress of poverty among Welfare mothers, forinstance.

Alternatively, a particular target group Could be selected among thevarious ongoing programs to estimate the differential program effects on thispoulation. In terms of research findings these two alternatives may be more

.--ixpensive than the unit study cost of the present type. Hdwever, they might bemuch lessrexpensive in terms of the amount of usable poiicy guidance theymight provide.

Re Search in Progress.

Vocational Ediscation. At the present time several major evaluations ofvocational education are ongoing. The Project METRO study by Max Eninger(197: II) is a nationwide analysis Of secondary vocational education funded bythe U.S. Office of Education. Theie data _are based on high school records and a

, mail labor market questionnaire Of the 1968 graduating class. Preliminarytabular analyses of the data ire in progress. Regression analysis of benefits will

. be conduCted but there are no estimates of costs hased on this sample.The National Planning Association has just been awarded a major contract

by the US. Office of Education toatudy the labor market andother nonecononticeffects of vOcitional education at the SecondarY level. An.effoit will be.Made toestimate the net: effects of vocational _education by program area and bytreatment such as iooperative educatini. 'Inis study is planned to run into 1973:Two cohorts of students-will be analyzedthose 10th graders who could havegraduated in 1968 and those who could have graduated in 1970. ApproximatelY20,000 Obseriations are planned for analysii. A mail labor market questionnaireis planned to collect the needed labor,maket.date. Da'ta frOM school records willalso be collected. Costs will not be collected except for a small Subsample ofschools.

Manpower Studies. Operations Research; Inc., in conkinction with theOffice of Ccondmic Opportunity .(0E0) is conducting a major benefit7cost studyof four manpower proerams which promises to be defmitive: The programs areMDTA institutionsf training; the Job Corps; NAB-JOBS; and the outof-schoolNYC. Data will be collected by Operations Research, Inc., and analysis will, bedone liy the Office of Econcimic Opportunity: .

This study 'shows great promise since the response rate his beenmaintained at over 95 percent and an'. elaborate .iampling pmcedure wasdeveloped to select a sample front a aet of large Metropolitan areas Whichthemselves were judgmentally selected.

An elaborate: person.Ally administered questionnaire is the main data basebut large masses of individual records frcim- training centers, and the like, arebeing collected on apprOximately 10,000 sample observations. . . .

The cost analysis is beiag conducted by the present author in Conjunctionwith Operations Research, Inc. Benefit analysis will be conducted 1,4, the OE0.

This study promises to be ahe definitive work on manpower programs todate, especially since it will allow marginal eConomic' comparisons awing majormanpower progfams on the basis of a commOn methodology throughout.

. 104iStz

Finally, the present author is conducting an- economic and institutionalanalysis of the Work Experience and Career Exploration Program (WECEP).WECEP is designed to provide occupational 'training and career explorationopportunities to disadvantaged, handicapped or otherwise alienated 14 and ISyear old high school youth in an effort to reduce the dropout rate.

This benefit study is nationwide in scope and will relyfor its data mainly.Oh records which are collected at the WECEP sites on the experimental groupand a preselected sample of controls. .

Preliminary analysis of the 1971 class is scheduled for the Spring of 1972while a final report incorporating the experience of the 1972 class as well is &liein the Fall of 1972.

Other Data Sources. For the energetic person,..both the data from theCoeman Report and Project TALENT remain to be properly exploited,

Finally, and perhaps most iMportantlY,.the.1970 Decennial Census has asection on vocational education based on a five percent iamplekThese data arecurrently coded and.aviait proper analysis.

. N.

The Future qf Benefit-Cost Analysis in Vocational Edniation

-The future of be ne fit-c o st analysis depends in part on what the presentstate of the art offers. So; the questiOn is, what does the typical benefiticoststudy now offer? As suggested 'aboie, the study will typically show only that aprogram appears to pay' or does not. (See Barth, .1971; Ill, on which thediscussion of the next few paragraphs is largely based.) The next judgment iswhether to expand or contract the program, depending on what the calculatednumbed indicate. However, the calculated rate of return or net present value .

does not tell you how much to expand or contract .the program. First, thebenefits or costs or both will' often be assumed constant over the relevant rangeof output and for all future periodi. &it, in the real. world, this is an unlikelySituation: With a 'failure to use nonlinear total cost or benefit functions and. ageneral lack of estimating such things as average cost functions, one hu no ideawhether one is operating in a range of increasing or decreasing returns or.costs orwhat the optimal scale of operation may be. One simply can't tell how fai toexpand or contract, and. knowledge of the future stream of costs and benefitsdoes not rAist.

Second, if one expands or, contracts a program on* the bash of thebenefit-cost analysis', it is not clear what one gives up in order to expand the

. program or what will replace the program in the event that its kvel of fundingiscut. Present benefit-cost studies do not address themselves to these hues, but itis clear that the broader' the expanse of alternatives one considers as iubstituteprograms, the less relevant any given efficiency .comparison becomes, since _theprogram objectives and outputs begin to differ .widely. Benefit-cost !studies as yet

. have not even extensively evalusted different production techniques for thesame type of output and program. H

Finally, most' Of the studies have not even cfAculated marginal coita and:benefits so they exhibit a fundamental inability -so aid in making choices amongalternatives.

, "With such fundamental problema one may rightly ask, why bothee withbenefit-cost analysis at. all? Yet, in spite of the above defects, benefit-costanalysis should and will continue to be done 'since it interjects a rational,systematic analysis into an area where judgment and implessionistic analyaisformerly operated with little contamination from orderly economic analysis or ac le a r-cu t spec ification of program objectives, outputs and evaluationmethodology.

However, we are clearly not yet at the, stage where benefit-cost mulishcan be widely adopted. and automatically integrated into the very fabric ofdecision-making at the federal or state educational level much less at the schooldistrict lrfel. For_ the present, too mars- methodological issues remain to berefilled, leaving the estimated measures of benefit too crude. Also, at present:itwould simply be too expenhive to continuously generate data good enough tobe -used in the resource allocation process on all specific local programs, even ifeconomic considerations had a weight of unity in the decision-making process(which they do not). For the time being, therefore, we may expect to see only a

---------ciantinuation of ex post analyses of specific programs or activities for veryspecial purpose-Cif-0i o r judgmentAwilbe some time even before extensive useof the type of methodological approach InIfirOperatio Research, Inc. - Officeof Economic Opportunity's study of four manpower prom --iswidelyadopted. Ihe cost of such types of comparative imaluations, Usually severalmillions,seems an effective deterrent to their widespread adoption at present.

:.

BIBLIOGRAPHY'

(Boldface type indicates pages which cite the references)

I. Conceptual buses

Abt, Clark C., and Miller, Peter S. Design for an Elementary and SecondaryEducation Cost-E ffectivineis ModeL Volume I, Model Description. Cam-bridge, Massachusetts: ABT Associates, Inc., June, 1967. 165 pp.

Design for an Elementary and SecondsCost 1ff ecriveness Model, Volume II, the Usee's Guide. Cambridge, Massa-chusetts: ABT Associates, Inc., June, 1967. Volumes I and II, 231 pp.(ED 014 152 'MF $0.65 HC $9.87)

An Overview of the Cost-Effectiveness Modeland Sub-models for the Evaluation of Mle I, ESEA Proposals. (Availablethrough ERIC as Technical Note, No. 14) .

Alberts, David S. A Plan for Megiuring the Performance of Social Programs; TheApplkations of Operations Research Methodology. New York: Praeger,1970.

Alkin, M.C. Towards An Evaluation ModelA. Systems Approach. Los Angeles,California: University of California, Center for the Study of Evaluation ofInstructional Programs, 1967. 27 pp. (ED 014 150 MF $0.65 HC $3.29).

Andcl, Norbert. "Some Notes on Equating Private and Social Cost: Comment,"Southern Economic JournaL 33:112. No. 1, July, 1966.

Ar.rnw,' Kenneth J. "Criteria for Social Inwstment," bitter Resources Research.1:1-8, No. J, 1965.

"Discounting. and Public Investment Criteria," in Allen V.Kneese and Stephen C. Smith, (eds.), Water Research. pp. 13-32. Balti-more: Johns Hopkins Press, 1966.

I Bibliographipai enhies followed by m ED or MP number in parenthesis are generallyAvailable in hard copy dr microfiche thtough the Educational Resources Information Center(ERIC). This availability is hidicated by the abbreviations; MF for microfiche and HC forhard copy: Disler from ERIC Document Reproduction Service (ERDS), P.O. Drawer 0,Bethesda, Maryland 10014. Payment must accompany orders totaling less than $10.00.Doctoral dissertations with a microfihn number are available hi microfilm (84.00) orxerographic cony ($10.00) from University Microfilms, Disinflation Copies Post Office Box1764, Ann Arbor, Michigan 48106. Biographical entries followed by as AD ot PB numiser inparenthesis are generally available in microfiche or paper copy through the National .

Technical Information Service (NTIS), Springfield, Vkgkds 22151. Microfiche price k $0.95 .per tide and paper copy price variei . with the length of the report. Ail orders must beaccompanied by a check or money ordei.

107'

"The Social Discount Rate," in G.G. Sailers and W.D. Wood,(eds.), Cost-Benefit Analysis of Manpower &licks, Proceedings of a NorthAmerican Conference, May 14-15, 1969. Kinpton; Ontario: Industrial Re.lations Center, Queen's University. pp. 56-75, 1069. Complete text 188pp. (ED 039 334 Document not available from EDRS)

.Auerbach Corporation. Evaluation Guide. Final Report, Appendix B. Philadel-

phia, Pennsylvania: Auerbach Corporation, March 8,'1968.

Evaluation Manual and Guide. Final Report. Philadel-phia, Pennsylvania: Auerbach Corporation, March 8, 1968.

Bailey, Martin J. "Formal Criteria for Investment Decision."journa/ of Political- Economy, pp. 476-488, October,1959. 12

-Balogh, T., eat Streeten, P.P. "The Coefficient of Ignorance," Bulletin of the

Oxford Universi6 Institute of Economics and Statistics, 25:99-107,May,1963.

Bamsby, Steve L. "An Analyiis of Data Requirements for Cost:Effectiveness andCost Benefit Analysis." Report to the Arizona Research CoordinatingUnit, ArizOna. January, 1970.

Baumol, William I. Economic Theory and Operitions Amlysis, \2nd ed.Englewood Cliffs, New Jersey: Prentice-Hall, 1961. 12

"On the Appropriate Discount Rate for Evaluation of Public projects,"'Hearings Before the Subcommittee on Economy in Govennselif of theJoint Economic Committee, US. Congress, 90th Cowen, lit Session,1967. pp. 152-159. Washington, D.C.: Government Printing Office, 1967.

"On the Discount Rate for Public Projects?' The Analysis and EvalUationcf Public Expenditwes: The PPB System, in U.S. Congress, JointEconomic Committee, Washington, D.C:: Government Printing Office, .

1969, Vol. 1, pp. 489-504, 9Ist Congress, 1st Session, Joint CommitteePrint.

Becker, Gary S. &men Capital: A Theoretical and Empirical Analysis, withSpecial Reference to Education. New York: Columbia University Press,1964. 18

"Investment in Human Capital: A Theoretical Analysis," Journal of-PolitkalEconomy, 70: 949, part 2. (Supplement), October, 1962.

Becker, Joseph M. In Aid of the Unemployed Baltimore: John Hopkins Press,1965.

Betry, Paul C:, et el. The Future of American Poverty: &KM laic Issues inEvaluating Alternative Anti-poverty Measure". Croton-on-Hudson, NewYork: Hudson Inst!tute, 1968. 390 pp. (ED 036 713, Document notavailable fiom EDRS). _

Besen, Stanley M.; Fechter, Alan E.; and Fisher, Anthony C.. "Cost-EffectivenessAnalysis for the War on Poverty,' " in Thomas A. Goldman, (ed.),

\

108 -116

-

.

,Cost-Effectiveness Analysis New Approaches in Decision-Making, NewYork: Praeger, 1967. 15 pp. (ED 024 877, Document not available fromEDRS). .

..' Blaug, Mark. "The Rate of Return on-Investment in Education in Great Britain,"The Manchester School, 33:205-61 No:3; September, 1965. 13

Bolino, August C. Manpower and The City. . Cambridge, Massachusetts:Scbenkman Publishing Company, 1969. 300 pp: (ED 037 635, Documentnot available from EDRS).

Borus, Michael E., and Task William R. Measuring the Impact of Manpower'Programs. A. Pittner. Policy Papers in Human Resources and IndustrialRelations, 17. Ann Arbor: The University of Michigan; Detroit: Wayne -State University, Institute of Labor and Industrial Relations, November,1970. 38 . ,

BOwen, William G. Economie Aspects of Education. Princeton,. New Jersey:Princeton University, 1964.

Bollby, Roger L., and Schriver, William Non-Wage Benefits of VocationalTraining: Employability and Mobility. Knoxville, Tennessee: University ofTennessee: 1969. 22 pp:.(ED 042 am IPAF $0.65 NC $3.29). , .

BoWnian, Mary J. "The Assessment of Human Investments as Growth Strategy,"in U.S. Congress, Joint Economic Committee. Federal Programs for theDevelopment of Human Risources: A compendium of papers submitted tothe Sub-Committee on Economic Progress: Washington, 'D.C4.; P.S.Government Printing Office, 1968. Vol. 1, pp. 84-99. 90th Cnniress, 2ndsession. Joint Committee Print.

C.

"The Costing Of Human Resource Development.," in E.A.. obinson '1

' and J.E. Vaizey, (eds.), The Eamomies of Education; New,. York: St.--Martin's Press, 1966. 0..421450. 29

_e...."Human Capital: Concepts and Measuresrin Selman J. Mushkin; (ed.);.EconOmics of Higher Education:Iii:69-92. Washington, D.C.: Departinent.

' of Health,Educatiand Welfare, Office of Education, 1962.. "Social Returns' io Education,". International Social Sceinie Journal.14:647-659, No. 4, 1962.

