TRAINING AND LINKS TO THE LABOR MARKETsiteresources.worldbank.org/INTSRILANKA/Resources/... ·...

93

Chapter Five

TRAINING ANDLINKS TO THELABOR MARKET

5.0 Introduction5.1. High unemployment ratesamong educated young peopleand their low skill levels areissues of critical policy concernfor GOSL. Youthunemployment, resulting mainlyfrom prolonged job search, is ofparticular concern given thehistory of social unrest overyouth joblessness. This iscombined with another policyconcern that school-leavers -grade 9, GCE O and A-levels,and university graduates areentering the labor market ill-prepared for the world of work.An important response fromGOSL has been to developtechnical education andvocational training to facilitatethe school to work transition andto reduce skills gaps and skillmismatches in the labor market.

However, the information baseon the technical education andvocational training (TEVT)system is weak, with no majorreviews of this sector in recenttimes. This chapter seeks to fillthis important information gapby undertaking an analysis ofTEVT to complement theanalysis of the general educationsystem.

5.2. The chapter focuses ontwo principal areas. First, itprovides an overview of theoverall policy framework for theTEVT sector. It examines: (i)the governance and financing ofpublic sector traininginstitutions; (ii) the supply ofpublic sector vocational trainingand technical educationopportunities, the range oftraining programs and coursesavailable, enrollments over time,

and their regional coverage; (iii)the environment for privatesector participation in training;and (iv) the evolution ofthinking about TEVT and itsrole in the economy, andformulation and implementationof training policies. Second, thechapter uses time-series datafrom the Labor Force Survey(LFS) to investigate the linkagesbetween technical education andvocational training on one hand,and labor market outcomes onthe other hand. This analysiswill focus on (i) the demand forpost-school technical andvocational training, especiallyamong school completers at thekey exit points of the schoolsystem and between gendergroups; (ii) how the demand fortraining varies over the work-life; and (iii) the labor marketoutcomes of investments in

different types of training,especially their impacts onunemployment, job searchduration, and earnings. Thechapter concludes with asummary of findings andimplications.

5.1 TVET InstitutionalFramework and PoliciesInstitutional Framework

5.3. The Technical Educationand Vocational Training (TEVT)sector in Sri Lanka is currentlymade up of an extensive systemof public, private and NGOsector training providers. Publicsector TEVT providers in Sri

Lanka used to function underdifferent ministries up to themid-1990s. In 1994 however,the TEVT sector was elevated toa ministerial function and thekey TEVT providers placedunder the supervision of oneministry, the Ministry of TertiaryEducation and VocationalTraining (MTEVT).18 As part ofthis reorganization, the Tertiaryand Vocational EducationCommission (TVEC) wasestablished and now functions asthe apex body for setting policyand regulating TEVT sectoractivities. In 2001, there wereabout 920 training institutesregistered with the TEVC

comprising of 556 institutions inthe public sector, 252 in theprivate sector and 112 in theNGO sector. In addition, asizable number of private sectorproviders operate in the marketwithout registering with theTEVC.19 The broad institutionalframework of the TEVT systemin Sri Lanka is presented inFigure 5.1.

5.4. Among public sectorinstitutions, the key providersinclude the Department ofTechnical Education andTraining (DTET), NationalApprenticeship and IndustrialTraining Authority (NAITA), Sri

94

18. Prior to that, under the Ministry of Labor and Vocational Training and the Ministry of Vocational Training and RuralIndustries.

19. According to the Ministry of Tertiary Education and Training (MTET), there were around 1,000 unregistered vocational andtechnical institutions operating throughout the country in 2003.

Figure 5.1. Institutional Framework of the TVET Sector

GOVERNMENT

Ministry of Youth Affairs &Sports

Priv

ate

Sec

tor

NG

Os

DT

ET

NA

ITA

Other Ministriesengaged in TEVT

Ministry of Tertiary Education &Vocational Training.

TVEC

SLI

AT

E

VT

A

NY

SC

Gov

ernm

ent

Dep

artm

ents

&S

tatu

tory

Bod

ies

Lanka Institute of AdvancedTechnical Education (SLIATE),Vocational Training Authority ofSri Lanka (VTA), and NationalYouth Services Council(NYSC). In addition, severalother Ministries, governmentdepartments and semi-government institutions alsoconduct sector specific skilldevelopment programs relatingto fisheries, agriculture,textiles/garments, transport,c o n s t r u c t i o n ,telecommunication and variousother manufacturing fields. Themajor public TEVT providersaccount for nearly 85 percent oftraining provided by the statesector:

• DTET delivers formalinstitution-based trainingat 37 technical collegesand affiliated institutionsin major provincial anddistrict capitals. Using thisregional net work, DTEToffers about 60 courses ona full-time and part-timebasis in four areas: a)technicians studies, b) craftstudies, c) business studiesand, d) general studies.The majority of programsare in the management andcommerce (32.5%),building and construction(29.7%) and electrical andelectronics (17.9%) sub-sectors.

• NAITA, the successor tothe NationalApprenticeship Board, hasisland-wide coveragethrough a delivery

network in 9 Provinces. Itoffers four majorvocational trainingprograms:a. enterprise - based

apprenticeships;b. institution-based dual

training;c. in-plant training for

other tertiary levelcourses; and

d. training of trainers.

Together, they account for about140 courses representing 20different trade groups.

