.10

REPORT R ESUMESED 011 578 VT 003 659

OCCUPATIONAL JOB REQUIREMENTS, A SHORT -CUT APPROACH TO

LONG -RANGE FORECASTING.BY- MEDVIN, NORMANBUREAU OF EMPLOYMENT SECURITY, WASHINGTON, D.C.

PUB DATE 67

EDRS PRICE MF.40.25 HC -$0.72 16P.'

DESCRIPTORS.... EMPLOYMENT TRENDS, *OCCUPATIONAL SURVEYS,

RESEARCH METHODOLOGY, MEASUREMENT TECHNIQUES, *EMPLOYMENTTRENDS, EMPLOYMENT OPPORTUNITIES,

THE AREA SKILL SURVEY, THE BEST KNOWN AND LONGEST USEDTECHNIQUE FOR FORECASTING LONG -RANGE OCCUPATIONALREQUIREMENTS, HAS COME IN FOR SEVERE CRITICISM ON THE GROUNDSTHAT IT IS TOO EXPENSIVE AND TIMECONSUMING AND THAT MOSYEMPLOYERS ARE NOT GOOD ECONOMISTS. A NEW TECHNIQUE ISDESCRIBED, THE EMPLOYMENT SERVICE UNFILLED JOB OPENINGS,OCCUPATIONAL OUTLOOK HANDBOOK APPROACH. ITS ELEMENTS ARE THE

UNFILLED JOB OPENINGS IN A LOCAL EMPLOYMENT OFFICE,IDENTIFICATION OF THOSE JOBS OPEN 30 DAYS OR MORE, AND THEADDITION OF A NATIONAL FORECAST (OBTAINED FROM THEOCCUPATIONAL OUTLOOK HANDBOOK) TEMPERED BY THE LOCALANALYST'S KNOWLEDGE AND SUPPLEMENTED BY A FEW ASSOCIATIONVISITS. IT IS ESTIMATED THAT THE CONDUCT OF SUCH A SURVEY TOSATISFY VOCATIONAL EDUCATION, MANPOWER DEVELOPMENT ANDTRAINING ACT, AND OFFICE OF ECONOMIC OPPORTUNITY NEEDS 'WOULD

TAKE A SINGLE MANPOWER ECONOMIST AN AVERAGE OF NOT MORE THAN

6 WEEKS FOR A SURVEY IN A METROPOLITAN AREA. TOTAL COST TOTHE EMPLOYMENT SERVICE FOR 150 AREAS, ALL AREAS TO BECOMPLETED IN 1 YEAR, WOULD BE UP TO $200,000. SKILL SURVEYS,

NOW USED, MAY COST AS MUCH AS $100,000 FOR ONE LARGE CITY.

THIS ARTICLE IS A REPRINT FROM EMPLOYMENT SERVICE REVIEW,JANUARY FEBRUARY 1967. (PS)

4

CoU.S. DEPARTMENT OF HEALTH, EDUCATION & WELFARE

r-t THIS DOCUMENT HAS BEEN REPRODUCED EXACTLY AS RECEIVED FROM THE

CD PERSON OR ORGANIZATION ORIGINATING IT. POINTS OF VIEW OR OPINIONS

STATED DO NOT NECESSARILY REPRESENT OFFICIAI. OFFICE OF EDUCATION

OFFICE OF EDUCATION

Lti POSITION OR POLICY.

Occupational Job Requirements:

A Short-Cut Approach to Long-Range Forecasting

ceD

0P-1 U.S. DEPART

Reprinted fr

EMPLOYMENT SER

la

em:

VICE REVIEW

nuary-February 1961

MENT OF LABOR W. Willard Wirtz, Secretary

`LIMAEN116111.41...4:11.1a11153M:MR17.7-10--47.:X.

,{1

Occupational Job Requirements:

A Short-Cut Approach to LonglangrForecasting

NORMAN MEDVIN

The following technique of making long-range occupational forecasts is oneof several proposals under discussion in the US. Department of Labor.

he past few years have witnessed the growth of a livelyinterest among the general public, an increasing concernby government and private administrators, and what canbe only described as feverisha activity on the part of somespecialized economists and econometricians, toward theproblem of creating a viable technique for forecastinglong-range occupational job requirements.

Collection of occupational data and its corollary spin-off, the evolution of long-range forecasting techniques,have lagged behind other aspects of the job market infor-mation program for obvious reasons. The expense incollecting data by occupation, burdensome employer re-porting, the technical ability to translate employer jobtitles into a standard occupational nomenclature, anddifficulty in creating acceptable forecasting techniques aresome of the more important reasons why this area ofknowledge has not progressed as far as other job marketstatistical programs.

The spate of manpower legislation in the first half ofthe sixties, however, placed an urgent priority on theexpansion of occupational information. in providingfunds for manpower training, the Congress foresightedlywrote in the Area Redevelopment Act of 1961 that beforeany training courses could be approved; there had to bea "reasonable expectation of employment" for the trainees

Norman Medvin is a Manpother Economist, U.S. Em-ployment Service, Washington, D.C.

upon course completion.1 This all-important phraSe wasrepeated in the Manpower Development and, TrainingAct of 1962 and the concept was carried over into theVocational Education Act of 1963.3 No longer was itpossible to deal in platitudes or vague references to jobopportunities. If a class in welding was to be consideredin Milwaukee, then a precise determination of employ-ment opportunities had to be made in advance foF thatspecific occupation in that specific place. And if thecourse was to run for a period of time, then the determi-nation of employment prospects had to extend certainlyup to, and preferably beyond, the end of the trainingperiod. Enter the need for a forecasting technique.

Current Research in Forecasting

Considerable experimentation has occurred and is tak-ing place in developing acceptable techniques for long-range occupational forecasting. As might be expected,the U.S. Department of Labor through two of its corn-

Area Redevelopment Act, Public Law 87-27, May 1, 1961, .sec. 16(b).

Manpower Development and Training Act of 1962, PublicLaw 87-415, Mar. 15, 1962, title II, pt. A, sec. 202(d).

Vocational Education Act.of 1963, Public Law 88-210, Dec.18, 1963, sec. 5(a) (4).

1

,z1.44-4NiF.zi4zralt4110C Mffigliejfe

PHE

MANPOWER

RESOURCE

cwoc

.CLNISAMAS

OF THE PORTLAND

METROPOLITAN AREA

ponent elements - -the Manpower Administration (whichincludes the U.S. Employment Service) and the Bureauof Labor Statistics---is deeply involved.

The USES and its affiliated State Employment Serviceshave basically relied on the area kill survey techniqueand training needs surveys as a means of estimating futureoccupational requirements. These in turn rest primarilyon the validity of an employer forecast of occupationalemployment in his establishment 1, 2, or 5 years hencewhich is summarized for an area or an entire State.An aspect of this approach is the development oftechnical school graduates and promotions to the job, allof which are arrayed against employer demand for a netdemand figure by occupation.4 Some 170 of these skillsurveys have been completed over a period of 6 to 8 yearsand 44 are in process.

Both the USES and its affiliated Wisconsin State Em-ployment Service are currently experimenting on a modelsystem of gathering occupational information of whichforecasting is one of the areas to be explored. The USEShas also engaged a private consalting firm to devise anacceptable forecasting plan .8

The Employment Service has been active in otheravenues of inquiry relating to long-range forecasting.The Michigan State agency in cooperation with the

Handboo n Employment Security job Market ResearchMethodsArea Skill Survey, DES No E--252, November 1965.

"Outline of Suppplemcnt to the Handbook on EmploymentSecurity Job Market Research Methods, BES No. E-252,November 1965." A preliminary paper for discussion purposesonly, prepared by John Fletcher Wellemyer Associages for theUSES, Nov. 15, 1966.