Buchanan; J.M.,.and Stubblebine, W.C. "Externality," Economics, 29:371-384:- No. 116: November, 1962. ..

Cain; Glen C., and Hollister, Robinson G. "Evaluating Manpower Programs .for. the Disadvantaged?' A piper presented at the North American Conference

.on Cost-Benefit Analysis of Manpower. Policies. May 14-15, 1969.Madison; Wisconsin: 'University of Wisconsin.

. "Methods for Evalwting Manpower Program's," in A. Weber,' et aL, (ed.),Public-Private Manpower Policies. Madison; Wisconsin: Industrial RelationsResearch Association, 1969. 210 pp, (ED 043. 807, Document notavailable from EDRS). e

'Capron, William M. "The Potential Role of Cosi fffectiviness AnalYsis for. Evaluation of Government DoMestic Programs.' An address.before the

117

Symposium on Cost-Effectiveness Analysis, Institute for Defense Analysid,Arlington, Virginia. June,15, 1965.

.

Carroll, Stephen J., and Pascal, Anthony H. Youth and-Work: Toward A Modelo f Lifetime Economic .Prospects. Santa Monica, California: . RandCorporation, 1969.

Center :for Human Resource Research. A Conference on-the Evaluation of theImpact of Manpower Programs, June 15-17, 1971. Columbusi, Ohio: TheOhio State University, 1971.

Chamberlain, Neil W. "Government Investment: How &ientific Can It Be?"Challenge. 14:32-35. July-August, 1966.

"Some Further Thought on the Concept of Human Capital," in G.G.Somers and W.D. Wood, (eds.), Vost-Beneftt Analyses. of ManpowerPolicies. Proceedings gi a'North American Conference. Ontario, Canada:Queen's University at Kingston,'Industrial Relations Centre, 1969.-188 pp.(ED 039 334, Document not available from EDRS).

"Some 8econd. Thoughts on, the Concept of Human Capital," TheDevelopment and Use of Manpower, Proceedings of the Twentieth AnnualWinter Meeting. Washington, D.C.,.Dec.-28729, 1967, pp. 1-13, Madison,Wisconsin: Industrial_Relatiorti Research Association, 1967. 408 pp. (ED041 100, Document- not available from EDRS).

ChaseSamuel B. (ed.) Problems in Public Expenditure Analysis. Washington,D.C.: The Brookinp institution, 1968. 31 . .

Ciriacy-Wan trup, S.V. ."Bene fit-Cost Analysis and Public Resources. Development," Journal of Farm Economics, 37: 676k89, Noveniber,.1955.

.

Coase, R.H. "The Problem of Secial Cost," Journal of Law and Economics.3:1-14.4Actober, 1960. .

Colm, Gerhard. "On the Problems of Program Appraisal in the Field ofDevelopment of Human 'Resources," U.S Congress Joint EconomicCommittee, Federal PIOgraMI for the Development of Human Resources:a compendium of Open submitted to ;he Subcommittee on EconomicProgress. Washington, D.C.: Government Printing Office, 1968, Vol. 1, p. .

23-37. 90th Congress, 2nd Session, Joint Committee Print. ,

Correa, H. "Optimum Choice Between General and Vocational Education,"Kyklos,18: 107-146, No. 1,1965.

Coster, L.R., and Ehnen, L.A. "Program Evaluation," Review of EducationalResearch, 38(4), 417-433. October, 1968.

Cotner, Melvin L 'Criteria and Problems in the Use of Quantitative Analytical .

Piccedures for .Public Devslopinent Investments," Journal of FarmEconomics. 44:1,085-1092, No. 4, November, 1962.

Cox, John A. "Apilication-of a Method of Evaluating Traiiing," Journal ofApplied PsYchology, 48: 84-87tApril, 1964.

110

1

Davie, Briice F. "Using Benefit-Cost in skanning and Evaluating -VescationalEducation," A paper prepared for David S. Buihnell. Washington, D.c.:Bureau of Research;4.U.S. Office of Education. November, 1965. 20 pp.(ED 016 077 MF $6.65 HC $3.29).

Davis, Otto A., 'and Whinston, Andrew B "Some Notes on Equating Private andSocial Colts" Smithery, Economic JownaL-32:113-125, October, .1965.

Davisson, William I. "Public Investment Criteria," Land kconomics. 40:1 53.16 2,1964.

.

Dearden, .John.- e6-st and Budget, Analysis. Englewood Cliffs, New Jersey:Prentice-Hall, Inch 1962.

Der Rossi, .I.A. Cost Analysis for Program Bw!getini. Santa Monica, California:The Rand Corporation, 1968.

Devine, El. "The Treaiinent of Intommensurakles in CoskBenefit Analysis,"Land Economics, 42:383-387. No. 3. August, 1945:1Reprinted by theInstitute of Government and Public Affaiis, University of California, LosAngeles, 1966.)

Dobell, A.R., and Ho, Y.C. "An Optimal Unemployment Rate," Quarterly. Journal ofEconomier, 81:675-683, November, 1967.

Dorfman, Robert (ed.). Meowing Benefits of Government Investments.Washington, D.C.: The Brcinkings Institution, 1965. .

Durbin, E.P. Manpower Programs as Markov C'hains. Manta Mqnica,Rand Corporation, Rand Memorandum, October, 1968- .

Dymod, . William R. "The Role of Benefit/Cost Analysis in FormulatingManpower Polici." in G.G. Somers and W.D. Wood, (eds.), Cost-BenefitAnalysis of Manpower Policies, Proceedings of a North .Americanconference. May 14-15, 1969. pp. 42-55. Complete Text 188 pp. (ED 039334, Document not available froin EDRS).

Eckaus, Richard S. "Economic Criteria for Education and Training," Revinv ofEconomics and Statistio 461 81-90, May, 1964.

"Investment in Human Capital: A Comment." The Journal of Political. Economy, 71:501-504. October, 1963.

Eckstein, Otto A. "A Survey of the Theory. of .Public E4enditure Criteria," inJames M. Buchanan (ed.); Public Rnances: Needs, Sources and Ulilization.pp. 439-94. National Bureau. of Economic Research Special ConferenceSeries No. 12, Princeton, NJ.: Princeton University Press, 1961. 12.

Enthoven, Alain "Appeallx.:.The Simple Mathematics of Maximization," inCharles J. Hitch and Roland.N. McKean, The Economics of Defense in theNucker Age, Neiv York: Atheneum, 1965, pp. 380-385: 35 .

"Federal Procurement Regulations (2iid ed., FPR Amendment 42, April, 1968),Part 1-15, Contract Cost Principles and procedures and Subpart 1-14. 2Principles .and Procedures for- Use.in Cost-Reimbursement .Type. Supplyand Research Contracts with Commercial Organizations;pp..1501-1520.31 .

. .

.

S.

e.;

. o

.Feldstein, Manin S. "The Secial Time Preference Discount Rate in Cost Benefit

-Analysis," Economic Journal. 74:360-379.,No, 294, JUne;. .

"Net Social Benefit ililculatfons 'and publie. Investment DeCisioni"Oxford Ecynomic Papers, 16:114-131, No. 1, Mareh, 1964:

and Flemming, IS. "The Problems f Tinie-Stream EvaluatiOn: Present.Value Versus Internal Rate of Return Rules," BUlktin of the Oxford'Ilnivenity Institute of Economics and ...Statistics,' 26:19-85. No: 1,February, 1964., .

. .

Freeman, Howard E. "Conceptual Approaches to Assessing Impacts.

.of Large-icaleintervention Prograni," American Statistical Anociation Social Statis-tics Secelon, Proceedings, 1964. pp. 192198. '". .

. Freeman, Myrick A. III. "Project Design .and Evaluation 'With MultipleObjectives." The Analysis and Evaluation of Publ c Expenditures:. The PE BSys:en WashingtOn, D.C.: Government Printing Office, 1969; Vol. 1. pp.565-578, in U.S. Congress, Joint Economic Committee, 91st Congress, 1st .

Session, Joint Committee Print. .

Glernan. Thomas IL, Jr. Evaluiting Federal Manpoiver Program: Notes 'and. Observations; Santa Monica, California: Rand CorpOration Memorandum

RM 5743-0E0, September, 1969. 55 pp: (ED 041 111 MF .$0.65 -NC$3.29). 6

Goldman, Thomas A. (ed.). Cost-Effectiveness Analysis:: New Approaches-in'''." Decision-Making. New York, New York:Tederick 'A. Praeger. 1967: . .

Gorham, William. "Allocating Federal. Resources Among- CoMpeting SocialNeeds,"...Healtli, Education, and .Welfarelndicatort .AUgull 1966, p..1-13.

Grosse, Robert. 'The Program Budget: Its Value to Education aiTellsra:, and Local . Levels,' in , Aronld. Kotz, (ed.), Occupational Educa

Planning end Programming. Meulo Park, California: Stanford RelearInstitute; September, 1967.. pp. 229-248. Complete Text 279 pp. (ED 017734 MF $0.65 HC $9.87). -

GubinS, Saffluel. The Impact of Age and Education ontlie Effectiveness ofTraining:' A Benefit Cost Analysis. Unpublished 'doctoral dissertation, JohnHopkins University, 1970.

Ilamen, W.Lee;and Weisbrod, Burton A. Benefits, Cost and FInance of Nigher.Education, Chkogo:. Markham PUblishing Co., 1969. C.

. Hsirberger, Arnold C. "Discussion:" Professor .Afrow on. the. SOCIAL DiscountRate," in G.G. SOmers. and W.D. Wood; (eds.), Con-Benefit.Analysis ofMmpower Polkies. Proceedings of North. American Conference.Kinpton, Ontariii: Industrial . Relations Center, Queen'i University at

, Kingston, 1969. .10 pp. (ED' 039 334, DoCument not available frornEDRS). , .

. Hxrdin, Elnar. "Benefit-COst 'Analysis of Occupatkinal Trainin_,- Comparison of ,Recent Studies," in G.G.Somers, and MD.." Benefit Arwlyni of Manpower Polcies, Proceedings of aWorth American

Conference, (May 14-15, 1969), Kingston, Ontario; Queen's. University,1.' e

.

112

.

Industrial Relations Center, 1969. Complete Teija 188 o. (ED 039 334,--;'Document not available from EDRS). , ..

-..___. "A Summary Guide for, Benefit Cost Analysis," in Arnold Kotz, (ed.),Occupational Education: Planning nod Programming. Volume 2, MenloPark, California: Stanford Research institute. .Septemb'er, i967. PP.379-386. Complete Text 279 pp..(ED 017 734 MF $0.65 IftS9.87):

, and Borus, Michael. "An EconOmie Evaluation of thc Retraining:, .

PrOgiams in Michigan: Methodological Problems of Research," AmericanStatistical Ass9dation. Proceedings of the Social Statistics. Section, pp;133-137, 1966.

,Haveman, Rolwt H. "Benefit-Cost Analysis: Its Relevance to Public Investment. . .

Decisions: Comment," Quenerly JoWnle of Econornics, 81:695-702. No.4, November, 1967:

, and tilla, John. "Unemployment, Excess Capacity, and Benefft-COstlnvestMe t Criteria," Review of Economics and Statistics. 49:383-392.,No. 3, Au at, 1967. 2, ..

Hines, Lawrence "The Hazarlds of Benefit-Cost Analysis as a Guide to Publicinvestment P cy," Public Finance, 17:101117, 1962.

Hinrichs, Harley E., d Taylor, Graeme M. (eds.). Program Budgeting. and .

Benefit Cost Analysis. Pacific Palisades, California: Goodyear Publishing'Co., 1969. : . . .

o /Hirsch, Werner; et al.' Spiliover of Public Education Costs and Benefits: Part 2, .

Los Angeles, Universay of California: Institute of Government and PublicAffairs, 1964. 484 pp. (ED 003 020 MF $0.65 HC $16.45).

Hirshleifer, Jack. "On the Theory of Optimal Investment Decision."_Jouptal ofPolitical Economy, pp. 329352, August, 1958. -12

et-al. Water Supply:. Economics, Technoloiy, and Policy. Chicago: TheUniversity of Chicago Press, 1960. 12, 19, 20, 34.

Hitch, Charles J.,' and McKean, Roland N. The Economics of Defense in theNude., Age. New York: Atheneum, 1965. ,35

Hovey, Harolct A. The Planning-Progrenuning Budgeting' Approach toGovernment DecisiokMaking. New York, New York: Frederick A.Praeger, 1968:

limes, Jon T.; et aL .The Economic Returns to Education. Eugene, Oregon:.. University of Oregon, Center for Advanced Studies in Educational

Administration, 1965. 46 pp. (ED 011 572, Document not available fromEDRS).

' Jahn, Juliui A. "The Measurement of Costs, .Utilization and EffectirneSs ofSocial Welfare Programa." Paper prepared for presentation at NatianalConference of Social Welfare, May, 1963. Philadelphia, pennsylvania.

Judy, Richard W. "CostE Theoretical and Method' (*logic' al-issues," in G.G.Somers, and W.D. Wood; (eds.), Cest-Beneftt 'Andy' s of ManpowerPolicies, Proceedings of aKorth AMericsm Conference,-Kinpton, Ontsio,

-.

11,

k

Levitan, Sar A. "Whit's Happening, Baby?-Essential Research for the War onPoverty." Proceedings of the Social Statistics Section. Washington, D.C.:'American Statistical Association; 1966.

0

Queens University at Kingstmi, InAustrial Relatio(ED 039 3341 Document not available from EDRS

Kershaw, Joseph A., and McKean, Roland N. SystemsResearch Memorandum 2473-FT. Santa MonicaCorporation, 1959.

Khatkhate, D.R. "Opportunity Cost and Its ApplicatioSocial Research, 28:60-70. No. 1, 1961.

Center. 1969. 188 pp.

naOsis end Education) .

California: The Rand

to Underemployment,"

Kotz, Arnold, (ed.). Occupational Education: Mannh,E and Programming. Vol.'1. Menlo Park, California: Stanfoid Research In titute,September, 1967.

Krutilla, John A. "Welfare Aspects of Benefit-Cos Analysis," Jilunsal ofPolitical Economy. 69: 226-35, June, 1961.

Lago, Armando: M. Beneflt-Cost Analysis of Manpower ?Wring Programs: aCritical Review. Silver Spring, Maryland: Operations' .Research Inc.,(Technical Memorandum) pp. 152-68. .

Laumann, Lydia Fischer. "Effects' of Project Headstart, Summer, 1965: ASecond Look at the Equality of. Educational Opportunity Study,"Discussion paper 47;69, Institute for Research on Poverty, Univeisity ofWisconsin, Madison, Wisconsin, August, 1969. 40

Ledebur, Larry C. "The Pioblem of Social Cost;' American Journal ofEconomics and Sociology, 26:399-415. No. 4, October, 1967,

Levine, Abraham S. "Cost-Benefit. Analysis and Social Welfare: An Explorationof Possible Applications," Welfare in Review. 4:1-11, February, 1966.

"Cost-Benefit AnalYsis .of the Work Experience Program," Welfare inReview, 4:1-9, Augist-September, 1966. . -//

Levine, Robert A. "Systems Analysis in the War on Poverty.! Office pfEconomic Opportunity. Paper presented to Operations Research Societyof America, May 19, 1966. " .

Levinson, Perry. "Evaluation of Social Welfare Prograt4; Two esearchModels," Welfare in Review, 4:5-12. December, 1966.

r

Loomis, W.G: A System for Pianniiii Implementing and Evaluating Vocational.Education. Salem, Oregon: Oregon State Department of Education,Division of Community Colleges and Vocational Education, OE 3-88-49.

. September 30, 1966.. . .

Maass, Arthur. "Benefit.COst Analysis: Its Relevancy,. to Public InvestmentDecision," Quarterly Journal of Economics. 79:208-226. May; 4966.

Manguin, Garth L. "Evaluaiing Federal Manpower Programs," Fedeial Programs'for the Devekpment of liumon ResOUIVOL A compendium of paperssubmitted to the SubCommittee on Economic Progress of the Joint.Economic Committee. : Congress, of the United States, Vol,' 1, 90th

114.

Congress, 2nd Session, Washington, D.C.: U.S. Government PrintingOffice, 1968.

. 7,Evaluating Vocational Ethication: Problems and Priorities," in ArnoldKoti; (ed.), °coups %mai !limning and Programming vol. 1, Menlo Park,Californii:-Stan search Institute;September, 1967. pp. 65.90. .

Margolis, Julius. "Secondary Benefits, isternal Econonnet-and-the-Justific. of Public Investment," Review of Economics and Statistics. 39:284-291.

No. 3, August, 1957.

. "Shadow 'Pucci for- Incorrect_ Or, Noneiistent Market Values," TheAnalysis and Evaluation of Publie Expenditure: The !TB System, in U.S:Congress,' Join t Econ omic Committee; Washington, DR.: U.S.Government Printing Office, 1969, VoL 1, pp. 533:546, 91st Congress, listSession, Joint Committee Print. . .

Marshall, F. lay, and Briggs, Vernon M., Jr. "The Administration of TrainingPrograms," U.S. Congress, Joint Economic Committee, Federal Programsfor the DeveloPment of Human Resources; a compindiurn of papensubmitted to the Subcommittee on Economic Progress. Washington; D.C.:U.S. Government Printing Office, 1968. yol. I, pp. 165.186. 90thCongress, 2nd Session. J7int Committee Print.

McGivney; Joseph H. "PPBS: A More Rational Approach to Decision.Making in- Vocational Education," American Vocational Journal. 44:63-5, May,

1969.

and . Nelson, William C. Program, Manning, Budgeting Systems forEducators. Vol. 1: An Instructional Outline. Columbus, Ohio: The OhioState 'University, The Center for Vocational and Technical Education,1969. 279 pp. (ED032 417 MF $0.65 HC $9.87).

McKean, Roland N. ::.ebsts and Benefits from Different Viewpoints," in H.G.Schaller, (ed.), Public'Expendkure Decision in the Urban Community.Washington, D.C.: &SWIMS for the Future. Distributed by the JohnsHopkins Press, Baltimore, 1963.

Efficiency in Govemmint Through Syste s Analysis. New York: JohnWiley and Sins, Inc., 1958. 19

McKechnie, Graenie H. "Discussions," in G.G. So rs and W.D. Wood, (eds),Cost Benefit Analysis of Manpower Policies. eedings of a NorthAmerican Conference, Kingston, Ontario: Queen University, IndustrialRelations Center. 1969. 188 pp. (ED 039 334, s ument not availablefrom EDRS),

Merrett, Stephen. "The Rate of Return to Education: A ritique," OxfordEconomic Papers. 18:289.303, November,.1966. .*

Mood, Mexander M., and Powers, Richard. Cost-Beneji t, Analysi of Educatk. Washington, D.C.: National Center for Educational Stati ics (DHEW)

1967. I"e 0

115

;

.0

Moss, Jerome, Jr. The balsa:lion of Occmpatiodal Progranu. Technical Report. No. 3, Minneapolis: University of Minnesota, Research Coordination Unit

in Occupational Education, September, 1968.

Musgrave, Richard A. "Cost-Benefit Analysis and the Theory of Public Finance,"Journal of Economic Literature, 7: 797406.. September, 1969. .

Neenan, William B. "Efficiency and Disiribution in Benefit-Cost Analysis,"Mimeographed. Ann Arbor, Michigan: Department of Economics,University orMichipn, 1969. 12 pp.

Novick, David7Progrem Budgeting: Program Analysis end The Federal Budget.. .

Boatok-Musachusetts:. Harvard University Pink 1965., .

Prest, A.K.; and Turvey, R. 'tCost-lienefit AnalylitrkSurveyThe_EconomicJournal 75:683-735,-Deceniber, 1965. 12

Preston, Maurice. "The Theory of SpilloVers, and Its Connection withEducation," Public Finance. 21:184-99, No. 1-2, 1966..

Quade, E.S. Cost-Effectiveness Analysis: An Introduction and Overview. Santa. Mónica,,California: Rand Corporation, 1965. - .

.Systents AndysitTechniques for Planning-Programming-Budgeting. SantaMonica, California: The Rand Corporation, 1966.

RaynaUld, Andre. "Benefits and Costs: Theoretical and Methodological Issues:Discussion," in C.C. Somers and W.D. Woods; (eds.), Cost-BeneJlt Analysisof Manpowtr Poikks; Proceedingi of a North -American Conference.Kingsttlit Ontario: Queen's University Industrial Relations Center; 1969..

. pp. 37141. Complete Text 188 pp. (ED 039 334, Document not availablefrom EDRS). .

Rider, Melvin W. 'Criteria for Judging the Effectiveness of-Training Programs."Stanford, California: Department of Economics, Stanford University,

. .

Mimeo, 21 pp.

Rees, Albert. "Review of lfuman (apital,' by Gary S. Becket". AnsericanEconomic Review. 55:958-60. Septiamber, 1965..

Rein; Martin, and Miller, S.M. "Poverty PrOgrams and Public Priorities,'Trans-Action, 4:6071, September, 1967. ,

Reuber, Grant L. "The Social Discount Rite: Discussion," in G.G: Somers andW. I). Wood, fedi.), Cost-Benefit Analysis Of ManpowerProceedings of a North -Amerkan .Conference. May 14-15, 1969.* pp.

_ 88-94. Kingston, Ontario: Quito's University, Industrial Relations_Centerfor -Stidfirirr-Vocational_ and Technical Education, 1969.-(ED 039 334;Document not avallabk from EDRS). -

RoW, Peter H.. "BOobitraPs end htfalls in the EValuation ofSocill *Action,Progranis," Proceedings Of theSodalStatiltics Section. Ifashingion, D.C.:

*,* Anierican Statistics Astociation, 1966.

"Practice, Method and 'Theory In Evaluating Social-Action Prograrns4', in. James :L. Sundquist (ed.), 00 Rghting poluty; Peispectives. from .

116

011

Experience, with the assistance of Corinne S. Schelling, New York: Bask, Books, 1969. pp. 217-234.

Schultz, Theodore W. "Investment in Human Capital: Reply," AmericanEconomic Review, December, 1961.

"The ate of Retun in Allocating Invesiment,Resources to Education,"Journalçf Human Resources. 2:293.309. No. 3, Summer, 1967.

"Reflektions on Public Approaches to Minimize Poverty," Investment inHuman Capital Series. .65:6. April 15, 1965. Chicago, Illinois: Universityof Chicago, 1965. Mimeographed. .

Sr.ott, Norman. "The Planning and Cost-Benefit Analysis of Social Investments,"International Institute for lAbor Studies Bulletin. Oebniary, 1967. ,

Scriven M. The Methodology of Evaluation. Washington, D.C.: Bureau ofResearch, U.S. Office of Education. 1966.

Sewell, David 0. "A Critique o rittaenefit-Analyals--of_Triiiiiing,'L MonthlyLabor Review: 9045-5. Sqternber, 1967.

_: Training the Poor: a Benejlt-Cost Analysis of Manpower Programs In theU.S. Antipoverty Program. Kinpton, Ontario: Queen's University.Industrial Relations Center, 1971. Research Series 12. 153 pp. (ED 050271, Document riot available froni EDRS).

Sewell, W.R.D., et al "Guide to Benefit..Cost Analysis," A paper prepared forthe Resources for Tomorrow Conference, MontrIal, October 23-28, 1961.49 pp.

Sh2ffer, Harry G. "Iwiestment in Human Capital: Comment," AmericanEconomic Review, Vol. 51:1026, December, 1961.

Sherwood, Clarence C.' "Methodological, Measurement and Social Action, Considerations Related to the Assessment of Large Scale Demonstrations

Piograms," American Statistical Association Social Statistics . Section.Proceedings, 1964. pp. 199.207.

Sloan, Stanlei. "Eviluating Training Programs for the Madvantaged," PeriOnnelAdministrator, 15:15-16. November/Deceraber, 1970. . .

Smithies; Arthur. A Conceptual Franiework for the Prognvn Budget. SantaMonica, California: Rand Corporation, 1964.

Sobin , Bernard. "Interest Rate Concepts for Analysis of GovernmentExpenditures," A paper read at the 30th National Meeting of theOperations Research Society of America., Durham, N.C. October 17.19,1966. Mimeo. 14 pp.

Solomon, Ezra (ed.) The Management of Corporate Capital, New: York: The

Solow, Robert M. "Review of 'Human Capital,' by GaryPolitical Ecvnomy. 73:552.53. October, 1965.

Somers, Gerald G. "Evaluation of Manpower Policies," Federal Programs kr. theDevelopment of Human ( Resource" A compeneum of papers submitted tothe Subcommittee on Economic Progress of the Joint EconomicCommittee, Congress of the United States,. Vol. 1, 90th Conpess, 2ndSession, Washington, D.C.: Government Printing Office, 1968..

"Research Methodology in the Evaluation of Retraining Programs,"Madison, Wisconsin: University of Wisconsin,IRRI Reprint Series, No. 61.Reprinted from Labor and Automation. Bulletin, No. 1, 1964. Geneva,1965. 21 pp.

. "The Response- of Vocational Education to Lab& MarketChanges, 'Voiational EducationSuppkment to the Journal of /lumen Resources,3:32-58 (a), 1968.

Spiegelman, Robert G. "A Benefit-Cost Framework for Education," in ArnoldKotz, (ed.), Occuptstiond Education: Planning and Progremming. Volume2.. Menlo Park, California: Stanford Research Institute. 2:359-377.September, 1967. Complete Text 279 pp. (ED 017 734 MF $0.65 HC

, s-dCostlen oeflt_MJAA"Evaluate Educational Programs. ProgressReport. Menlo Park, California: StanfrifdlteTeirc1rInstituteT1

Steiner, Peter 0. "The Role of Alternative Cost in Project Design and Selection,"Quarterly Jounwl of Economics. 79:417430. No. 3, August, 1965.

StOber, William J.; Falk, L.H.; 'and Ekelund; Robert IL," Jr. "Cost and Bias inBenefit-COst Analysis: Comment," Southern Economic Journal.34:563-568. No. 4, April,_1968.

Stromsdorfer, Ernst W. "Economic Concepts and Criteria for Investment inVocational Education," in Arnold Kotz, (ed), Occwitional Education:Planning and Programming. -Vol. 2, Menlo Park, California: StanfordResearch Ingitute, September, 1967. Pp. 331-357. Complete Text 279 pp.(ED 017 734 MF $0.65 HC $9.87).

Hu, Teh-wei; and Lee, Maw Lin. "Special Problems. in the EconomicAnalysis of Investments in Human Resources,".Socio-Economic PlanningSchnee', 5:275-283, 1971.

Str ong, Alan, and Whittingham, Frank. A Proposed Methodology forCost-Benefit Analysis of Goverment Sponsored Trainintin-Indusby.Toronto, Ontario: Department of Labor, Research Branch. 1970.

Suits, Daniel B. "Use of Durruny Vaiiables in Remakes' Equations," Journal ofthe American Statistical Anociation, 52:548-551, 1957.

Swift, William 1., 'and Weisbrod, Burton A. "On the MonetarY Value ofEducation's intergeneration Effects," Journal of Poktkal Economy.73:643-649. No. 6, December, i%5.

Taylor, Graeme "Office of Economic Opportunity: Evaluation of TrainingPropams," in Harley H. Hinrichi and Graeme M. Taylor, (eds), Program

Budgeting and Benefit-Cost Analysis, Pacific Palisades, California:Goodyear Publishing Company, 1969.

--

Technomics, Inc. The Feasibility of Cost-Effectiveness for Title I, Public Law89-10. Final Report: Chicago, Illinois, January 31, 1966. 185 pp. (ED 013.370 MF $0.65 HC $6.58). . .

Tella, Dorothy; Dobson; Allen; and Commins, William. A Model for ManpowerDaring Evaluation. McLean, .Virginis: Planning Research Corporation.November, 1970.