2

akal...........111miorwroMM

Battelle Memorial Institute is developing a model fore-casting technique for the State of Michigan and forDetroit separately which attempts to measure supply aswell as dexnand.6 The State agency is also experimentingwith current occupational counts, frequently the base foremployment projections, through a conversion of thelistings in the R. C. Polk city directories,

The California State Employment Service experi-mented with an innovative approach in which an attemptvias made to marry the decennial census data with aneconomist's knowledge of local area manpower and

`economic conditions.?lie Manpower Administration, through funds avail-

able to it under title I of the Manpower Developmentand Training' Act, awarded individual contracts to theColorado and Temple Universities to experiment withlong-range occupational forecasting techniques, Both ofthese universities developed models of local occupationalemployment projections based primarily on the use ofdecennial census data.8 In neither case was there anyadjustment for available supply of manpower.

One of the more important techniques for projectingnational occupational levels is one developed by the Bu-reau of Labor Statistics which Eivolves preparation of aneconomic model by building up demand for each productin an industry, manpower requirements of each industryconsistent with this demand, and finally, the developmentof an occupational matrix. For a grid which has 116industries in the caption and some 150 occupations in thestub, BLS technicians have painstakingly over a 16-yearperiod filled in the cells so that today the have what istantamount to 116 industry manning tables for 1960, thatis, occupational distributions of employment for theseindustries. By applying their know-how of industrialand occupational trends, they have constructed a dupli-cate matrix for the year 1975 to be released in early 1967.By obtaining the difference from one pattern to the other,they in effect have a demand chart by industry and oc-cupation over the 15-year period. A third matrix for the

`Michigan Manpower Study, November 1966, MichiganEmployment Security Commission (research conducted by Bat-telle Memorial Institute, Columbus, Ohio).

Manpower Resources, San Francisco-Oakland Bay Area,1960-70, California Department of Employment, May 1963.

° Methodology for Projection of Occupational Trends in theDenver Standard Metropolitan Area by L. Fishman, W. E.Roberts, C. M. Franks, and W. McCormick of the University ofColorado is an evaluation of the state of the art of projectingoccupational demand for local manpower. The authors includea theoretically "ideal" methodology and data collection techniquewith each of their assessments of the precision and adequacy ofthe currently used occupational forecasting methods.

Projective Models of Employment by Industry and by Occupa-tion for Small Areas: A Case Study (Part I) and a Manual forthe Development of Estimates of Future ManpoAri;i Requirementsfor Training Purposes (Part II) by L. T. Harms, R. James, andR. C. Springer of Temple University presents a detailed descrip-tion of manpower forecasting models derived by the authors andan example of the models' application to manpower estimatesin one particular county. Pt. II of the report is a "hau-to-do-it"step by step description of the procedures involved in makinglocal area employment estimates.

4.4-444

lo see footnote

911.1,1,17 -..4-444,444.1.741.111.4114.4447411111.11111.1

Employment growth in mostoccupations will speed up

over next half decade.

Occupational groups

4

Professional andtechnical workers

Clerical workers

Service workers

Percent change in employment

-30 -20 -10

Managers, officials,and proprietors

Craftsmen

Sales workers

Operatives

Nonfarm laborers

0 10 20 30

Farmers andfarm workers

-30 -20 -10 0 10 20 30

V 1965-70 employment of nonfarm laborers is projected at about the

absolute level of the past decade.

Source: U.S.Departrnent of Labor.

41111011WMAMMtgtevAn.Z.VitriAVAnneaetac.igmlnigtoncznmeMMIVeza0

year 1965 is scheduled for completion in 1967. In eachof these models, there is a lack of manpower supply figuresfor occupations other than professional and some highlyskilled .°

Obviously, there are many efforts in the occupationalforecasting field which have not been mentioned. Somelike the Arizona Employment Service approach may beconsidered variants of the area skill survey approach.Others, like in the New York State Department of Labor,are an effort to apply the concepts of the BLS matrixapproach to a local situation. Some techniques like oc-cupational listings fror1i the Internal Revenue Servicequarterly contribution reports and a new technology ap-proach which assumes that a technologically advanced

For information on the matrix, set "Long-Term ManpowerProjections." Institute of Industrial Relations, University ofCalifornia, 1966, "Estimates of Employment Requirements byOccupation for Future PeriodsData Sources and ModelDevelopment," Harry Greenspan.

-1/Nbay.

plant foreshadows the average occupational distributionfor, the industry some years hence are in an embryonicstage or not yet sufficiently tested.1°

Major Difficulties Remain

A gap which is inherent in studies to date, with suchexceptions as the area skill survey, Michigan-Battelletechnique, and, to some extent the BLS model, is theabsence of labor supply figures. It is entirely possible, forexample, for employment in an occupation to increaseover a period of time but at the same time for the existingforces in the job market to supply those workers. Thus,one may project an enormous demand for file clerks butbecause any high school graduate could be part of thepotential supply, a pi.ospective shortage would be unlikely.It would be highly dubious to plan training courses on thebasis of demand alone. Perhaps more importantly, onemight have difficulty in using only a demand figure tosatisfy the "reasonable expectation, of employment" clausesince the funding of training courses and successful jobsearch is tied to a shortage concept.

Existing techniques suffer from other limitations.They are (1) extremely expensive (a skill survey for alarge city may run to $100,,000) ; (2) extraordinarilycomplex (the Temple University study, in our opinion,could hardly be applied at numerous local levels with ex-isting funds, skills, and data processing resources) ; (3) tootime consuming (skill surveys which should theoreticallytake 4 to 6 months to produce preliminary and usableresults most frequently take much longer. Some of theother studies in the experimental stage have not yet beenapplied but an objective judgment leads to the conclusionthat they, too, would take inordinately long) ; (4) notapplicable to a local :cene (the BLS matrix is an excellentnational planning document but is tied to local situa-tions without additional specific ac\tptation) ; and (5) notfeasible on a frequent basis becausli of cost, complexity,and time.

A note of urgency compounds the situation. The Voca-tiOnal Education Act of 1963, in al: orating hundrods ofmillions of dollars for vocational training, specificallygives a mandate to the Employrrpent service system to pro-vide job market information to the vocational schoolsystem. The State Employment Serves. no addi-tional funding and with the traditional and costly s I

survey technique available to them, have in many in-rstances been unable to furnish the information which bylaw they are obligated to provide.

There have been a number of instances in which theState vocational education people have attempted to getthe information themselves, or have financed private con-sultants or universities to obtain the information for them.It is clearly paradoxical that the Employment Servicewith the most know-how in the manpower field fre-

tnote 8 on the University of Colorado Study. Thisreport contains an extensive bibliography of manpower researchin the occupational forecasting field.

3

(

quently finds itself sitting on the sidelines while moneyis being distributed to others with far less experience todo the job. In one of the States, an outside consultantwho had been hired to make job market surveys for thelocal education authorities approached the USES inWashington on techniques to accomplish the task andthen offered to hire the responsible staff member to makethe surveys.

No additional budgeting is in prospect for the Em-ployment Service for skill surveys until fiscal year 1968at the earliest, and it is problematical that such fundswill be forthcoming then. In the meantime, the needsof the vocational education authorities grow more press-ing by the day. In a handful of States, the State voca-tional education group has cooperated with the StateEmployment Service in funding part of the costs of askill survey but this cost sharing, on the basis of presentevidence, will continue to be exceptional.

Need fear New Technique

There is little evidence, based on the availability ofcurrent techniques and known experimentation, that aviable method of long-range forecasting of occupationalrequirements on a local basis will be operational on anational basis before calendar year 1968. Excluded, ofcourse, is the skill survey approach which is operationalbut which is handicapped by lack of funding.. There canbe no argument that a new technique is needed as quicklyas possible to permit the Employment Service to dischargeits function of providing long-range occupational infor-mation to the vocational education system that would alsoserve the Employment Service itself in carrying out itsresponsibilities under the Manpower Development andTraining Act.

Two Myths

Before proceeding with the task at hand, it is necessaryto demolish two myths which frequently divert manytechnicians la achieving a modus operandi in the long-range occupalional forecasting field.

The first is tAe one which concerns "emerging occupa-tions." Many prefaces to research designs and preamblesto legislative proposals in the field of occupational jobmarket information have as one of their cardinal objec-tives the searching Out of new occupational opportunitiescreated by changing technology. Laymen repeat thephrase and give substance to the idea, conveying thenotion that to ferret out these elusive but all-importantdevelopments is to take a giant stride in harnessing theopportunities of tomorrow. There is a practitioner in theforecasting field who places major emphasis on this ele-ment, in the course of which he allocates a significantproportion of his total expenditures to interviewing.hundreds of employers for a series of qualitative judg-ments, many of which are akin to curbstone opinion andguesses. Even if these judgments were reasonable andeven if. it were possible to quantify them, it does not seemthat the emerging occupations could have more than aslight impact on the job market for years to come.