ThOmas, J.Alan. An Economic Approach to Systems Analysis Los Angeles,California: Paper prepared for the annual meeting of the AmericanEducation Research Association. February, 1969. 15 pp. (ED 029 380 MF$0.65 HC S3,29).

Thurow, Lester. Investment in Human Capital. Behnont, California:. WadsworthPublishing Compsay, tnc. 1970. 22

Turvey, Ralph. "On iergences Between Social Cost and Private Cost,"Economic& 30:309,313. No. 119, August, 1963.

United Nations Reseuch Instiftite for SocW Development .and Office of Social,Affairs. Cost-Benefit Analysis of Soda Projects. Report of a meeting ofexperts held in Rennes, FINICe. September 27 tO OCtober 2, 1965. ReportNo. 7, Geneva, April, 1%6.

U.S. Conguss; Joint on valuation oPublic Expenditures: The PPB Systan. A CompersdLon of papers.Washington, D.C.: Government -1Printing Office, 1%9. Vols. 1, 2, and 3,91st Congrees,'Ist Session, Joint Committee Print. .

, Subcomtnittee on Econoiny ht Government; Economic' Analysis of PublicInvenrrent Dedsions: Interest Rate Policy ard Discounting Analysis:Hearings. 90th Congresi, 2nd Session, 1%8.

, Subconunittee on Economy in Gournment; "Interest Rate Guidelines- for Federal Decision-Making," Harings, 90th Congreu, 2nd Session,

January 29, 1968. Washington: 1J.S. Gournment Printing Office, 1%8.

, Subcommittee on Economy in Government. ThePlanning-Programming-Budgeting System:. :Progress end Potentials:Hearings 90th Congress, 1st Session, 1%7.

US.. Office of Education. "Evaluating the Program," Organisation AndOperation of a Local Progran of Vocational Education. Waddngton,.D.C.:Office Of Educ.stion, Department of Health, Education and Welfare. 1%8..pp. 76.79. COMpkte Text 96 pp. (ED 022 061 MF $0.65. IfIC 83.29).

US. Office Of Manpower. Policy,. Evaluation and Research: An Aimed. toCost-Effectiveness; Studies in Manpower Program; planning stiff paper.Washington,- D.C.: U.S. Office of Manpower Policy, Evaluation mdResearch, 1966.

Veilserhill, Ralph. Methods ks the Embalm Of Programs for Poor Youth.Chicago, Illinois: National Opinion Returch Center,.1968.

:

119

Ward, J.T. "Cost-Benefit Analysis" Journal of Public Administration. NewZealand: March, 1967.

Weil, R.L., Jr. "Allocating Joint Costs;" American Economic Review, pp.1342-1345, DeceMber, 1968. 35 °-

Weisbrod, Burton A. "Benefiti of Manpower Programs: Theoretical andMethodological Issues." in G.G, Somers and W.D. Wood, (eds.), CostBenefit Analysis of Manpower Policies, Proceedings of a North AmericanConference. Kingston, Ontario: Industrial Relations Centre, Queen'sUniversity at Kingston, 1969. 188 pp. (ED 039 334, Document notavailable from EDRS).

. "Concepts of Costs and Benefits," in 5.11. Chase, Jr., (ed.), Problems inPublic Expenditure Analysis," Washington, D.C.: The Brookings Institution,1968.

- "Conceptual Issues in Evaluating Training Programs," Monthly LaborReview. 89:1091-1097. Ot tober, 1966.

"Expenditures on Human Resources: Investment, Income, Redistributionor What?" U.S. Congress Joint Economic Committee. Federal Programsfor the Development of Human. Resources; a compendium of paperssubmitted to the Subcommittee on Economic Progress. Washington, D.C.:Government Printing Office, 1968. Vol. 1, pp. 80-83. 90th Congress, 2ndSession. Joint Committee Print.

External Benefits of Public Education, An Economic Analysis. Princeton,New Jersey: Princeton University, Industrial Relations Section, 1964. 27_ "Measuring cono , nancial Aidand National 'Purpose. College Scholarship Service, pp. 12-27. Princeton,New Jersey: CaB, 1962.

.

Wholey, Joseph S., et al. Federal Evaluation Policy; Analyzirg the Effects ofPublic Program. Wr.shington, D.C.: Urban Institute, 1970.

Williams, Walter, "Developing an Evaluation Strategy, for a Social ActionAgency, Journal of-Hunam Resources. 4:451-465. Fall, 1969.

Wiseman, Jack. "Cost-Benefit Analysis in'Education," The Southern EconomicJournal. 32 (2,part 2). Supplement: 1-12, July, 1965. 24

H. VOCATIONAL EDUCATION.. .1

Almaiode, Richard L A Follow-Up Study of Junior College Hotel andV Restaurant Education in Florida. Tallahassee, Florida: Floridi State

. UniVersity, School of Business', September, 1968.. .

Anderson, Ernest. F. Differentiarpsts of Curricula in Congrehensive JuniorColleges. Urbana, Illinois: Buteau of Educational Research, University ofIllinois, 1966. 94 pP. (ED. 013 085 MF $0.65 HC

120

Altman, James W., and Dueker, Richard 1..,-An Analysis of Cost and PerformanceFactors in the Operation and Administration of Vocational Programs inSecondary Schools. Final Report. Pittsburgh, Pennsylvania: AmericanInstitute for Research in Behavioral Sciences, October, 1967.

, and Morrison, E.J. School and Community Factor: in EmploymentSuccess of Trade and Industry Course Graduates, Pittsburgh, Pennsylvania:American Institute for Research, Al R-E26-8 I/66-FR, August, 1966.

American Institutes for Research. Prolect TALENT, Palo Alto, California.

Anderson, Ernest F., and Spencer, Jimes S. Report of Selected Data andCharacteristiis, Illinois Public Junior Colleges, 1966-67. Springfield,Illinois: Illinois Junior College Board, 1967.

Austin, John J., and Sommerfield,- Donald A. An Evaluation of VocationalEducation for Disadvantaged Youth. Muskegan, Michigan: MuskeganPublic Schools, Report No. BR-6-1831, April, 1967. 212 pp. (ED 016 097MF $0.65 HC 59.87).

Blume, Paul Roundtree. An Evaluation of knitutional Vocational Thainbt:Receiied by American Indians Through the Muskogee, Oklahoma AreaOffice of the Bureau of Indian Affairs. Unpublished doctoral dissertation,Stillwater, Oklahoma: Oklahoma State University, 1968. 261 pp. (ED 023862, Document not available from EDRS).

. .

Bushnell, D.S. "Evaluation Research Studies," Americo Vocational Journal,41:17-18, 1966.

Cap, Robert N. Cost Analysis of Selected. Education hograms bi lova,. DesMoines, Iowa: Iowa State Department Of Public Instruction, 1968.

..Cost Analysis Of Selected Education Programs in the Area Schools ofIowa. Ames, ono: owiiSlgeUtilversity7)fScience-rd3eelnoloch1968. 23 pp. (ED 025 639 MF $0.65 HC $3.29).

California lob Training and Placement CouriCil. Final Report of tire California .

Job Training and PlaceMent CounciL .Sacramento, California: CaliforniaDaining and Placement Council, 1968.

Carroll, Adger B. Value of Human Capital Crsated by Investments in TechnicalEducatiOn. Unpublished doctoral dissertation. Raleigh, North Carolina:North Carolina State.University, 1966.

, and Ihnen, Loren A. Costs and Returns of Technkal EdsicatiOn: A PilotStudy. Raleigh, Mirth Carolina: North Carolina. State University. July,1966. 52 pp, (ED 015 247 MF $0.65 HC $3.29). 56 : .

Costs and Returns for Investment in Technical SchOoling by a Group ofNorth Carolina High School Gradates.. Economics Research Report No. 5.Raleigh, North Carolina: Department of Economics, North Carolina StateUniversity, December, 1967.

"Costa and Returns for. Two Years 'of Post-Secondary TechnicalSchooling: &Pilot Study," Journal of Political Economy. 75:862-73. No.6, 1967.

121

Cohn, Elchiman; Hu, Teh-wei; and Kaufman, Jacob J. The "Added Costg-of____L__Vocatioratl Education in Michigan, University Park, Pennsylvania: Institutefor Research on Human Resources, The Pennsylvania State University,November, 1971.

Coleman, James S. et aL Equality of Educational Opportunity, National Centerfor Educational Statistics, U.S. Office of Education, Washington, D.C.:Government Printing Office. 1966.

Conley, Ronald W. "A Benefit-Cost Analysis of the Vocational RehabilitationProgram," Journal of Human Resources, 4:226-252; Spring, 1969:

Corazzini, Al "The Decision to Invest in Vocational Education: An Analysis ofCosts and Benefits," Vocational Education Supplement to the Journal ofHuman Resources, 3:88-120, 1968. 53 .

.

Vocational Education, A Study 4 Benefits and Costs (A Case Study ofWorchester, Massachusetts) Princeton, New Jersey: Industrial RelationsSection, Princeton University, 1966. 33, 92

"When Should Vocational Training Begin?" in Cathleen Quirk and CarolSheehan, (eds.), Research in Vocational and Technical Education(Proceedings of a Conference, June 10-11, 1966) Madison, Wisconsin:Center for Studies in Vocational Education, University of Wisconsin, -1967, pp. 193-201. Complete Text 287 pp. (ED 021 976 MF $0.65, HCnot available from EDRS).

Ciancara, Joseph G. Input-Output Relationships Among Selected Intellectual. Investments in Agriculture, Unpublished doctoral dissertation,

Minneapolis, Minnesota: University of Minnesota, 1964.

Davie,- Bruce, "Benefit-Cost Analysis of Vocational Education: A Survey," inArnold Kotz, (ed.), Occupational Planning and Programming, Vol. 2,

en309-330.

. "Cost-Benefit Analysis of Vocational Education: A Survey," Hearingsbefore the General Subcommittee on Education and Labor, House ofRepresentatives, 90th Congress, 1st Session on MB. 8525 and Related

3 Bilk A .Bill to Amend the Vocational Education Act of 1963. Pitt I.Washington, D.C.: Government Printing Office, 1968. pp. 105.117.

Drouet, Pierre. "Vocational Training Costs: Results of a Pilot Study and an'Essay in Methodology," International Labour Revlev y7:115-133,Febmary, 1968.

t..

Duis, Harold, and Sander George. "Employment of Vocational ProgramGraduates," AmericanEducation. 4:29. February, 1968.

Dupree, Robert L. A Cost-Benefit Study of Post-High School TechnicalEducation 'in Oklahoma. Stillwater, Oklahoma: Oklahoma StateUniversity, May, 1968, 50 pp. (ED 026 488, Document not available fromEDRS).

- 122 139

Eggenberger, Ulrich L. An Analysis of High School Vocational Agriculture fromEvaluations of Graduates in West Texas. Unpublished doctoraldissertation. Ames, Iowa: State University of Iowa, 1964.. 166 pp. (ED020 334, Document not available from EDRS).

Eninger, Max U. The Process end Productof T &I High School Level VocationalEducation in the United States: The Product. Pittsburgh, Pennsylvania:American Institute tor Research, AIR-D7 1-1 9/65-FR, September* 1965.445 pp. (ED 012 315 MF $0.65 HC Sl 6..45). 37,70 .

The_Procestand_hoduct.. of T & I High -School LevelVocationalEducation in the United States: Vol. II The Process Variable. Pittsburgh,Pennsylvania: Educational Systems Research Institute, 1968. 654 pp. (ED024 797 MF $0,65 HC $23.03).

. Project Metro-Evaluation Data do Vocational Educational Progrens inMajor Metropolitan Areas. Vols. I, II; *id Tit Pittsburgh, Pennsylvania:Educational Systems Research Institute, 1971, 104

Fernbach, Susan, and Somas, Gerald G. An Analysis of the Economic Benefitsof Vocational Education at the Secondary, Post-Secondary and JuniorCollege Levels. Preliminary Report. Madison, Wisconsin: Center forStudies in Vocational, Technical Education, University of Wisconsin. May,1970. 49, 56

Garbutt, DOuglas. Training Costs with Reference to the Industrial Training Act.London: Gee and Company, Ltd. 1969. 210 pp. (ED 032 481, Documentnot available from EDRS).

Goff, Maurice L. Survey of Presrist Methods of Fcillow-Up - of PublicPost-Secondary School Graduates in Cooperative end PreparatoryVocational Programs end Development of a Follow-Up Procedure.Unpublished doctoral dissertation. Laramie, Wyoming: University ofWy

EDRS).

Grigg, Charles M. et al Vocational Rehabilitation of Disabled Public AssistanceClients, Urban Research Center, Publication No. 6, 19C8. Tallahassee,Florida: Tallahassee Institute for Social Research. Florida Rae University,1968.

Haines, Peter, et al How High School Cooperative Trainees Fair in the LaborMarket, Phase D: A Follow-Up Study of 1965 Graduates Ten._MonchsAfter Graduation. Educational Research Series No. 39.*.East Lansing:Michigan State Unirs4ty.--- .

Hansen,y.L.;-Welibrod; B.A.; and Scanlon, WI "Schooling and Earningi of,-- ----Low Achievers," American Economic Review, Vol..LX, pp. 409418,

June, 1970. 78

Education in Secondary Schools, Pal Alto, California: American Instituteorfre'Hawkeidge,DavidGgetaLAStudyedhognmufor nivtional

for Research, AiR-848-1/70-FR, January, 1970..

123

e

Hirsch, Werner, and Marcus, Morton. "Some Benefit-Cost Considerations of. Universal Junior College Education," National Tax Journal. pp. 48-57.

March, 1966.

Hu, Tehowei, et al. A Cost Effectiveness Study of Vocational Education: AComparison of Vocational and Non-Vocational Education in,SecondarySchools, Anal Report University Park: Pennsylvania State University,institute for Research on Human Rekfurces. March, 1969. 37, 53, 71

Lin; and Stromsdorfer, Ernst W. "Economic Returns toVocational and Comprehensive High School Graduates," The Journal ofHuman Resources. 1:25-50, Wnter, 1971.

Kaufman, Jacob J. and Lewis, Morgan V., The Potential of VocationalfduRtion: Observations and °Conclusions Based on a Study of ThreeSelected Cities in Pennsylvania. University Park, Pennsylvania:Pen nsylvinia State University, Institute -for Research on Human-Resources, 1968. 171 pp. (ED 023 902 MF $0.65 HC $6.58): 53

et al. An Analysis of the Comparative Costs and Benefits of VocationalVenus Academic Education in Secondary Schools. University Pirk,Pennsylvania: Pennsylvania State University. Institute for Research onHuman Resources. 1967.

The Preparation of Youth for Effective Occupational Utilization, theRole of the Secbndary School in the Preparation of Youth forEmployment. taJniversity Park, Pennsylvania: Pennsylvania StateUniversity, institute for Research on Human Resources. February, 1967,448 pp. (ED 011 060 MF $0.65 HC $16.45).

Employment. University Park, Pennsylvania: Penntylvania StateUniversity, institute for Research onliuman Resources. February, 1967.

Kay, Evelin R. Inventory of Vocational Education Statistics Available in FederalAgencies, National Center for Education Aitatistics, US.Office cf Education,Department of Health, Education, and Welfare, OE-80069, Washington,D.C.: Government Printing _Office, May, 1970.

Kirby, Frederick C, and Castagna, Paul A. Benefit Cost AnaOsis of TAT Phase I,Worker Training. Training and Technology Project. Special Report.Training and Technology Project Phase I. Oak Ridge, Tennessee: AtomicEnergy Commission, July, 1969. 34 .pp..(rD 034 039 MF $0.65 HC13.29).

Kraft, Richard H.P. Cost-Effectiveness Analysis of Vocational-TechnkalEducation Programs.' Tallahassee, Florida: Department of EducationalAdministration, Educational Systems andylanning Center, Florida StateUniversity, 1969.

124

Kunce, Joseph T. "The Effectiveness of Poverty Programs, A Review,"Columbia, Missouri: University of Missouri, 1969. Regitinal RehabilitationResearch Institute, Research Series; No. I. pp. 127-171.

Lee, Maw ,Lin; Hu, Teh-wei; and Stromsdorfer, Ernst W. "Costs and Returns to -

Vocitional Education," American Vocationil Association Yearbook,1971, forthcoming. . .

Lindman, "Erick 1....____Rnancing Vocational Education in the Public School,National Educational Finance Project, Special Study No 4,11.os Angeles,California: Graduate School of Education, University of California, ,LosAngeles, 1970.

Martin, Albert H., and ThoMblad, Carl E, Report of Selected Data andCharacteristics of Illinois Public Junior Colleges, 1960-70, Springfield,Illinois: Illinois Junior College Board, 1970.

Miles, Guy. H. Preliminary Phase: Effects. of Vocational Training end OtherFactors On Employment Experience. Minneapolis, Minnesota:North StarResearch and-DevelopmetA.Institute, April, 1966.

Morsch, William C. Study of .Community. Colleges and Vocational .TrainingCenters: Cost Analysis. Washington DC.: Bureau of Social Science

.. Research, Inc. 1970.- 92, 103. ...

linenNotth Carolina State- University at Raleigh-. Costs end Returns for Inves t in. Ind trial Education. A Pilot Study, Raleigh, North Card : U.S.

. Depar fit of Labor, Office of Manpower, Automation and Training,ri.d.: -

Persons, Edgar A.. "Adult Educelon: Does it Pay?" Apicultreal Education*Paine. 41:42-43, August, 1968.

. The Farmer and His Educational Investment: What Are the Relationshipsof this Investment to Fern Success? Unpublished doctwal dissertation.Minneapolis, Minnesota: University of Minnesota,. 1966. DissertationAbstracts, 27: 1622-A; 1966. (ED 020 346, Document not available fromEDRS). -

et aL, An Economic Study of Investment Effects of Education inAgriculture. Final Report. Project No. OE 427 65. St. Paul, Minnesota:Department of Agricultural Education; University of Minnesota, April,1968. 175 pp. (ED 021 989 MF $0.65 HC $6.58). .

. Investments in Education for Farmers, Summary of an Economic Studyof the .Investment Effects of Education in Apiculture. St. Paul:Departs-tient of Agricultural Education, University of Minnesota, January,1968. 54 pp. (ED 021 951 MF $0.65 HC S3.29). ,

Pe terson, Leroy J. "Cost-Benefit Theory in Vocational and TechnicalEducation," Madison, Wisconsin: University of Wisconsin, Center for

: Studies in Vocationai Adult and Technical Education, Unpublishedmanuscript, 1969.

123125

eost-High School - Vocatiiinal-Technical EducatiOn in Wisconsin. Madison,. 4.

, Wiiconsin: Wisconsin State Board of Vocational and Adult Education,1960, 1962: .. s . - ., .

Priebe , Don ald .. , "A Follow-np of Vocational Agriculture Graduates,"Agricultieral Education. 40:273, June, 1968. ,. .

0.

"Project Talent: A S.Yeat follow-up information on high school .graduates of1960." An Analysis of Vocational EduCation in o'er Secondary Schools.Washington, D.C.: U.S. Office of Education, Office of Program .Planning

. and Evaluation, July,1969. ., .

.

Reinhart,. Bruce,. and Blomgrek,- Glen H. Cost-&nefit Andysis-llrade and .

Technical Education.-Final Report..Los'Angeks: University of California.Division of Vocationil Education, August, 1969: 82 pp. (ED 034 056 MF$0.6511C $3.29). .

Report of the, Analysis Group,NEW Vocalional Education Review Task Force,Vols. I and II, Prepared for Dr. Jamit Albert, Oeputi.Auistant Secretiry,Evaluation and ?4,3nitoring, DepartMent of Health, Education, andWelfare. Silver Sprig) Maryland: Operations Research) Inc., September 25,

, 1970:

Robergs, Roy. W7 Evaluation of the Effectiveness of InSbleCtiOn in Agricultural... Mechanics for Vocational Agriculture Students in Arkanis.who Enter-7\ Non-Ferrning Occupations. Fayetteville, Arkansas: .University of Arkansas,

College of Education, July, 1965: 53 pp. .(ED 013 875 MF $0.65 tIC: -.

. , .

'Scales, Eldridge E. Current Operating CMS of Two!Yeer Colleges in the South,Atlanta, Georgia: Southern AssociatiOn Of Caleps and Schools, 1969.

8iftriver, William R., and Boitilby5ROger L. The Effects of GoCetional MiningOn. Labor Force Experience.' An Analysis of the TennesseeVocational-Technical School . Systeme -Memphi:, Tennessee: Center ,for 'Manpower Studies, Memphis State University, February, 1971. 149Pp.,(ED 051 388 MF $0.65 HC $6.58)..

Sharp, Lanre . M., and Myint, Thelma. Graduates. of ,Vocationil-TerminalPrograms bi Junior Colleges. Washington, DX.: Bureau of Social Science'

. Research. September,. 1970. .91

SOciar Educational Research and Development Inc. An Identification and .

. Analysis of Effective $econdary Level, rocitionitl Programs for theadventaged. Final Report. Washington, D.C.:.; U.S. Department of

Health, Edueation, and Welfare, Office of EducatiOn, December, 1968.314 pp. (ED 029 105 MF $0.05 S13.16). .

Somas, Gerald- G. "Vocational Education. RePort. of's Conference Sponsored ..

by the Brookinp Institution," Supplement to The Journal of limoResources, 3'. 1968. 140 pp.

; Sharp, Laure M.; and.Myint; Thelmi,The Effectiveness of Vocatkinal andTechnical Programs: A National Follow-Up Study. Madison, Wisconsin:

.

126

:."

0

Center for Studies in Vocational and Technical Education, University of. WiscOnsin. 1970.

Stephenson:Don. The Effectirenen of eke Contra Costa College Dental AnistingProgram' -in heparin its Graduges for .Employment. San Pablo;:,,California: *Coiltra Costa College, 1967.

Stromsdorfer, Emst W.; Hu, Teh-wei; and Lc*, Maw Lin. "Theoretical andEntpirical .Problems in the Analysis of the Economic Costs of Vocational'Educattonf .Proceeditegs of the" Social StatistiCs Section. Washington;D.C:: 'American Statistical Association, 1968. .

et aL "Vacational-Technicel Eiucation as an Investment," AnseticanVoCational Associetkm Yealmok, 1971, Forthcoming.

Swanson, 'Chester Leadenkip Role, Functions, Procedures and Administrationof VocationalTedatical Education Agencies at the State Levels. Volurie3. Worm Con AnalYses of Vocationd-Technks1 Education in a Amor

. College and in a Junior College and a a Unified School lIntrict.California: School of, Education, Universiti of California, March, 1969.

. .

Sianson, J.C., and Kramer, E.G. "Vocational. Ethication Beyond- the kligh. SchOol,".Nationd Social Study Education Yearbook' 1964. Part 1. 1965.pp. 168,185.

, . 4

Taus:is:Michael K. "AnSconomic Analysii of Vocational Education in the New9 York 'City High SChools," Jowl* of Hunan Reaources, 3:59-87.

Supplement. 1968. 53, 102 :Thompson, .R.J. "`OUtcomes in . Vocational and Technical Education in the

Junior College," National Business Education Yearbook, 1:371-83..1963.

US. Bureau of the Censui. Household Question's*. - 5 Percent;.Content.'Reinterview study, (E-G), 19th Decennial Census, U.S. Department ofCommerce. Washington, D.C.: Goventment Printing Office, 1970.

Reconciliation QUestionnitire' (Siage .1 A Content. ,Reintervient Study(E-9),,i9th Decennial Census. U.S. Department of Commerce, Washingiok

GoVernment Printing Office, 1970. .`

US. Office of Education. Vocational. and Technical Education: A Review of .

Activities In Federally Aided Programs,, Fiscil Year. 1963. Washhigton,D.C.: U.S. Department ofilealth, Education ane Welfare, 1964. . -*

Venn, G:, and Marchese, T. J., Jr. Man, Educatkm ind Work, Post SecondayVocational and Technical Education.- Washington, D.C.: American CounCilon Education, 1964. ,

. .

Vocational-Technical Education. 1968. A. Summary Report of a Study of theEffect of the Area Vocational Technical Schools in the State of Minnesota.'Minneapolis, Min neeota: Minnesota Rematch Coordiaiting Unit,:Occupational Educatiosi;JUly, .1968 p .

Wood, Eugene S. An Evatu31' Illinois Post-Hight. School EdiwationalProgratns is Agikulture.'. Publication No. . 28, Carbondale, Illinais:Southern Illinois University, School oftApieultUre:September, 1967." "-

1,35

..

Zaharchuk; T., and TOyiwsky, R. Cost kid Benefit Study of Post-SecondaryEducation in Ithe Province otentarid; Stool Yea', 1968-69. Phase 1-A,Vol: 1, Toronto, Canada: Systees Research GroUg; April, 1971.

111. MANPOWER PRC,OS

A and R Reports No. 11:,"Job Corps Benefit/Cost Study,". Evaluation andResearch Branch, Plins and Evaluation Division, Plans and ProgramsDirectorate, Job Corps, Office, of Economic Opportunity, Washington,

t.D.C. (iro date).

Baldwin, Stephen E. The Impact AA Governmental Programs on.' theEmployability of the .}TOth in the Seattle Csabor-Market. Unpublished.,dottoral. dissertation', Seattle, Washington: University tof Washington,1968.

Barth, Peter S. "An Assessment of FourVanpower Programs," presented beforethe Operations Research SocietypDallas, Texas, May 5;1971. 105

A Benefit Cost Analysis of the, South Carolina MDTA Pipgram. PreliminaryReport, Columbia, South Caroline: South Carolina liniversity, ButeaU ofBusiness and Economic Research. (no date).

Besen, Stattley M., et al. "Cost-Effectiveness Analysis for tho War On Poverty:, "in .Thomas A. Goldman, (ed.), Cost. Effectivenels Analysis,Chipter 10,New York; Frederick A. Praeger, .1967. 15 pp. (ED, 024 877, Dociimentnot aVailable from EDRS). . . .

.

Bjorkquist, David C. Effects of Field and Job-Oriented Technkal Retraining onManpower Utilization of the Unemployed. University Park, Pennsylvania:-

. ,.

Pennsylvania State University, August, .1968. 119 pp. (ED 023 910 MF

-....-.$0.65 HC $6.58): ,

4:371-4' 2

Borus, Michael E. "A Benefit.Coat Analysis of the of Rettaining theUnemploYed," Yale Economic Essays.

"The Cost .of' Reiraining the. Hard-Core Unernployed; An EconomicEvaluation of the 1963 and 1965 Amendments to . the. Manpower

,.,..:.. ' DeVelopment and Training- Act," Labor Lew Journal, )6:574-583.. ,,i.i....- September, 1965. ' .'.

The Economic -Effietiveneas of.Retr-ining the Unemploied. Research. , . I .

Report to ,the Fedetal Re3erve Bank.. of Boston; No. ; 35. Boston,. Massachuretts: The Federal Reorve aar:.:- of Boston, 1966. 28 .', ,.. ,. . .

The Economic Effectivenesk of Retraining the Unemployed: A Study ofthe ..Betielits and Costs of Retraining the Unemployed. Based on the

. Experience of Worken in Connecticut.. Unpublielsect doctoral. dissertation,:Yale Universitif ;1964. 228 pp. (ED 010 865 MF $0.65 HC $9.87). 69.

.

.-.:"The. Effects of Retraining the Unemployed in Connectecut,r_in GeraldG. Somers,. (ed.), Retraining the Unemployed pp. 125-148. Madison,

128

:4-

Wisconiin: University .of Wisconsin Press, /1968. Complete .Text 350 pp.(ED 021 198, Document not available froM EDRS).

. "Time Trails in the Benefits frOm Retraining in Connecticut,"Proceedings of the Twentieth Annual Winter Meeting of the IndustrialRelations Research Association. 20:36-46; 1967. 15 .pp..(ED 028 319,Document ma available from EDRS).