4

Major Occupational Groups ofWorkers, Actual 1964Employment and Projected 1975 Requirements

[Number in thousands]

.........Occupational group

-198,1

EmploymentProjected

1975Requirements

Per-cent

change1964-7e---,

Total, All occupationalgroups

White-collar workers

Professional and technicalManagers, officials. and pre- .

prietors.Clerical workersSales workers

Blue-collar workers.-- .......Craftsmen and foremenOperativesNonfarm laborers

Servics workersFarm workers

Num-ber

Per-cent

Num-ber

Per-cent

70, 357 100.0 88, 700 100, 0 26

'x1,125 44.2 42, 800 48.3 38

8, 550

7, 45210, 6674, 466

25, 534

12.2

10.815.26.3

36.3

13, 200

9, 20014, 6005,800

29, 900

14.9

10.418.58.5

33.7

54

2337ZO

17

8,98612, 9248, 6249.2564, 444

12.818.46.2

13.26.3

11, 40014, 8003, 700

12,6003, ii00

12.818.74.2

14.13.9

2715

(1)3521

1 Less than 3 percent.NOT*: Projections assume a 3percent level of unemployment in 1975.

Percents do not add to totals due to rounding.

Source: U.S. Bureau of Labor Statistics.

The fact of the matter is that the job structure changesrather slowly and that the overwhelming number of jobsand subsequent opportunities are in already existing oc-cupations. Outside of the unusual emergence of newindustries infrequently over a half century, it is a factthat most new industries call largely for existing skills.Moreover, emerging skills occur in very small numbersand when they do, it is most usual for the machine manu-facturer or the machine user to train the worker for thejob. By the time the number of workers in the occupationincreases substantially, it has undoubtedly been identifiedin the mainstream of job market information.

Evidence of the numerical importance of the older jobsin the economy is shown by the fact that since the in-ception of the MDTA program, institutional training in10 occupations, none of which were new to the postwareconomy, accounted for the bulk of the enrollments inthe program.11

The second myth concerns the drive for preciseness inforecasting projections. Models. stand or fall on howclose the projections are to the actual. Econometricianscompute changes to one decimal point. Evaluationstudies are conducted repeatedly for accuracy.12 Endless

"Statement of Stanley H. Ruttenberg, Assistant Secretary ofLabor and Manpower Administrator before the General Sub-committee on Education of the House Committee on Educationand Labor on Vocational Education Act of 1963, Aug. 76, 1966.

"An Appraisal of Area Skill Surveys in Battle Creek Mich.,and Trenton, NJ., November 1965, John Fletcher WellemyerAssociates, Washington, D.C.

5

arguments ensue on whether the economist or the em-ployer should make the forecasts on the assumption thatone is more accurate than the other.

Of course, any technique which produces a precise, orreasonably precise, long-range forecast is devoutly to bedesired. It is our contention, however, that this strainingfor exactitude in the field of counseling and training isspurious and lirgely unnecessary.

What does it mean, for example, when an estimatedneed in an occupation is 500, that the true need is 250 or750 (assuming it were possible to ascertain the figure),and that there are only physical facilities or funds for twoclasses of 25 each? Again, why the heavy emphasis onan elaborate econometric model when either before thestudy takes place or after it is. completed, the economistprograms a major correction factor based on judgment orother nonquantitative data? This was an actual occur-rence in a heavily financed study which tried to arrive ata technique at a point far removed from the area actuallyinvolved by using Census and other data. It was sub-sequently conceded that there should have been local con-tact to obtain "the intimate knowledge of an area whichonly residents have or which can be produced by visits."An obvious plea to temper statistics with mercifuljudgment.

We conclude with respect to this myth, and as it isapplicable to the field of job forecasting to satisfy train-ing requirements, that it is sufficiently accurate to achieveonly the Proper direction of employment change. In ad-dition to direction, it would be useful to have a magnitudeof need and this can be achieved without satisfying thepresumed need for preciseness. Moreover, if a tech-nique can be devised to repeat a survey quickly and in-expensively, say every 6 months or a year, how muchbetter would that approach be than one which is lockedinto an estimate arrived at by a survey which because ofits costs, can be repeated only at 3- to 5-year intervals.

Elements of a New Technique

The proposed methodology can perhaps be described asthe "Employment Service unfilled job openingsOccupa-tional Outlook nandbook" approach after its two basiccomponents. Two to three years from now, its namemight change to the "Job VacancyOccupationalMatrix" approach since these two components ,will bethe improved versions of the earlier elements.

Unfilled job openings are those live job orders given byemployers to the Employment Service which at the timeof count remain unfilled because of a shortage of ap-plicants or inability to find applicants who meet the em-ployer's job specification.13 Unfilled job openings are col-lected on a quarterly basis by the Employment Service in80 of the largest metropolitan areas in the country. Plans

This is not precisely true since some job orders mny havebeen given to the Employment Service so close to the time ofcount that there had not been sufficient opportunity to attemptto fill them.

are already underway to extend this collection to 150areas comprising some 70 percent of the Nation'semployment.

Job vacancies as used here are the universe of jobopenings in an area. , As such they differ from ES unfilledjob openings in that the latter represent that portion ofthe universe of job vaca.nies given to the EmploymentService.

The OCCUPATIONAL OUTLOOK HANDBOOK iS theBureau of Labor Statistics' outstanding contribution tothe training and guidance fields. Published biannually,the Handbook provides current and, long-range informa-tion on occupations concerning 90 percent of the 16 mil-lion employed in professional, managerial, and technical .occupations; nearly all of the 4.5 million sales workers;about half of the 10.7 million clerical workers; and about40 percent of the 9.3 million service workers.

The BLS occupational matrix ( see above for descrip-tion) will provide quantitative information on currentemployment and lcg-range forecasts of employment byoccupation. Most of the Lrecasts in the current Oc-CUPATIONAL OUTLOOK HANDBOOK are qualitative innature.

The final basic tool in the proposed technique is thenew and recently issued third edition of the Dictionary of

occupationaloutlook

MP STATES EPAtINO(T d DM, W. NMSwam of Labor Statistics Arthur M. Rosa. Conualoolorior

Mao& /14.140

5

OcOupational Titles.14 The Dictionary contains 18,000defined job titles arranged in an innovative structurewhich has unique connotations for the proposed fore-casting methodology.

ES Unfilled Job Openings

ES unfilled job openings by detailed occupation is anadministrative statistic which has been available in theEmployment Service for many years. Yet it was onlyrecently, as a result of the first experimental job vacancystudies in the United States in fiscal year 1965, conductedby the U.S. Department of Labor that the true dimensionof the unfilled job openings figure became known.

The job vacancy data showed that for the countryas a whole, ES unfilled job opening, in April i 965 ac-counted for 30 percent of the universe of job vacancies.15In the second round of experimentation, in April 1966,the ES share of total job vacancies had increased to 37percent. This average ,proportion of the job marketwhich incidentally exclude domestic service, varied byarea in April 1965, ranging from 13 percent in the lowestarea to 44 percent in the highest."

The relationship of ES unfilled job openings to totalOccupational needs in a community was literally a researchbombshell to the Employment Service system and to manypublic and private economists. Domestic and foreignmanpower experts were surprised to learn that the ESjob market exposure compared favorably with some of thehighly regarded systems abroad where the employmentservice is an instrument of an active manpower policy andcompletely acceptable to the entire community.17 It isthe ES unfilled job openings' penetration of the jobmarket that constitutes the core of the proposed newmethodology.

In applying the technique, the first step is to obtain alisting of the ES unfilled job openings in an area fromthe local public employment office.

These will most frequently be in 3-digit occupationalgetups 18 whether based on the second edition or the thirdedition of the DOT. It is a matter of relative ease for

'Dictionary of Occupation Titles, 1965 vols. I and II, thirdedition, U.S. Department of Labo'r' Bureau of EmploymentSecurity.