,' et aL 'PA Benefit Cost Analysis of the Neighborhood Youth Corps:`iheOut-of-school program in Indiana," Jownal- of Humeri Rewurcet5:139-159. Spring, 1970. 69 .

Brazziel, William F.. "Effects of General Education in Manpower Programs,"Journal of Humeri Resowcei 1:39-44. Summer, 1966: .

Iluenaveritura, Angeles. Follow-Up Study of MDTA E and D Project Conductedb" the Michigan Cetholic Conference, Lansing. Washington, DC.: Bureauof Social Science Research, Inc.,' Mar* 1967. 2 pp. (ED 015 .315 MF$0.65 HC $6.58).

*Cain, Glen, G. Benefit/Coif Estimates for Job Corp' adison, Wisconsin:University of Wisconsin, Department of Economics, InMitute for Researchon Poverty. 1967. 52 pp. (ED 037 495 MF $0.65 HC $3.29). 63

, and Somers, Gerald. "Retraining the Disadvantaged Worker," in CathleenQuirk and Carol Sheehan, (eds.), Research in Voattional and TechnicalEducation, Madison, Wisconsin: Center for Studies in Vocational andTechnical Education, University of Wisconsin, 1967.,pp.27-44. CompleteText 287 pp. (ED 021 976 MF $0.65 HC not available from EDRS).

and Stromsdorfer, Ernst W. "An Economic EvaluatiOn of Government,Retraining Programs in West Virginia," in G. G. Somers, (ed.), Retrainingthe Unemployed, .Madison, WiscOnsin: University of Wiwongin Press,1969. 350pp. (ED 021 198, Document not available from EDRS). 86

Celebrezze, Anthony J. Dwining Activities tinder the Manpower Development. naining Act. Washington, D.C.: U.S. Department of Health, Education,

, and Welfare: Circular No. 706, FebrUary 28, 1963.

Center for the Study of Unemployed Youth. A Study of gte Menming andEffects* if the NeOboritood. Youth Corps: On Nam- Youths Who areSeeking' Employment. New York: Graduate School of Social Work, New

. York University; March, 1968. '.

Chamber of CoMmeice.Of -the United States of AMerica.. Youth and the War onPoverty; an evaluation of the Job Corps,i.Youth Corps, and Project Reed

. Start Washington, D.C.:Chamber Of Commerce, 1966.

CheRer; Herbert A: "The :Retraining Decision in MessaChusetts: 'Theory and.. ,

Practicer, in G.G. Soineis, (ed.), Retralibig the Unemployed,:Madison,Wisconsin: University of Wisconsin Press;.,1968. 350 pp.' (ED 021' 198,Document not available from EDRS). ' 1.

Cohen, Bonnie :R., and Yonkers, Ann H. Evaluations of the We. on Poverty.', Educational Fromm:, RMC Report. . Bethesda, Maryland: Resource

129

-Management Corporation. 1969.. 146 pp. -(ED 041 -979 MF $0.65 HC$6.58).

Cohen, Malcolm S. "The Direct Effects of Federal Manpower Programs inReducing UneMployment," Journal of Human Resources, 4:491-507. Fall,

.1969. .

Commins, William EV. Sociel Secwity Dan: An Aid to. MaripoWer ProgramEvaluation Washington, D.C.: Pripareirforthe Department . of Labor,Planning Research Corporatirm, November, 1970.

,Comptroller General's Report. to ihe Congress, Review of the Commwsity

Action Program in the Los Angeles Area Under the Econornk OpportunityAct. Washington, D.C.: Office of Economic Opportunity, B-162865, '

March 11, 1968. 32 .

Cornehen, L.A. "EConomics of Training the Unemployed," School Life.47: 17-18. October, 1964. Condensed in Monthly Labor Review, 88:672.

. June, 1965. .

Ehresman, Norman b.; Evanson, Lyle S.; andfischa4 Frank W. Analysis of theEffectiveness of the MDTA. Institutional Programs .in North Dakota,1962-66. Research Re)sylNdo: 12, Grand Forks: Center' for Research in/Vocational and l'echnic ucation, University of North Dakota, 1968.

.Flores, Froilan. An Historical and Cott Anelyes of Manpower Development endPaining Act Programs in the Washoe County (Reno) School .District,Unpublished 'doctoral disiertation, Salt Lake City, Utah: University ofUtah, 1968. 203 pp. (ED 030 069,Document not available from EDRS).

Gibbard, Hard() A. and Somers, Gerald G. "GovernMent .Retraining ofthe Unemployed in West Virginia,"' in G.G. Somers; (ed.), Retrgining theUnemployed, Madison, Wisconsin: University of Wisconsin Press, .1968. .

350 pp. (ED-021,198, Iktument not available from EDRS). 69

Goldfarb, Robert S.- "The Evaluation of GrivernmeniPrograms: 'The Case ofNew Haven's Manpower Training Activities," Yak Economic Essays. 9 .

:(2): 59-104; Fall, 1969.

Gough, LOwell, A., and Rowe, Harol. d R. A Shady of Anton Anodsted withOutcomes Of MDTA Aricultwal Education Pmjects in the Somerset Arm'

. Final Report. Lexington, Kentucky: Kentucky Research! CoordinatingUnit for Vocational Education, January, 1968.. 25 pp. (ED 021.130 MF$0.65 HC $3.29).

. ..

Greenberg, D.H. Employe's end Manpower TrainingliSgrams: Din Collectionend Analysis. Santa Monica, California: Rand Corporation. October, 1968.

Greerdeigh Associates, inc. The Job Opportweties in the.. Business SectorProgram: An Evaluation of Impact Ten . Stenderd bktropolitenStatistical Areas. Washington, D.C.: Greenleigh Associates, Jule, 1970::

Gurin, hiner Oty Neyo Youth en a Job Trebling Project: A Study ofFicton Rekted to Attrition end lob Success. Final Report. Anh Arbor:

University. of Michigan. Institute for Social Research, December, 1968.155 pp. (ED 026 530 MF $0.65 HC 66.58).

A National Attitude Stud); of Trainees in MDTA Institutional Programs.Final Report to US: Deportment of tabor Manpower Administration andU.S. Department of Health, Education, and Welfere. Contract No.OS,64-47. Ann Aibor: 'University of Michigan, Survey Research Center,Instittlte for Social- Rematch. August, 1910.1268 pp. (ED . qss 893,Document.ftot available from EDRS). 1

Hamburger, Martin, and Locucio, Ralph. A Study of ManPoswr Developmentand Diking' Act Program in New .York State. New Yotk, New York:New York University. Center for Field . Research . and School Services,November, 1969. 195 pp..(ED 038 525 MF $0.65 HC $6.58).

Hardin, Einar. "Benefit-Cost Analyses of Occupational Tiaining Programa: AComparison of Recent Pudiea," in G.G. Somers; ald WD. Wood, (eds.),Cost-Benefit Analysis of. Manpower Policies; Proceedings of NorthAmerican ConferenCe. May 14-15, 1969. pp. 97-118.*

and Bonn, Michael E. &anomie Benefits end Colts of ReasinkgCourses in Mkhigen. East Laniing, Michigan; Michigan State University,:.Decembei, .1969. 437 pp. (ED 043 808, Document not available fromEDRS). 39, 06, 91, 102 .,

Harris, Liwir and Associates; 'A Study of job Corp No ShOws: AcceptedApplicants Vapid Not Go to a Drining Center; March, 1967.

A Study. of theltatus of August, 1966 job Corp Terminies-12 MarthaAfter Termination. October, 1967. (Revision).

a

t

,

Harrison; Bennet. Eduartinn, Trainerg' and. the Urban' Ghetto: Unpublisheddoctoral disseitation. Philadelphia, Pennsylvania: University ofPennsylvania, 1970;

Hoffnwn, N. March. Follow-up Study of MDTA E and D Proiect at BluefieldState Colkge. Waohingion, D.C.: Bute* of Social Science Research, Inc.,May, 1967.

Hoos, Ide R. (b" Retriining in the United States: Problemi -and Progress,"International labor Review pp: 414426. Naernbet, 1965.

. Retraining tire Work Force. Berkeley and Los Anples, California:University of California Press, 1967.

Johnson, Lo. une A. Followup of MDTA E. and D. Project Conducted by.7lakegee 'Institute. Washington, D.C.: Bureau Of So.;ial Science Rematch,Inc., May, 1967.

Follow;llp Study of MDTA E and D hoject Conducted byAgricultWaland Industrial State University at Nashvilk Wuhington,D.C.: Bureau ofSocial Science Researck, Inc., July, 1967.

Johnson, Paul R., and Robbini, Richard D.An Evaluation- of the Labor MobliityDemonstration Protect of the North Cavl* Fund. Raleigh, NorthCarolina: North Carolina StattpUniversity, 1967.

-

. .

Jones,' James A. Operation Retrieval. Chapter VIII. Research in Experimentaland Demonstration Programs for Disadvantaged Youth. U.S. Departmentof Labor, Manpower Administration. Washington, DC.: 1966. 36 pp. (ED

. 014 607 f4IF $0.65 liC $3.29).

Kaufman, jacob J. Research, Development, and Demvsstration in AdultIMining and Retraining. Final Report. University Park: Pennsylvania State .

University, September, 1966,

Kreps', Juanita M. "Manpower and Welfare Programs: Benefit-Cost Analysis:DiscussiOn," Industrial Relations Research Association, Proceedings of thel7thAritsral Meeting, Chicago, Illinois. December 28-29; 1964.

Langdon, Charles W.-ManpoWer Development Training in Alabama. Unpublisheddoctoral dissertation, University, Alabama: University of Alabama.

Leasco Systems and Research Corporation. Quantitative Analysis of theConcentrated Employment Program, Vol. II, Technical ReportMSG-101/69, Silver Spring, Maryland: Waco. Systems' and ResearchCorporation, August 13, 1969.

Levine, Abraham S. ."Job Training Programs for the Disadvantaged; how canthey become effective?" Welfare in RevieW, 8:1-7. January-February,1970.

Levine, Robert A. "Manpower Programs in the War on Poverty,"- in CZ. Somersand W.D. Wood, (eds.), Cost Benefit Analysis of Manpower Policies.Proceedings of a North American Conference, Kingston, Ontario: Queen'sUniversity 'at Kinpton, Industrial Relations Center, 1969. 188' pp. (ED039 334, Document not available from EDRS).

.

Levitan, Sar A. "An Antipoverty Experiment: The Job Corps," Industrial .

Rektions Research Association. Proceedinis. Winter, 1966. .

Antipoverty WOrh and Training Efforts: Gosh and Reality. Policy Papersin Human ResourcM and Industrial. Rektions. No. 3. Ann Arbor,-Michigan: University. of Michigan, Institute of Labor and IndustrialRelations,,1967.

. "Manpower Aspects of the Economic Opportunity Act," IndustrialRelations Research Association- Proceedings, 1967. Winter. pp. 172-181.AlsO in US: Congress .kint Economic Committee. Fedend Programs farthe 'Development of (Human Resources; A Compendium of paperssubrnitted to the Subcommittee on Economic Progress. Washington, D.C.:Govemnsent Printing Office, 1968. Vol. 1, pp. 195-201. .

&am*, Programs for the Disadrantsgac Berkeley, California:*University of Califomia. June 21, 1966. 18 N. Mimeographed.

, and ManguM, Garth L. Federal Train* and Work Progmnu in theSixties. Ann Arbor, MiChigai: University of Michigan,- Iniiitute of Laborand Industrial Relitions, 1969..464 pp. (ED 034 860, Document notavailable from EDRS); ' 4 '

t.}

132

London, H. H. How Fare MDTA Ex-Trainees? An E4hteen Months' Follow-upStudy of Five Hundred Such Persons. Columbia, Missouri: University ofMissouri, December, 1967. 219 pp. (ED 022 054, Document not availablefrom EDRS).

Main, Earl D. "A Nationwide Evaluation of MDTA Institutional Job Training,"Journal of Human Resowces. 3:159-170. Spring, 1968. 29, 62, 69

. A Nationwide Emeluation of MDTA Institutional Training. Report No.118, Chicago, Illinois: National Opinion Research Center, University ofChicago, October, 1966.

Mangum, Garth L. Contributions end Costs of Manpower, Development endTraining Polky PaPerS hi HuMin Resnurrez and bodyeedisi5. Ann Arbor, Michigan: University of Michipn; Institute of Labof andIndustrial Relations, 1967. 85 pp. (ED 021 949 MF $0.65 RC $3.29).

"Manpower Programs in the Anti-Poverty Effort," ExaMination of the'War on Poverty. Staff and Comultant ..Reports. Prepared for theSubcommittee on Employment Manpower and Poverty. U.S. Senate. 90thCongress, 1st Session, August, 1967.

Foundation of Federal Manpower Policy. Baltimore, Maryland:opkhis Press, 1968. .

"Second Chance in the TrasisitiOn from School. to Work,"power Symposium. th Princeton University, 1968. The

m School to Work; a report. 1968. pp. 231-269. *-

The John

ngxv.1±, Garth

ce ion Mtion

Miles, G H. Optimising the Benefits of Nekiebontood Youth Corps Protects forR Youth: ?hoe I, Final Report. Minneapolis, Minnesota:- North Star

search and Development Institute, 1968.

S. y 'of R - - Literature Relevant to Opti ing the Benefits ofNe . Youth Proleas for R Youth. Minneapolis,Minnesota: North Star ch and Developmen Institute, 1968. ,

Mincer, ,Jacob. "On the Job Training. s and Some Implications,"Journal of Politkel Economy. Supplement. 70:50-79. October, 1962.

Mishnun, Florence, and Delehanty, John. Treinbrit hi Serrice Occupations Underthe Manpower Development end Doting Act. Washington, D.C.: U.S.Department of Labor, Manpower Administration. Office of ManpowerPolicy, Evaluation ind Research. Manpower Research Bulletin, No. p,1966. 19 pp. (ED 012 875 MF $0.65 IIC $3.29).