15 Hearings before the Subcrinmittee on Economic Statistics ofthe Joint Economic Comm;tee 89th Cong., May 17 and 18,1966. Statement of V. O. CI:amid, Director, Office of Man-power Analysis and Utazation, U.S. Employment Service.

"Findings and Implications of the Job Vacancy ExperimentalProgram, condsdted during FY-1965 and FY-1966, USES,a prelimtnar; paper dated Sept. 8, 1966.

" "The Measurement and Interpretation of Job Vacancies";g.. anference report of the National Bureau of Economic Re-search, Columbia University Press, New York, 1966.

"AU DOT occupations are grouped into nine broad categoriesidentified by the first digit 0-9. These nine categories are dividedinto 84.two-digit divisions, and the divisions, in turn, are subdi-vided into 603 distinctive three-digit groups. The latter aresubdivided into some 3,100 six-digit occupations containing some18,000 defined jobs.

6

the Employment Service to provide occupational detailbeyond the 3-digit level.

The most important element in the unfilled job open-ings figure is not the absolute level of the openings by oc-cupation but a corollary figure, namely the number of un-filled openings that have been unfilled for 30 days ormore." It is the number and percentage that unfilledopenings 30 days or more is to total unfilled openings forany 3-digit (or 6.digit) occupation over a period of timethat provides the indispensable clue to current shortagesand to a long-range projection in that occupation or oc-cupational group.

These unfilled jobs open 1 month or more and theirrelation to total unfilled openings for each occupationshould be evaluated for successive quarters, for a mini-mum of one year if possible, since seasonality will varythe figures of some occupations substantially.

If the hard-to-fill percentage for occupation overa. period of quarters ranges over 40 percent for anumber of quarters, there is evidence to indicate thatthe occupation in large measure is in a continuing short-age situation, certainly up to the most recent compilation.Time occupations should be identified on a listing andwill form the cohort for the next step:

Starting with the list of hard-to-fill occupations, andusing the actual number of hard-to-fill openings, that is,those unfilled 30 days or more, as a criterion, the listingshould be subdivided into a number of categories. Thesubdivisions will be subjective depending on the size ofarea and an analysis of the array. In a large metro-politan area, for 'example, those occupations with morethan 100 hard-to-fill openings which tend to .be repeatedover a period of time should go into the first category.Into the second category might be placed those whichhave between 20 and 99 hard-to-fill openings. The re-mainder of the occupations will constitute the finalcategory.

In medium-sized and smaller areas, the categoriesshould be scaled down by rule of judgment.

The Wage Factor

The ES reporting program not only collects hard-to-fillopenings but also the reason for their being hard-to-fill.'The category of reasons for hard-to-fill are (1) shortageof qualified workers; (2) low wages; (3) working condi-tions; and (4) other. Low wages as defined here is nota substandard wage since the Employment Service is notpermitted to accept such job orders. Rather it is a lowwage in relation to other jobs in the area.

On the other hand, if job vacancy data are used ratherthan unfilled job openings, a certain number will behard-to-fill because they pay substandard wages for theoccupation. On a national basis, this proportion amountsto a little under a fifth of the vacancies.2° Consequently,

"These duration data are collected quarterly by the ES as partof its 80-city survey.

"Hearings before Subcommittee on Economic Statistics, pp.74-75. (See footnote 15 for full reference.)

4,

4

a refinement may be introduced into the vacancy data byusing the hard-to-fill figure less those attributed to sub-standard wages.

Are jobs which cannot be filled because of low wagesin relation to the community wage structure to be con-sidered as unsuitable for training? 'This problem hasbeen the subject of an evolving philosophy with manyramifications and is covered in a number of administra-tive orders and policies within the Labor Department'sManpower Administration? 1 Among other things, thepolicy is not to fund a MDTA training course for anoccupation which pays below the national minimumwage. On the other hand, jobs in short supply becauseof a low wage which, nevertheless, pay above the mini-mum wage and the prevailing wage in the communityare considered trainable occupations from the standpointof the law. The means taken to assure that trainees willaccept such jobs is to screen them and elicit their attitudesin advance. Frequently, these trainees are disadvantagedand the prevailing wage, although low, is more than theyhave been accustomed to receiving.

While the foregoing is obviously not intended to be asophisticated discussion of the wage problem in assessingopportunities, it highlights a difficult problem which theEmployment Service faces up to all the time in suchdeterminations.

Expansion Via Job Vacancy Data

The next step in the progression of this technique is torelate long-duration unfilled job openings to the universeof hard-to-fill vacancies in the area, if such data are Avail-able. This relationship will produce a series of factorsby which a "blowup" of the unfilled job openings data touniverse proportions will become possible. Dependingon the scatter of the data, the blowup factor could beapplied at the one-digit occupational level, or the three-digit, or even the six-digit level. Because of seasonalchanges, care should be exercised to use like time periodsin expanding the data.

It is assumed, of course, that job vacancy surveyswould be patterned after the Department of Labor ex-perimentation which recommends collection of vacancydata by detailed six-digit DOT titles, a breakout of hard-to-fill vacancies ( jobs vacant 30 days or more) , andwage data for each job vacancy.22

The Congress once again turned down departmentalrequests for an extension of the vacancy program to 80major metropolitan areas for fiscal year 1967. An assess-ment of prevailing administration and congressional moodin the light of low unemployment rates and Vietnam ex-penditures would seem to indicate that funding of an ex-pansion of the job vacancy program beyond the currentexperimental pace is not in the offing. Neyertheless, thereare a number of mitigating circumstances which can be

Manpower Administration Order 33-65. "Wage Rates Ap-propriate to MDTA Projects in Occupations Not Covered bythe Minimum Wage Act," Dec. 29,1965, and U.S. EmploymentService Program Letter 1994; Mar. 25,1966.

" Instructions for Obtaining Area Job Vacancy InUSES, Mar. 15,1966.

. . .

used to harness this tool for greater usefulness in thecurrent technique.

First, the current experimental surveys for which Con-gress has appropriated funds for 3 successive years coverssome 16 areas compring one-fourth of the Nation's workforce. By itself, the relationship of job vacancies to jobopenings by occupation for so large a. segment of thecountry is a major contribution to establishment of ablowup factor for more accurately determining areaneeds.

Second, the same experimental program has the poten-tial for expanding the vacancy program. As now costedout, the appropriations permit conduct of Surveys in 16areas two times a year. If each of these 16 areas were toconduct one survey per year, the money so saved couldbe used to introduce at least a like number of areas tothe program. By so doing, areas with perhaps as muchas 40 percent of the Nation's work force could be estab-lishing job vacancy benchmarks against which to expandthe Employment Service unfilled job openings figures.

The 0 ccupational Outlook Handbook (OOH)

The OCCUPATIONAL OUTLOOK HANDBOOK iS a docu-ment of prime importance containing some 850.pa,ges o5data covering outlook for more than 700 occupations.Projections are on a national basis and most frequentlyof a qualitative nature. forecasts are compiled by agroup of trained economists in the Bureau of Labor Sta-tistics utilizing resources within the Labor Department.To this is added the resources of other governmental agen-cies the trade associations and unions.

This major work, updated every 2 years, enjoys nationaldistribution and is a basic tool for all Employment Servicecounselors. Employment Service familiarity with theOOH makes it an ideal component for use in the pro-posed techniques.

The next step, after an array of occupations in shortsupply by category of volume is achieved, is to make anassessment of long-range opportunities in those occupa-tions. This is accomplished by reference to the OCCUPA-

TIONAL OUTLOOK HANDBOOK with its national projectionsby occupation. Against each of the occupations listed inthe array, a notation would be made of national long-range opportunity, adding by some prearranged code thesize of the occupation and the intensity of growth andshortage. A simplified approach would be for the Bureauof Labor Statistics to make a determination of long-rangedemand based on size and intensity of need for all theoccupations listed in the Handbook as they have alreadydone for skilled occupations.23 Once done, this wouldbe distributed to all users and would have ihe effect ofsubstantially cutting down on the time needed for indi-vidual determinations and also provide uniforinity in theinterpretation of the data.