Mobile Area Committee for Training and Development, Inc:Project MACTAD.Final Repon. Mobile, Alabama: Mobile Area Committee for Triining andDevelopment, Inc. 1969. ..

Muir, Allan H., et al. Cost Effectivenns Analysis of On-thellob and InstitutionalCourses, Washington, D.C.: Planning ResearckCorporation, June, 1967.62

0

Neighborhood youth Corps. Evaluation of Neighborhood Youth Corps Projects.Abstract. Darien, Connecticut: Dunlap and Associates, Inc. February,1966. 10 pp. (ED 015 286 MF $0.65 HC $3.29).

Nosow, Sigmund. Refraining Under the Manpower Development Training Act: AStudy of Attributes of Nimes Associated with Successful Refrdning,Lansing, Michigan: School of Labor and Industrial Relations, Michitan,Michigan State University,-January, 1968.

Operations Research Inc. CEP Evahation MethodOlogy, Phase I Report,Technical Report 617, Draft. Silver Spring, Maryland: OperationsResearch, Inc. July 13, 1970.

c ,Montgomery and AZthur Smithies, (eds.), Pub& Policy, Volume 13,Cambridge, Massachusetts: Harvard University Press, 1964,

A-

Fume, Herbert S. "Disciusion," in G.G. Somers and W.D. Wood, (eds.),_Cost-Benellt Analysis of Manpower Policies. Pioceedings of a NorthAmerican Conference, Kingston,. Ontario: Queen'e University, LidustrialRelations Center, 1969. 188 pp. (ED 039 334, Document not availablefrom EDRS). . .

Pejovich, Svetozar, and Sullivan,. WilliaM. The Role Of Technical Schoolt inImproving the Skills and Earning Cep** of Rum, Manpower: A Case

.. Study. Washington, D.C.: Office of Manpower Policy, 'Evaluation andResearch, U.S. Department .g3f Labor, September, 1966, 24 pp. 'FinalReport. (ED 015 329 MF $0. 5 HC $3.29).

Planning Research Corporatioelwit:Effictiveness Analysis of On-tIsefob andInstitutio Trnal ek* Washington, D.C.: Office of ManpowerPolicy, Evaluation and Researili. 1967.

Public Policy, 1964. Cambridge, Massachusetts: Harvud University Preu, 1964.

Rasmussen, Dale Bruce. Detemsinants of Rates of Reruns to Investment inOn-theJob Training. Unpublished doctoral dissertation, Dallas, Texas:Southern Methodist University, 1969. 159 pp. (ED 036 698, Document.not available from EDRS). .

Resource Management Corporation. Evaluations of the War on Poverty. Preparedfor General Accounting Office under Contract No. GA-65C. Bethesda,Maryland: Resource Management Corporation. 63

Retraining the Unemployed, Part I. "The New England Experience," NewEngland Busbies, ReWew, Federal Ram Bank, 1962. Part 2. "Interest inTraining," New England Business Review, Federal Reserve Bank ofBoston, September, 1962. Part 3. "RetrainingA Good Iniestment:' inNew Englond Divines: Review, April, 1963.

aReynauld, Andre. "D1scuuion,7 in G. G. Somers ind W. Donald Wood; (eds.),

. _Cost-Benefit Analysis of Manpower Policies. Proceedkr ,of tt North

. American Conference. Kingston, Ontario: Queen's University at Kingston,

1.34

Industrial RelationsSenter, 1969.-188 pp. (Ep, 039 334, Document notavailable from EDR'S).

Ribich, Thomas: 'Education, and Poverty. Washington, D.C.: The BrookinpInstitution, 1967; 268 pp. (ED 028 194, Document not available fromEDRS).

Robin, Gerald D. An Aisemment of the In-Public 'School Neighborhood YouthCorps Projects in Cincinnati and Detroit, with Special Reference toSummer-only old Year-round Enrollee& Fins I Report and IntSchedule SuppleMent. Philadelphia, Pennsylvania: National Analysis, 1February, 1969. 38

.

Roomkin, Myron. An Evolvation of Adult Bask Education under the ManpowerDevelopment and Trwining Act in' Milwaukee, Macomb:. Unpublisheddoctoral dissertation. Madison, Wisconsin: University of Wisconsin,. 1970.

. .

Schleck, Robert W. Antipoverty' Prograns Under the Economic. OpportunityAct. New York, New York: Tax Foundation, 1968. .

Schultz, George P., and Weber, Arnold' R. Strategies for the Displaced Worker:Confronting Economic Cliange. New York: Harper and Row, 1966.

Scott, Loren C. "The EconomiC Effectiveness of On-the-Job Training: TheExperience of the Bureau of Indian Affairs in Oklahoma,"Industrial andLabor Relations Review 23:220-36, No. 2. January, 1970. ..

Sewell,- David 0. "Discussion," in G.G. Somers and W.D. Wood, (eds.),Cost-Benefit Analysis of Manpower Policies. .Proceedings of a NorthAmerican conference. Kingston, Ontario: Queen's.University at Kingston,Industrial Relations Center; 1969. 188 pp. (ED 039 334, Document notavailable from EDP). 29, !!

Silverman, Leslie J. Follow-UP Study of MDTA E end D Project at NorthCarolina A and T State Colkge. Washington, D.C.: Bureau of ScicialScience Research, Inc., March, 1967. .

Folk;w.up Study of Project Upiijk, the MDTA E and D ProjectConducted by Florida A & M Unhirsity. Washington, D.C.: Bureau ofSocial Science Research, Inc., July; 1967. ,74 pp. (ED 013 967 ME $0.65HC $3.29).

Smith, AD. "Active Manpower and:Redundancy Policies: Their ,Costs andBenefits," International Labour Review. pp. 49.60. January-FebrUary,1967; .

Smith, Ralph E. "Foregone Earnings During Manpower Retraining," WorkIngPaper 350-11, TheUrhan latitute, January 28,;1970.

- r,

Software Systems, Inc. A Job Corps Study of selanve Cost Benefits: Volumes Iand IL Washington, D.C.: Software Systeins, In'O. April, 1969.:

Solie, Richard J. "An Evaluation of the Effects 4. Retraining.in Tennessee:"Retraining the Unemployed, Gerard G. Somarsf(cd.), MadisOn, Wisconsin:,University of Wisconsin Press; 1968. 0 , t. ,

it

"Employment Effects of Retraining the Unemployed," Industrial andLabor Relations Review, 21 (2):210255, January, 1968. 69

Somers, Gerald G. "Economic Implications of United States Training Programs."Madison, Wisconsin:. Department of Economics, University of Wisconsin.August 15, 1968. Mimeographed.

.

"Evaluation of Manpower Development Programs," in Edward B.Jakabauskas and C. Phillip Baumel, (eds.), Human Resources. Ames, Iowa:Iowa State University Press, 1967. pp. 143-152.

"Evaluation of Work Experience and Training of Older Workers,"Unpublished report to the National Council on the Aging, Madison,-Wisconsin: University of Wisconsin, Industrial: RelatiOns ResearchInstitu te, 1967.

. ..

"Government SUbsidized On-the;Job Trainirt," U.S. &mite, Committeeon Labor and Public Welfare, Subcommittee on Employment endManpower, Hearings,' September, 1965, Washington; D.C.: GovernmentPrinting Office, 1965. 12 pp.

. "Our Experience with Retraining and Relocation_,_"in RObert A. Gordon,(ed.), Totorrd a Manpower Policy. New York, New York: John Wiley andSons, Inc., 1967. pp. 215-248. Complete Twit 31 pp. (ED 023 944. MF$0.65 HC $3.29).

"Retraining: An Evaluation of Gains and Costa," in Arthur M. Ross,(ed.), Employment Policy ond the Labor Market, Berkeley, California:University of California Press, 1965. pp. 271-98.

. "Training the Unemployed," in Joseph M. Becker, (ed), In Aid of the.Unemployed, Baltimore, Maryland: Johni Hopkins Press, 1965. pp.227-251. -0. The Training and Placement 4 Older Workers: an Evaluation of FourCommunity Projects. Prepared for the National Council .on the Aging andthe Office of Manpower Policy, Evaluation and Research, U.S. Departmentof Labor. Madison, Wisconsin: Unirisity of Wisconsin, 1967.

.. "Vocational Retraining of the Unemployed: Art Evaluation," IndustrialRelations Research Institute Report. Madisik, Wisconsin: IndustrialRelations Research Institute, Center for Stdies in Vocational andTechnical Education, University of Wisconsin, April, 1966. pp. 6-8.

, and McKechnie, Graeme H. "Vocational Retraining Programs for theUnemployed," Proceedings of the Industrial _Relations ResearchAssociation. 20:25-35. December, 1967.

, and Stromsdorfer, Ernst W. "A Benefit-Cost Analysis cf ManpowerRetraining," Proceedings of the Industrial Relations Reseirch Association.17:172485. Winter, 1964. 40,

A Cost-Effectiveness Study of the In-School and Summer NeighborhoodYouth Corps. Madison, Wisconsin: Industrial Relations Research Institute,

or

Center for Studies in Vocational and Technical Education. The Universityof Wikonsin, 1970.

. "A Cost-Effecti eness Analysis of the In-School and SummerNeighborhood Youth orps: A Nationwide Evaluation," in U.S. Congress,Joint Economic Conmlttee,L Compendium on Cost4ienefit Analysis;forthcoming.

Seminar Progress Rert on the Study:A Cost-Effectiveness Analysis ofthe In-School and Summer Neigh od Youth 'Corps, Publi6 PolicySeminar.. Series, The Utban Institute, U.1: Working Paper No. 350-22,

_ Wasliington, D.C., November; ,1970.

, a41 Wood, W. D., (eds.). Cost-Benefit Analysis of Manpower Policies.-: Proceedings of a Nonh American Conference. Kingston, Ontario:

Industrial Relations Center, Queen's University at Kingston; Madison,'Wisconsin: University of Wisconsin, Center for Studies in Vocational andTechnical Education, 1969. 35

owes, Carl E. An Instrument 4 Evaluate MDTA Institutional Trlining Protectsin Arizona. Phoenix, Ariz a: Arizona State Department of VocationalEducation, 1968. 39 pp. (E 025 638 MF $0.65 HC $3.29).

Stern, James L. "Consequences f Plant Closure,"Journal of Iiwnan Resources.Winter; 1972. Forthcoming. 62

Stevenson, Diantha. Eollow-up Ssdy of MDTA E and D Projects Conducted by it.Morgan State College, WaJiington, D.C.: Bureau of Social ScienceResearch, Inc., May, 1967. 57 . (ED 015 287 MF $0.65 HC $3.29).

Stromsdorfer, Ernst W. "Determinaits of Economic Success in Retraining theUnempj.ked," The Journal c'f Human Resources,' 3:139-158. Spring,1968. ,69

"Dismission," in C.C. Somers and W.D. Wood, (eds.), Cost-BenefitAnalysis\ o,f Manpower Polk-lei Proceedings of a North AmericanConferince,; Kingston, Ontario: Queen's University at Kingston,

. Industrial Relations Center; 1969. ."Probleths' in Estimating Social Costs and Benefits' in Education: The

Case oT the In-School Neighborhooa Youth Corps," InternationalFederatkin 'of Operational Research Socieiies International Cost-Effectiveness Conference. Proceedings. Washington, D.C. April 12-15, 1971.

. Forthcoming.Systems Development Corporation. Evaluation of 'the JOBS Program by Nine

Cities, final Report. Falls Church, Virginia:. Systeriu DevelopmentCorporation, September, 1969. ;

The North Carolina Fund. Manpower Improvement through Community Effort.Final Report Durham, North Carolina: U.S. Department Of LaborContract. No.,82-32-30.-1967.

Therkeldsen, Paul; ind Rno, "Cost-Benefit Evaluation of the BernalilloCounty, NeW Mexico. Work Experience Project," Welfare in Review. .

: March/April; 61-12,.1968.

145137

Thomas, Brinley; Moxham, John; and Jones, J.A.G: "A 'Cost-Benefit Analysis ofIndustrial Training," British Journal of Industrial Relations, 7:231,164.No. 2auly, 1969. 1

Troboff, Benjamin Michael. Employment Experience Aftr MDTA`Training: A

. o Study. of the Relationships between Selected 7)Wnbl Characteristics andPost-training Employment Experience. Unpubl14%e1 doctoral dissertation.Atlanta, Georgia: Georgia State College, 1968.

U.S. Congress. Review Of EconoMic Opportunity Progrenrs. Comptroller General\ of the United States. Made Pursuant tolitle H of the 1967 Amendments

to the Economic Oppoitunity 'Act of 1964. Report to the Congress of theU.S., 91st Congress, 1st Session; Washington, D.C.: GoVernment PrintingOffice, 1969.

, .

Senate, Committee on Labor and PubliC Welfare. SUbcommittee onEmploymeut Manpower and Poverty. "Manpower PrOjfams in theAntipoverty Effort," by Garth L. MangUm. Examination of the War onPoverty. pp. 235, et. seq. Staff and Conaultants Reports, Vol. 11, 90thCongress, ist.Session. 1967.

U.S. Department of Health, Education, and Welfare. Assistant Secretary forProgxam Coordination, .Office of the U.S. Department of Health,Education and Welfare, Human Investment Programs; Adult BasicEducation, Work Experience, rill Paining. Washington, D.C.: 1966.

U.S. Depariment of Labor, Ateepower Report of the President, Washington,D.C.:.Government Printing Office, Maich, 1970.

. Manposeert Research, and :Doling Report of the Secretory of Labor,Washington, D.C.: U.S. DepartMent Of Labor. Mut, 1965.

The Influence of MDTA Delning on &Wisp.- Wishington, D.C.:Manpower Administration, 1968, 40 pp. (ED 026 542 MF $0.65. HC$3.29) . .

U.S. Department of -Labor, Manpower Administration. 'The NeighborhoodYouth, Corps. A :Review of Reeearch," Mimpower Rama Monograph,No. 13. Washington, D.C.: Government Printing Office, .1970. 60 pp. (ED040 279 MF $0.65 HC not available from EDRS).

.