At this point, the economist brings to bear a series ofjudgments and interpretations, the same as he would

"Estimated Need for Skilled Workers, 1965-75,"MONTHLY LABOR REVIEW, April 1966, U.S. Departmentof Labor.

7

.s.

Ib

Estimated Job Openings for Skilled WorkmenSelected Occupations, 1965-75

in

WIWI* NUI111111

/OD (3/410004%

I...wgon,OCCUPATION

"PROPORTION OF JOB OPENINGS RESULTING FROM-

NET GROWTH DEATH ANO RETIREMENT

55 Fronting craftsmen. exceptbookbinders

245 Carpenters v ,s,

115 .5nitionary engineers

105 Allrauni mac Wats A

0 Ilricelsetrs

I 3 3 Electricians

230 Automotive mechanics

4.040 Craftsmen. foremen, and kin.Wed woollen. all Industrie

40 TV and re ;.o servicemen

40 Rumness machine servicemen

_t

100 75 50 25 0 25PERCENT

1

50 75 100

Source: U.S. Bureau of Labor Statistics.

under any other technique employed by him. Where thenumber and percent of hard-to-fill is large and has beenlarge for some time, and where the Handbook indicatesnational expansion, it is entirely. reasonable to assumethat expansion will occur in the area.

Is there any danger in applying the national trend tothe local scene? On the whole, we think not because thebulk of the areas where the occupation exists of necessitywill follow the point of central tendency; namely, thenational projection. Moreover, and this point needs em-phasis for it is the crux ef the matter, it has already beendetermined in advance that these occupations exist inlarge numbers ire the community and have already beenadjudged hard-to-fill on a past and current basis. Con-versely, the danger of applying a national projection to alocal area would be real where the occupation existed invery small numbers and did not have a hard-to-fill history.

Let us examine existing situations in which a nationalforecast might be applied to a local area, and determinewhether such situations can be rationalized. Keep inmind that in each of the illusdations, the occupations existin important numbers irt'Ole community and that therecurrently exists, and has fa: a year or more, a conlitionof shortage:

1. The cress- industry occupation.If the national pro,jection indicates expansion for such occupations as book-keeper, stenographer, or office machine operator, thebroad-based nature of the occupation in the communitywould present little chance for deviation from the nationaltrend.

2. The specific industry-occupation tied to broad-gagedincome trends.If the national projection shows expan-sion for such occupations as bank teller, barber, andoccupations in cleaning establishments, the risk of apply-

8

sr'

ing growth to a specific community (where shortage al-ready exists) is quite minimal,.

3. The specific industry- occupation tied to nationalgrowth and institutional factors.Would anybody ques-tion the application of national growth in nurse and otherhospital occupations to a local community where there isalready abundant evidence of shortage in these sameoccupations?

4, Another possibility.What about the occupation ina specific industry in which the national long-range pro-jection is favorable, in which there are current localreports of shortage of certain workers in that local indus-try, but which the long-range projection for the plantin the community is not favorable. We have in mindthe Studebaker situation in South Bend where there mighthave been needs reported for tool and die workers andmachinists up to the day it was announced that the plantwas going out of business. The economist in this case,knowing the precarious status of the corporation, wouldmake an adjustment reducing the size of the local need.

.But suppose, as actually happened, the intention 'toclose was a closely guarded secret? The answer is obvious.No other known technique of forecasting can or wouldhave produced the right answer either. As we will eithersay or imply throughout this paper, the unfilled openings -OOH approach is better than most forecast devices inmany instances and no worse than most such devices inother instances. Saying it another way, the unfilledopenings -OOH technique is capable of making the sameerror as other techniques but with much less cost; trouble,and effort.

5. The industry - occupation which is expanding in alocal community but for which the national prognosis iscontraction.An illustration could be the location of amechanized meatpacking plant in a new community toconsolidate the operations of one or more inefficient plantselsewher. The local economist can be reasonably securein his judgment that the recently constructed operationis competitive with the best. in the industry and hencewould have to be regarded in a different light from theemployment downtrend in the industry as a whole. Ifsuch a plant exhibited hard-to-fill openings over a periodof time, rule of judgment would seem to indicate a positivetraining decision.

Application -If the ES unfilled openings -OOH tech-.nique to the DiC "jet of Columbia metropolitan area af-forded insight. mt.) the applicability of the 00H data tothe District of Columbia's unfilled job openings series.Of the 65 occupations, some three-digit, some six-digit,from the unfilled openings listing that warranted seekinglong-range information, usable data were available for 50of them in the OOH. The 15 for which there was noinformation were mostly semiskilled and unskilled jobssuch as messengers, charwomen, demonstrators, porters,guards and watchmen, and other groups of unskilled suchas laundry workers, warehousemen, and packing, filling,and labeling workers.

It is not altogether surprising that these occupationsshould have been omitted from the OOH. They requiredlittle or no training tim e and there may have been a pre-vailing opinion that enough workers would always be

sr

Employment of Nonagricultural Wage and SalaryWorkers, by Industry, 1964, and Projected Require-ments, 1975 1

[In thousands]

Actual 1964employment

Projected 1975requirements

Pcrcentchange

Total 50,156 75,875 30

Mining 633 620 etContract construction 3,056 4,190 37Manufacturing 17,259 19, 740 14

Durable goods 9, 813 11, 500 17Ordnance and accessories._. 247 250 (1)Lumber and wood prod-

ucts , except furniture az 550 9Furniture and Attune-- 406 510 26Stone, clay, and ghee prod-

ucts 612 876 10Primary metal industries 1,231 1,290 5Fabricated metal products 1,187 1, 460 23Machinerr3r 1,606 2, 050 28Electrical equipment and

supplies 1, 548 2,000 29error nation equipment 1, 606 1,730 8

Motor vehicles andequipment. 755 800 6

.6.irauft and parte 604 515 5Instruments and related

products 339 510 38Miscellaneous manufactur-

ing industries'

399 475 19Nondurable goods 7,446 8,240 11

Food and kindred products. 1, 746 1,855 5Tobaczo manufacturers-- 89 80 10Textile-mill products 891 880 (I)Apparel and related prod.

ucts 1,302 1,525 17raper and allied products_ 625 775 '024

Printing, publishing, andallied products 951 1,100 10

Chemicals and allied prod-ucts 877 1,125 28

Petroleum refining and re-lated industries 183 160 13

Rubber and miscellaneousplastic products 434 580 34

Leather and leather prod-ucts 348 350

Transportation and public utilities 3,947 4, 425 12Trade, wholesale and retail 12,132 16,150 33Finance, insurance, and real estate 2,264 3, 725 26SerViCee and miscellaneous 8, 569 12,275 43Total government_ 9, 595 14, 750 54

Federal government 2,348 2,525 8State and local government-- 7, 248 12,225 69

*Projections assume an unemployment rate of 3 percent in 1975.'Less than 3 percent.Now.: Because of rounding, sums of individual items may not equal totals.

Source: U.S. Bureau of Labor Statistics.

available for unskilled jobs. In fact, omission of manyof these types of occupations from the OOH, from BESsill surveys, and from State ES occupational guide pro-grams lent support to contentions of the Office of Eco-nomic Opportunity that the Department of Labor wasnot sufficiently sensitive to the needs of the poverty-type,disadvantaged workers for whom these jobs could providenumerous opportunities.24

"The Manpower Administration has for approximately ayear prior to the date of this writing reoriented its training policy.The policy now provides that some two-thirds of MDTA ex-penditures .are to be allocated to making disadvantaged groupsemployable and one-third to training for higher skill occupationalshortages.

The experimental job vacancy and unfilled job open-ings surveys demonstrated clearly that there were manyjobs of a semiskilled and even unskilled nature that couldnot be filled for at least 30 days, with the difficulties notbeing attributed to low wages.