Office of Manpower, Automation,.and Training. Measuring. the Benefitsaryl Costs of Retraining norms for Unemployed Workers. East Linking,Michigan: Michigan State University, n.d. .. , Report of the Secretary of Labor on Research and Paining byAccordance ts4th &aim 309 of the Manpower Development and TrainingAct, Trottnyitted to Congress, March, 2964.

U .S... Office of Edncation...Eiehrelion end Benefit-Cost Relationthip of .

Manpower.. Training; Propane in New. York State. New. YOrk State..:Department of Labor, October,- 1967.

U.S. Senate, Cenimittee on Labor and Public Welfare, Subcommittee onEmployment, Manpoiver and Poverty. The JOBS Program, BackgroundInfonnation, 91st Congress, 2nd Sersion, April, 1970.

Virginia State College. Factors in Workers' Decisions to Forego Retraining under,the Manpower Development and Maiming Act. Norfolk, Virginia: U.S.Department of Labor, 1964.

Walther, Regis H. Methodological Considerations in Evaluative ResearchInvolving Disadvantaged Populations; a study of the effectiveness ofselected out-of-school Neighborhood Youth Corps Programs. Washington,D.C.: George Washington Univereity, Social Reattach Group, 1968. 34 pp.(ED 028 223 MF S0.65 HC $3.29).

;14

,

. A Study of the Effectiveness of Selected Out-of-School NeighborhoodYouth Corps Programs: Implications for hogram Operations andResearch. Washington, D.C.: Social Research Group. The GeorgeWashington University, Februaty, 1969. 23 pp.- (ED 028 222 MF $0.65HC $3.29).

et al A Retrospective Study of the Effectiiersess of Out-of-SchailNeighborhood Youth Corps Prograntsbi Four Urban Sites. Washington,D.C.: George Washington -University, November, 1963. 157 pp. (ED 020407 MF $0.65 liC $658). ,

Walther, Regis H., and Mignusson, Margaret L. A Study of the EffectivenessSelected Out-of-School Neighborhood Youth CornMetlsodOlogjcal Considerations in Evaluative Research InvolvingDisadvantage d Popirletion. Wuhington, D.C.:. Social 'Science ResearchGroup, cooing Washington University, May, 1968.

Young, Stanley. Manpower Training: Some -Cost Dimensions. Amherst,Massachusetts: Labor. Relations and Research Center, University ofMassachusetts, 1968..51 pp. (ED 027 390 MF $0.65 HC $3.29).

O,`;'

IV. GENERAL WORKS. . .

Advisory Council on Vocational Edulation.."Vocational Education": The Bridge. Between Man and His Work," 'Publication 1.Nores1and Working Papers

Conceming the Administration of Programs Authorised under VocationalEducation Act of 1963. Ptak' Law 88410; as Amended Subcommitteeon Education, Committee on Labor and Nblic Welfare, U.S. Senate, 90thCongress, 2nd' Sesiion Nov., 1968. 248 pp. (ED 028 267 M $0.65 HCnot available from EDRS). . I ;

"Vocational Education: The :Bridge Between- Map and His Work,'Publication 2. Notes 'and Working Apes Concerning the Administrationof Programs Authorised Under Vocational Education Act of 1963. Publk,Law 88.210, as Amendid. SubcOmmittee on Education, Committee onLabor and Public Welfare, U.S. Senate, 90th Congress, 2nd Session, Nov., .

1968. 248 pp. (ED 028 267 MF $0.65 HC not available from EDRS).

&doe', Melvin. 'Vocational Edsication. Chicago, Illinois: National Society forthe StUdy of Education, 1965, . .

II

,

.

Becker, Gary S. Human Capital, New York': Columbia University Press, 1964.

Brindon, George L., (ed.), "Research Visibility. Research in Vocational and.Technical Education," American Vocational Journal, .44:41-51 14o. 2.February, .1969,

Correa,'Hector. The Econcrics of Hamm Resources. Amsterihm, Holland: TheNorth Holland Publishing Company., 1967. ..

Evaluation in Ve;cational Education Research Summary. SacrameotO,California:Coordinating Unit for Occupational Research and Davelopthint. 1967. 81.pp. (ED 025 587 MF $0.65 IIC $3.29).

Ginsburg,.WoodroW; Robie, Edward A.; and Taylor, David P. "Manpower andJob Education and Training': Discussion," The Development and Use ofManpower Proceedings of tfte Aventieth Annual 'Pinter Meeting,Washington, D.C., December 28-29, i967. pp. 47-59. Madison, Wisconsin:Industrial Relations Research Association, 1967.

Griggs, Charles M., et al. Vocational Rehabilitation for the' Maim ihm;illaneconomic and sociological evaluation. Lexington, Massachusetts: HeathLexington Books, 1970. 275 pp.

Lampman, R.J. "Towards an Economics of Health, Education and Welfare,"Journal of Human Resource& 1:45-54. Summer, 1966.

Meade, ward I., Jr., and Feldman, Marvin I. "Vocational Education: Its Placeand Process," The Jownal of Human ILTurces. 1:70-74. Summer,1966.

Monroe, Walter Scott, (0.). Encyclopedia of Educational Researck New York:The Macmillan Company, 1950.

Puce!, D.J., and Evans, R.N. "Manpower Economics and Vocational EdUcationl'Ametican Vocational Journal, 39:18-19, April, 1964.

.

Robinson, E.A.G.; and Vaizey, J.E.; (eds.). The Economics of 'Education.'Proceedings of a Conference held, by the Internelonal EconomiCAssociation, New Yorki.New,York: St. Martin's Press, 1966.

Russell, Vocational 'Education. President's AdviscnY Committee onEducation, Staff Study No: 8, Washington, D.C.: Government Printing ,

Office, 1938. . .

Schultz, T.W. "Investing in tOor People:. An Ecoltmist'i View" American.*Economic Review 55:510-529. No, 2. Mai,. 5:

Somers, Gerald G., and Little, J. Kenneth. Vocational Education: Todiy andTOMOROPI. Madison, Wisconsin: Center for Studies in Voca3ional andTechnical.Education, University Of Wisconsin, .1971.. -

Sundquist, James L "Jobs, Training, and Welfaie for the Under' Cliu,'4 in ,

Kermit Gordon, (ed), Agenda for the Nation: Pepin OA Domenic andForeign Policy Issues. Washington, D.C.: The Brookings Institution, 1968.pp. 49-76. .

140 ''.

! 0

Tweeten, Luther G. The .ilde *of Education tel A ling Ram; Poverty.. Agricultural EconomiC Report No. 114. Washington, D.C.: Economic

Research ServiCe, U.S. Department of Agriculture, June, 1967. 60 Pp. (ED. 020 039 MF $0.65 HC $3.29). . - '

U.S. Department of Health, Education,and Welfare. Education for a ChangingWorld of Work; Report of the Panel of Consultants in Vocational .

Education. Washington, D.C.:- Govethment Printing Office,1964. 307 p0.. (ED 019 500 MF $0.65 HC 113.16). . .

. .

Vaizey, J. The Present State .of the Economics of Edmcation. London, England::..Faber.and Faber, Inc., 1963. ..

Venn, -Grant, and Marchese, Theodore, J. Jr. Akn, Education,' and[Work.Washington, D.C.: Amerkan Council on Education,'1964.

William, Walter. OE0 Illanpow e randkanpoly e r Related Research Was1Lington,D.C.: Office of Economic Opportunity, U.S. Department of Labor, 19671

V. (B;;;LIOGRAPHIES

Abstracts of 'Research Materials in. Vocational and .Teeknical. Education.Colimbus, Ohio: The Ohio State University, The Center for VocationaEducation. 1967-1972. . .

An Annotated Bibliography of Benefits and Coits in the rub& Sector.Philadelphia, Pennsylvania:, Research for Bitter Schools, Inc., 1968. 254pp. (El) 026 744 MF $0.65 HC $9.87),,

Blaug; Mark. EconomicsOfEduortion: A Selected AnnotatedBibliograplsy. NewYork, New York: Pergamon Pfeil, 1966. .

. A Selected' Annotated 'Bibliography in the Economics of Education:Education' Librariei Bulletin, Supplement, No. 8, London; England:

. 'Institute of Education, University of London, 1964. .

Case, C. Marston, and Clark, Stephen C. A BibliograpNcal Guide toOperations,Analytis of Education. Technical Note' Number 43. Washington, D.C.:National Centei for Educational Statistici, U.S. Office of Education.September. .

Corplan Association. A Bibliogra. phy of 'Published and Unptiblished Vocationalend Technical EducatiOnal Literature.. Chicage, Illinois: Illinois. Research .

and DevelopMent COordinating Unit, June, 1966: .,

Darcy,' Robert L. Economics of limn= Resowees: A Biblthgraitt,, 4ktheas,Ohio: Ohio University, Department of EcOnomks, January; 1968.Mimeographed.

Deitch, Kenneth M.; McLoone, Eugene P.; compilers. The Economics ofAmerican Education: A Bibliography, Including Sdected Ms/Or Referencesfor Other Nations. Blooinington, Indiana: Phi Delta Kappa, 1966. 69 pp.(ED 024 108 MF $0.65 $3.29).

0

9.

Edding, Friedrich. Bibliography of the Economics of Education. Paris, France:0.E.E.C., 1961. Mimeographed.

.Eidell, Terry L., .and John A., (compilers). Annotated BibiloirePhy onthe Evaluation or Educational Programs. Eugene, Oregon: OregonUniversity, 'Eugene, Oregon: ERIC Clearinghouse on EducationalAdministration, November, 1968. 19 pp. (ED 625.. 857 MF $0,65 .HC$3.29).

Homer, Jamei T., (ed.). Summaries of Studies in Agricultural Education, CentralRegion, 1968-69. A n Annotated Bibliography of Studies in AgricultwalEducation. Lincoln, 'Nebraska: University of Nebraska, Department ofAgricultural Education; 1969. .

Little, J. Kesineth. Revie w and. Synthesis of Research on the PlaceMent andaf Vocational Edmonton Students. Columbus, Ohio: The Ohio

State University, The Center, for 'Vocational and Technicil Educition;FebruarY, 1970. 54 pp. (ED 037 543 MF $0.65 HC $3.29).. .

'McGivney, Joseph H., and 'Nelson, William C: Program; Banning, Budgeting.. Systems for Educators, Volume III: An Annotated Bibliography. Final.

Report. .Ribliography Series No.. 3. Coltimbus, Ohio: The Ohio State..University, The Center for Vocational and Technical Education, 1969. 57

. pp. (ED 035 756 MF $0.65 11C $3.29).

Microfiche Collection of Documents Reported in Abstracts ,of Research Ma-totals in Vacctional and Technical Education, Wintir, 1967. Columbus,

. Ohio: .The Ohio State University, The Centeofor Vocational and TechnicalEducation, 1967. '1 1,214 pp. (ED 015 349 MF $23.10 HC not availablefrom FDRS).. .

Nelson, William C.Progrnm; Planning, Budgeting Systemi jot Educators; VolumeIV. A Research Bibliography. ColuMbus,.OhiO: The Ohio State-University,The Center for Vocational and Technical Education, 1970. 1 17 pp. (ED038 512 MF $0.65 Hc $6.58). .

Poverty end Hump Resources, Abstracts. Univitiity of Michigan-Wayne StateUniversity: Industrial Relations/InstitUte of Labor.. Volumes 1-4.1966-1969.

, .

Rand Corporation. A Bibliography of Selected Rand Publications. Santa Monica,California: Rand Corporation, July,'1970. , .

Rearman, Elisabeth H. Bibliography, of Cost-Benejli Analysis and ManningProgramming-Budgeting. McLean, Virginia: Research AnalysisCorporation, February, 1966. . .

Sharp, Laure M., and Kramegor, Rebecca. The Use of Follow-Up Studies in the&vitiation of Vocational Education. Washington, D.C.: Bureau oi SocialScience Research, Inc. May, 1966. 42 pp. '(ED 010 072 MF W.65 HC$3.29). , . .

U.S. Department of Labor. Cost-Benefit Analysis: Theory old Application to .

Manpower Paining Programs: A Bibliography. Washington, D.C.: U.S.Department of Labor, May, 1971..

U.S. Office of Education. Technical end Vocational Education in the USA. ABibliographical Survey. Education Studies and Documents, No. 36, Paris,France: UNESCO, 1959.

University of Minnesota, Laboratory for Research in Social Relatiofis. AnAnnotated Bibliography of Research Related to Unemployment andRetraining. Washington, D.C.: U.S. 'Department of Labor, Office ofManpower, Automation and Training, 1965.

University of Wisconsin, Bibliography No. 12, Addendum No. 6: Cost BenefitAnalysis. Madison, Wiscomin: Center for Studies-Vocatioiral andTechnical EdUcation. January to April, 1968,

.11ibliography No. 16, Addendum No. 6: Manpower. Center for Studies inVocational and Technical Educati,Januaty-April, 1968.

Bibliography No. 18, Addenda No. 6: Retraining. "January-April, 1968,,Mimeographed.

.

cent, Howard L. Selected Bibliography.-.-Application of Economii Analysis`4n' -Operailons Research to Problems in Educational Plannint

Wa

Lieson, D.C.: U.S. Department of Health, Education, And Welfare,

' National nter for Educational Statistics. 1966. ,

VocatiOnal and Technical Education Reviews the Literature for the 6-Year-period since October, 1956: Review of Educational Research, October,1962. 32:363-433.

. ,

Warmbrod, L Robert. Review and Synthesis of Research on the Economics of tVocationd-Technical Education. Columbus, Ohio: The Center for \Vocational and Technical Education, The Ohio State University,

.1November, 1968. 62 pp. (ED 023 937 MF $0.65 HC $3.29).

Wood, W. Donald, and Campbell, H.F. Cost-Benefit Analysis and the Economicsof Investmeat in Human Resources: an Annotated Bibliography. Kingston,Ontario: Industrial Relations Center, Queen's University, 1970..

e.

*U.5.13ovornpiont Printing Offices 1972 755-7113/4 Rogion No. 5-11