The ES unfilled job openings -OOH technique, by itsdesign, picks up this generally neglected group. If theseoccupations have a substantial number of hard -to -fill jobsas shown by the unfilled job openings survey, they areincluded in the original list of occupations for study. Thehard-to-fill figure is considered a measurement of neededsupply, no matter how many such workers may be listedin the active file of the employment office. If wages isnot the factor for inability to fill the job, and it is unlikelysince the Employment Service cannot accept job orderspaying less than the prevailing wage, it can be concludedthat a shortage of workers exists at the time of the surveyunder prevailing conditions. If the shortage has pre-vailed for three or four successive quarters, and one as-sumes further that relatively full employment conditionswill prevail, the information thus obtained can provideinformation on job opportunities and training needs forlesser skilled workers.

Occupations wit% Small ES Penetration

A major difficulty in applyiing this technique is relatedto those occupations in which the Employment Servicedoes relatively little business and in which the numberof unfilled openings is consequently small. Thus, banksmay not be likely to come to the Employment Service forrecruitment. Local utilities, railroads, craft unions, etc.,may similarly conduct the bulk of their recruitment else-where. For a relatively large metropolitan area, for ex-ample, this situation may show up in the unfilled open-ings tabulations as nine openings for printing pressmenwith six of them indicated as hard-to-H.25 A perusal ofunfilled job opening lists in a number of cities shows manyentries of this type.

Were these entries to represent little or no demandfor workers in the community, they would not cotcernus. If, on the other hand, the unfilled job openings num-ber is small in the Employment Service but large in thecommunity, they are of direct importance to the forecast-ing process. The Employment Service people in the com-mumty will know how to make a distinction between thesetwo groups because they have learned over time andthrough job development efforts which are the employ-ers, industries, and occupations where they have smallrepresentation.

Apart from local office knowledge and judgments,however, there are certain safeguards written into thesituation. In terms of total actions in the job market, thebulk of such actions take place in a relatively few oc-cupations. Thus, only 43 (three-digit) occupations com-prise 66 percent of the total job vacancies and these same

2* Actual listing for the District of Columbia metropolitan areafor June 1966. This appears to be a clear case of under -re pre-sentation, since the area is a relatively large printing center.

9

Job vacancies and unfilled job openings: Selected three-digit occupations' as a percent of each major occu-pational. group

Major occupational groupPercent of

all jobvacancies

Percent ofall unfilledopenings

Total, all groups, 43 occupations

Professional and managerial;, 11 occupations (including 6 categories of engineers)Clerical and sales, 11 occupationsService,2 4 occupationsSkilled, 8 occupationsSemiskilled, 5 occupationsUnskilled, 4 occupations

66 58

6278525871

3 56

656447506427

1 List of 43 selected 3-digit occupations, second edition of DOT.Profirssional and Managerial: Engineers (metallurgical,

chemical, civil, electrical, industrial, and mechanical), socialand welfare workers, trained nurses, draftsmen, laboratorytechnicians, healers and other medical occupations.

Clerical and Sales: Bookkeepers and cashiers, clerksqierteraloffice, general industry clerks, office machine operators,secretaries, stenographers, and typists, stock clerks, salesmen-insurance, salesclerks, salespersons, salesexcept to consumers.

Service: Waiters and waitesses, kitchen workers, attendantshospital, porters.

Skilled: Machinists, tool and die makers, machine shop

occupations account for 58 percent of the ES unfilled jobopenings. Hence, for those occupations numericallyimportant in any community, the local EmploymentService office most probably will have representation.The table and notes above express this information.

Nevertheless, it will still prove desirable to ferret outthose opportunities existing in relatively large numbers inthe community which are not reflected in ES unfilled jobopenings. To do this, the local job market analyst inconjunction with employer relations representatives,placement specialists, and counselors (in the local employ-ment office) would make a List of occupations and in us-tries, and from them of major establishments and localtrade associations representing those groups in which theEmployment Service does little or no business. By way ofillustration, such a list might include the local telephonecompany, the electric light and power utility, the con-struction trades council, the one or two railroads servingthe community, etc. An aid in compiling such a listwould be the occupations appearing in small numbersamong the unfilled job openings. It is relatively easy tospot these occupations and their related associations."

The author tested this hypothesis by scanning the District ofof Columbia metropolitan area unfilled job openings list. Famil-iarity with the employment pattern in the area enabled him toidentify quickly such occupations as upholsterer, pressman,electrician, plumber, practical nurse, bank teller, surveyor, tech-nician, and telephone operator for which it would have beenonly a formality in most instances to obtain the associated orga-nizations or establishments.

occupations, welders and flame cutters, carpenters, plasterers,mechanics and repairmenmotor vehicle, mechanics andrepairmenother.

Semiskilled: Textile products occupations, machine shopoccupations, occupations in the treatment of metals, chauffeursand drivers, occupations in laundering, cleaning, etc.

Unskilled: Occupations in meat products, packing andfilling occupations, transportation equipment laborers, ware-housing and related occupations.

2 Excludes domestic service occupations,.s May be some distortion because of sample inflation.Source: U.S. Employment Service.

The next indicated task would be a series of visits tothese groups and for judgments to be made of the demandand supply situation in these occupations. Judgmentsbased on these interviews would range from the simple tothe complex.

For example, it should be relatively easy to determinefrom the telephone company the job prospects for tele-phone operators, linemen, and a few other important andunique occupations.

At the local hospital or hospital association, it wonid bedifficult to conceive of any reason why a hospital admin-istrator could not list a number of occupations in whichhe is currently experiencing recruitment difficulties oranticipates problems. Keep in mind that the Employ-ment Service interviewer will already have many answersbefore he talks to the hospital administrator. He willknow of the difficulties in the clerical and stenographicgroups, and in the supportive occupations such as main-tenance men, cooks, dishwashers, porters, etc. He willmostly be seeking knowledge in those occupations peculiarto hospitals.

When the Employment Service man speaks to thecarpenter union official, he will be inquiring only after theorganized craftsman and apprentice. He will alreadyknow the job market for the nonunionized carpenter andhelper because these groups are frequently well repre-sented in the local offices penetration patterns.

In still another illustration, suppose it's the printingoccupations in which there is one employer association, aunion local, and two or three large employers and a host

of smaller ones. If the employer group or labor unioncould not adequately furnish the answers, then con-ceive ably some telephone inquiries to 'the two or threelarger ones, utilizing elements of the training needs surveyapproach 27 of the Employment Service so successfullyused to implement the Manpower Development andTraining Act would be appropriate. This would be themost time-consuming aspect of the study and would prob-ably be an infrequent occurrence. It must be t:orne inmind that exact parameters of need are not necessary.Where the figures are indefinite but large, a numter ofclasses could be set up without risk. Where the numbersare small, perhaps one class could be initiated, either invocational education, or for MDTA institutional or on-the-job training.

Admittedly, this is the segment of the proposed tech-nique for which precise experience and experimentationare needed. In a very large metropolitan area, wevisualize perhaps some 20 scheduled visits. Between twoand three could be scheduled per day and would representthe largest single output of time in the study. In smallareas, these visits would be scaled down accordingly.Visits should be structured, that is, a questionnaire shouldbe prepared and sent out in advance. It might provefeasible for the analyst to be present at a meeting of theassociation or union to conduct a group interview.

An on-going job vacancy information collection pro-gram would obviate the necessity of association or em-ployer visits for purposes of this technique because thedata so collected would repreesnt the universe of oppor-tunities in the irea.

Unfilled Openings and the Business Cycle

The number of ES unfilled openings in 1966 reachedits highest peak in the history of the Employment Service,corresponding to the equally unprecedented boom in theNation's industrial establishments. Similarly, ES unfilledopenings peaked in 1953 and 1956, 2 years in whichemployment opportunities were also plentiful. On theother hand, and as might be expected, the number of ESunfilled openings showed cyclical declines during therecession years of 1954 and 1958.

It would be reasonable to ask, therefore, whether theapplication of the proposed technique would suffer whenunfilled openings dropped off in periods of economicrecession.

A reduction in numbers of ES unfilled openings, be-cause it would produce smaller magnitudes, would prob-ably subtract somewhat from maximum usefulness of thetechnique when compared to its use in a brisk job market.

. Of course, low points in a business cycle tend to affectother techniques of long-range forecasting. Employerforecasts in periods of recession tend to underestimateneeds substantially. And for those economists relying ondecennial census data to establish a projection, how many

4 incipient dangers would be present if the census count

MDTA HandbookChapter 11,, Training Needs Survey,March 1965.

took place in a year which was above or below the `norm"for the period?

The fact of the matter, however, is that a downtrendin the level of employers' orders would not basicallydetract from the usefulness of the ES unfilled openingsOOH technique. We believe that in the kind of jobmarket existing in the decade of the fifties and sixties,and probably in the decade of the seventies, there will belittle diminution in the nature of structural unemploy-ment. Mismatching'of worker and job will undoubtedlystill be considerable. Consequently, we would continueto expect a large proportion of hard-to-fill jobs.

There is some statistical support for this position. Anexamination of the job vacancy data in April 1965 andApril 1966 reveals that even in those metropolitan areaswith relatively high unemployment rates, the proportionof hard-to-fill jobs (those unfilled for 30 days or more)was relatively high. Specifically, in April 1965, Los An-geles and Charleston, (W. Va.) had an unemploymentrate of almost 6 perceni and New York almost 5 percent.,These were the three highest rates in the 16-city sample.In none of them was the proportion of hard-to-fill jobs forthe area as a whole less than 35 percent.

In April 1956, the same three cities still had the highestunemployment rates in the experimental study but theunemployment rate had fallen to approximately 4.4 per-cent for each. Again, the proportion of hard-to-fill jobsranged between 40 and 50 percent for each.

It is not unreasonable to assume, therefore, that unfilledjob openings will be capable of pinpointing the types ofoccupational needs at relatively high periods of unem-ployment. Moreover, a repeat study once a year (or morefrequently) will enable the analyst immediately to besensitive to changes in the job market and the meaningsuch changes would have for training plans andopportunities.

High-Employment Occupations in an Area

What of those occupations in which there is a consid-erable number of workers employed in the area and 'whichdo not show up in the hard-to-fill listings? If nothingelse, these occupations generate needs for replacement.Even if there were no long-term growth, it would not beprudent for vocational schools to eliminate training for,say, stenographers.

These occupations can be spotted by referring to theoccupational counts by area in the decennial census.Those with the largest numbers should be listed for furtherexamination. To the extent they duplicate those alreadylisted because of a hard-to-fill history, they present noproblem.

What if the high-employment occupations have nohard-to-fill history? It seems to us that the OOH outlookwould be controlling. If the outlook is for growth, wewould recommend continuance of the vocational educa-tion classes in the locality. If the occupational outlookis for neither growth nor decline, we would also continuetraining but watch placement experience carefully. If,on the other hand, the prognosis was for a decline nation-

II

ESTIMATED CAREER OPENINGS FOR CRAFTSMEN, FOREMEN, AND KINDRED WORKERS RESULTING FROM GROWTH OF EMPLOY-MENT REQUIREMENTS AND FROM RETIREMENTS AND DEATHS) 1965-75

[In thousands)

OccupationEmploy-

meut, 1965 iEmploymentrequirements,

1975

Career openings, 1965-75

Totalopenings

Netgrowth

Deaths andretirements

Craftsmen, foremen, and kindred workers, all industries

Selected building trades, all industriesIn construction industry

CarpentersIn construction industry

Bricklayers, stonemasons, and tile settersIn construction industry

Cement and concrete finishersIn construction industry

Electricians.In construction industry

Excavating, grading and road machinery operators..In construction industry.

Painters -

In construction induatryPaperhangers

In construction industryPlasterers

In construction industryPlumbers and pipefitters.

In construction industryRoofers and slaters

In construction industryStructural metal workers

In construction industry

Mechanics and repairmenAutomobile mechanicsAppliance servicemen_Television and radio servicemenBusiness machine servicemenIndustrial machinery repairmenVending machine mechanicsOther

Machining occupations ,.

All-round machinistsTool and die makersSetup menOther

Printing craftsmen, except bookbindersCompositors and typesettersElectrotypers and stereotyper,Engravers, except photoengraversPhotoengraversPressmen and plate printers

OtherStationaryengineersBakers 'Locomotive engineersMillwrightsSheet-metal workers 2Miscellaneous skilled workers

Foremen, not elsewhere classified1111=111011111111h .

9,220 11, 400 4, NO 2,180 1,860

2, 6901,915

8506652061806055

385165245190410285

11105050

34522060557040

2,33577019511575

16520

995

64032514545

125

290170

14111580

2, 000255105357550

1, 480

1,265

3,0802,175

8956702302000,7575

44520033026543530514136055

4252657065

10060

3,29588025014010518525

1, 710

7303E016050

160

290155

61515

100

2,38027395408065

1,825

1, 625-N,...

960670

24516560502530

13565

1251061309576

2015

1609020204530

1,390230864040407

950

234105451465

5520663

35

78585

. 10152025630

620

390260

455

25201520603585752520

33

'10+5

804510103020

960110552530205

715

9035155

35

015840

20

380201055

15345

360

570410

200160353098

75304030

1057543

1010704588

138

430120301410202

235

14070309

30

55le223

13

405652010159

285

260

The 1965 employment estimates for total skilled workers and the majoroccupation subgroups are from the Bureau's Monthly Repovt on the LaborForce, December 1965, table A-19. p. 42. The 1965 estimates for individualoccupations were developed in the course of the Bureau's occupational out-look research program. using a wide variety of sources including unpublishedCurrent Population Survey data, employment data derived from BLS wageand, other surveys, union membership figures. data collected by governmentregulatory agencies, and miscellaneous employment information obtainedfrom trade, associations and other private groups. Actual 1965 employmentlevels by industry and estimated industry occupational staffing patternswore an important tool in making the 1965 employment estimates. In addi-tion to employment data, use was made of related economic data in estimat-ing employment in individual skilled occupations. For example. in develop-ing estimates of vending machine mechanics in 1965 the number and kinds ofvending machines in use and their estimated-maintenance requirements wereconsidered in arriving at the final employment estimate of these workers.

Source: U.S. Bureau of Labor Statistics.

12

The classification of individual occupations into given groups is that used Lnthe Current Population Survey. It differs somewhat from the occupationalclassification system followed in the Occupational Outlook Handbook which isbased on an independent analysis of the job content of individualoccupations.For example. in the Handbook, millwrights are classified as mechanics andrepairmen and not as "other craftsmen; ' maintenance electricians are classi-fied as mechanics and repairmen and not as building trades craftsmen.

2 Includes an estimate of the number of street-metal worke-r in the con-struction industry and those performing construction-related work in allother industries. Excluded are workers such as aviation metaismiths,pattern cutters and developers in the motor vehicle Industry, and templatelayout men in industry, as classified in the U.S. Census of Populationcategory, "Tinsmiths, coppersmiths, and sheet-metal workers."

Nora: Because of rounding, sums of individual items may not equal totals.

ally and there was not a hard-to-fill experience locally,we would recommend contraction of training in theabsence of other extenuating circumstances.

The New Dictionary of Occupational Titles (DOT)

The third edition of the DOT, released in 1966, pro-vides a different structure of occupational classificationfrom the two editions preceding it. Among the innova-tions introduced was one which has significance for thecounseling field and for vocational education and MDTAtraining programs. The structure as now devised, groupsoccupations according to some combination of body ofknowledge required, purpose, industry, materials workedwith, product, and/or service.

The end result is a structure which clusters occupationswith like characteristics under a common heading andalso which groups occupations in a "skill ladder."Specifically, with respect to clustering, under metal ma-chining are included the following 10 three-digit occupa-tion groups : Machinists and related, toolmakers and re-lated, gear machining, abrading, turning, milling andplaning, boring, sawing, and metal machining occupa-tions, n.e.c. The latter three-digit grouping contains the"production-machine operator", which together withentry-level jobs in the other three-digit groups, feed intovirtually any of the 10 three-digit groupings.

With respect to the skill ladder, each three-digit occu-pational group has within it the range of occupations pro-ceeding from the top skill in the group down the promo-tion ladder to the entry occupation. Thus, for the three-digit occupational group "machinist and related,", there'are included the highly skilled machine shop foreman,the all-round machinist, layout man, machinist apprenticeand instrument-maker helper. These occupations areonly a sampling of the full range of occupations includedin the three-digit level of complexity skill ladder.28

Since each of these three-digit occupational groupsrelated to one another in terms of materials worked with,product, etc., it follows, therefore, that at the entry levelof all of these occupational groups, the training requiredis somewhat similar. Consequently, job opportunities atthe entry level would be the number of needs listed forall of the occupations in the 10 three-digit groups. Forsurely, if there is an unfilled need for a machinist, aboring machine operator, and an entry production-machine operatorthree in allthen presumably in theabsence of skilled supply, the need would be to train threeentry workers in metal machining occupations. If anemployed boring machine operator had sufficient skill toqualify for the machinist's job through upgrading, thatwould still leave a need for three workers; i.e., two boringmachine operators and one entry production-machineoperator; assuming no adequate supply, the trainingwould still be in terms of three entry workers in metalmachinery occupations. To pursue the point for thesake of clarity, if both a machine operator and an entry

vol. giird edition of the Dictionary of Occupational Titles."For full listing of occupations in three-digit group, see p. 121,

production-machine operator were upgraded, the netneed would remain at one machine operator and twoentry production-machine operators; assuming no supplyfor the machine operator, the training need would remainat three entry workers in the metal machinery occupations.

If the above thesis is correct, it follows that any surveywhich estimates total opportunities for a list of skills suchas baker, machinist, or electrician is only measuring apart of the need. The forecasting technique would haveto consider the needs of the entire family of occupationsboth horizontally to include a number of related six-digitoccupations and vertically to include the entire three-digitskill ladder.

There are 603 three-digit occupational groups in thethird edition of the DOT. In the case of metal machin-ing, it was feasible to group 10 such three-digit groups-600 through 609for a family cluster. However, thefeasibility of grouping other three-digit groups is a matterfor different handling. Some clusters might consist ofonly one or two three-digit groups. The task of dividingthe 603 three-digit groups into homogeneous clusters isnot too difficult and would take a trained occupationalanalyst only a matter of a week or two. This clusteringdetermination would need to be done only once afterwhich it could serve all corners until such time as theremight be a major change in the definition and structureof occupations.

The theory of occupations clustering provides thejustification for adding together one or more three-digitcodes. In the unfilled job openings-00H technique, thesame theory justifies the addition of related groups ofthree-digit occupations. This would result in fewer oc-cupational identities but in larger numbers attached tothose groups which emerge.

Let us assume further that a vocational educationcourse or an institutional course under' MDTA actuallytrains a student to the level of a skilled worker, or forthat matter to any level identifiable by a six-digit code inthe new DOT. All that is needed is to break out thatsix-digit code in the listing of unfilled job openings orjob vacancies and to correlate it with the OOH outlook.

Here then, in summary, is a. grouping technique whoseusefulness for long-range forecasting and for vocationalplanning and MDTA is well-suited.

The educator who is anxious to obtain a cur-riculum which exposes the student to a skill clusterrather tharAto an occupational pinpoint will find thatthe new DOT provides this mechanism;

The job market analyst, whose problem is tomeasure the opportunities in a total field of workwith similar job characteristics an skill transfer-ability, will achieve greater reliability in hisestimates;

At the same time, with slight modification intechnique, the DOT can provide both the educatorand the job market analyst with information on

. opportunities in specific occupations (e.g., key punchoperator, stenographer, auto mechanic, and welder)if the job market demands such workers and thetraining curriculum can produce qualified trainees.

13'

There is another development in the field of clusteringwhich deserves mention. The U.S. Office of Educa-tion is currently evolving a clustering technique based onoccupational titles in the third edition of the DOT.This technique is predicated on starting with the coursesgiven in the vocational education system and classifyingunder each of them the occupations which are part ofthe "family!' Such a grouping would merit special con-sideration because it is tied to a pragmatic situation,namely, the 'actual courses as they are structured andidentified in the vocational education system.

Postulates of a New Technique

An assessment of any new technique should take noteof the following basic observations:

1. The foremost consideration is that any . techniquedeveloped for widespread use to produce quick resultson a continuous basis must be a relatively simple one.Budgetary considerations inhibit, if not rule out, an ex-pensive approach, Lack of technical know-how andbudgetary resources probably rule out those econometricmodels produced thus far.

2. The quest for preciseness and for emerging occupa-tions in the forecasting area is not necessarily meaningfulfor training purposes.

3. The unfilled job openings (job vacancy) 0ocupA-moral. OumooK HANDBOOK (matrix) approach is notwithout its limitations. Its limitations, however, are noworse than those in other known techniques and theresults, in our judgment, are at least as good and perhapsbetter.

4. A forecasting system not predicated on the measure-ment of supply as an offset to demand is of overallquestionable value. The hard-to-fill (openings unfilled1 month or more) concept, albeit a simple one, is anadequate and realistic supply measure.

5. Any occupational forecasting system which does nothave basic data originating on the local scene runs achance of being divorced from reality. The unfilledopenings -OOH approach starts with, and indeed is predi-cated on, a local area's experience, currently and over aperiod of time. Of course, unfilled openings data maybe analyzed, geographically, on an area larger than astandard metropolitan area. In fact, we recommend astatewide survey, broken down into a number of smallercomponents, namely geographical units of SMSA's.

6. The third edition of the DOT is a technical devicewhose potentialities for providing information on cluster-ing and avenues of promotion should be ha messed to along-range forecasting technique. The U.S. Office ofEducation clustering concept, translated into DOT terms,may be a useful supplement, orpossible substitute.

7. Applying the unfilled job openings-OccupAnoNALOUTLOOK HANDBOOK approach requires judgment of thelocal employment office economist or job market analyst.The job market analyst on the scene, dealing with datagenerated out of his own operations and with continuouseconomic review of his area, is in the best position to drawvalid conclusions. Moreover, by placing this responsi-bility in the local office, the Employment Service will be

14

discharging its mandate under the law to provide suchinformation.

Conclusion

Development of a technique to forecast long-rangeoccupational requirements to satisfy the needs of theManpower Development and Training Act of 1962, theVocational Education Act of 1963, and the EconomicOpportunity Act of 1964 has bedeviled manpowers econo-mists for years.

The area skill survey technique, ,the best known andlongest used, has come ih for severe criticism on thegrounds that it is too expensive and time-consuming, andthat most employers are not good economists.

The dilemma has been that up to now no one (in ourjudgment) has come up with a viable technique of fore.Casting long-range occupational requirements. The of-fice of Manpower Policy, Evaluation, and Research inthe Department of Labor has funded a number of con-tracts with universities under title I of the ManpowerDevelopment and Training Act to produce acceptabletechniques. A modest estimate of these costs totals about$150,000. The resulting techniques are extremely ex-pensive and difficult to perform. We leave an assessmentof their technical adequacy to other reviewers.

The Bureau of Employment Security, estimating thatskill surveys cost an average of $20,000 each, requestedfor Escal year 1968, $3 million over a period of 2 yearsto conduct such surveys in 150 of the largest metropolitanareas. A like amount would be requested .every 2 or 3years. It is true that these surveys would also have pro-duced employment benchmarks for a limited number ofoccupations by area, a product not visualized in the ESunfilled openings -OOH approach.

The ES unfilled openings -OOH approach, in our judg-ment, can and will work. It is relatively simple and canbe transmhted quickly. Its elements are the unfilled jobopenings in a local employment office, identification ofthose jobs open 30 days or more, and the addition of anational forecast tempered by the, local analyst's knowl-edge and supplemented by a few association visits.

It is possible that an indeterminate number of areas,because of an unusually low number of ES unfilled open-ings and the absence of job vacancy data, may not besuitable for the application of this technique. The scopeof this limitation rests on individual examination of eacharea's data.

We estimate that the conduct of an unfilled job open-ings -OOH survey to satisfy vocational education, MDTA,and Office of Economic Opportunity needs would take asingle manpower economist an average of not more than6 weeks for a survey in a metropolitan area. Total costto the Employment Service for 150 areas: Up to $200,000and undoubtedly less as experience is gained, all 150areas to be completed in 1 year and repeated at IL-yearintervals. This sum does not include collection of ESunfilled openings data which are now collected in 80 areasand. are scheduled to be expanded to 150 areas, with orwithout the application of this technique.