• VTA began vocationaltraining in 1995 as thesuccessor to the trainingarm of the Department ofLabor which pioneeredvocational training forout-of school,unemployed youth. Itcaters primarily to ruralunemployed youth,providing employment-oriented short courses at207 Rural VocationalTraining Centers (RVTC)which work closely withrural committees and localorganizations. In addition,14 District VocationalTraining Centers (DVTC),4 National Institutes and 9Special Centers offercertificated courses oflonger duration.

• SLIATE, established in1995, caters to A-levelqualified studentsinterested in TEVT. Itoversees 6 AdvancedTechnical Institutions and

3 Advanced TechnicalCenters, and conductssome higher leveltechnical courses in 10Technical Colleges thatwere previously run byDTET.

• NYSC is another majorpublic TEVT provider butunder the supervision ofthe Ministry of YouthAffairs and Sports(MYAS). The courses atNYSC are offered at twolevels: Level 1 (basiccourses) and Level 2(semi-skilled craftsmancourses). In addition,NYSC also provides jobmarket information andcareer guidance servicesto rural youth through itsRural Centers.

5.5. The non-state TEVTsector, while quite smallinitially, has grown andexpanded into many areasduring the post-liberalizationperiod. Private training institutesare now well established inoccupational areas such asmachining, welding, radiorepair, motor repair, electricalwiring, refrigeration and airconditioning, television,computer and communicationstechnology, tourism and hotelindustry. Many private providersoperate on a fee basis, especiallyin urban centers for book-keeping and accountancy,computer technology and officemanagement. The NGO trainingsector covers many religious andvoluntary organizations that

95

offer craft-level training, fee-free or on a nominal fee basis,targeting unemployed youth,rural women, school leavers,and semi- or unskilled workers.The information base on privateand NGO sector TEVTproviders - their enrollments,regional distribution, courseofferings, and operations - islimited and coverage relativelyincomplete.

Trainee Enrollments5.6. Figure 5.2 presentsinformation on training outputof the major public TEVTproviders from 1990 to 2002.Student intake grew from 32,612to 67,612 between 1990 and2002 representing an annualgrowth of 8.9 percent. Thefigure shows the growingdominance over time of VTA asagainst the declining share ofNAITA in the TEVT sector. The

DTET share appears to havestabilized after a significant dropin 1996 with the takeover ofsome of its training programs bySLIATE. Though not shownhere, in terms of gender, theproportion of femaleenrollments in TEVT courseshas remained relatively constantat about 40 percent, clustered inseveral traditionally femaleoccupations (e.g. dressmaking,ISM operator, beauty culture,secretarial jobs); recently,however, female participationhas increased as new courseswere offered in English forcommerce, industry and furthereducation (69%), accountingtechnicians (65%), computerapplications (73%) andcomputer programmer andoperator (57%).20 In general,female participation in TEVT islow relative to their participationrates in school and highereducation.

96

20. Based on DTET data for 2002.

Table 5.1 TEVT Enrollments by Province - 1990 to 2002

Source: Technical and Vocational Education Commission.

Province 1990 1995 2000 2001 2002

Western 29.23 30.54 30.07 29.86 28.60Southern 16.03 15.82 16.38 8.77 16.71Sabaragamuwa 12.44 11.15 11.65 16.62 11.63North-Western 11.64 10.53 9.66 2.32 9.93Central 11.52 12.06 9.35 8.95 9.31Eastern 9.29 8.91 8.92 9.93 9.16Uva 5.60 6.17 6.87 4.41 7.46North Central 3.96 4.35 4.53 7.55 4.49Northern 0.30 0.48 2.56 11.58 2.71Total 100.00 100.00 100.00 100.00 100.00

Figure 5.2 TEVT Enrollments 1990-2002

0

10000

20000

30000

40000

50000

60000

70000

80000

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

Enr

olm

ents

SLIATE

NYSC

VTA

NAITA

DTET

Source: Technical and Vocational Education Commission.

97

5.7. Table 5.1 reports totalstudent enrollments acrossprovinces by the major publicsector TEVT providers. Thetable reveals a regionalconcentration of TEVTactivities in the Western,Southern, Sabaragamuwa,North-Western and CentralProvinces. The WesternProvince alone accounts forabout 30 percent of enrollmentswhile the other four provincestogether account for more than45 percent of total enrollments.This closely follows the heavyconcentration of regional grossdomestic product (RGDP) in theWestern Province (47 percent)and in the other four provinces(37 percent). Among the majorpublic sector providers, VTAleads in serving rural youthbecause of its regional networkof RVTCs supported by Districtand National Training Centers.In 2002, these RVTCs (19,351trainees) and DVTCs (3,157trainees) accounted for about

one third of all enrollment bymajor public TEVT providers.

5.8. Table 5.2 presents therecurrent and capitalexpenditures by the major publicTEVT institutions in 1983, 1993and 2002. Between 1983 and2002, total expenditures bymajor public sector TEVTproviders increased from Rs.230.4 million to Rs. 1527.3million, representing an averageannual growth of 35 percent innominal terms. At constantprices, however, this reflectsonly a marginal increase of 1.3percent per annum. Even thisincrease is mainly due toadditional investments providedby the Asian Development Bank(ADB) through the SkillsDevelopment Project (SDP). Interms of total governmentexpenditure, the share of majorpublic TEVT providers wasaround 0.31 and 0.36 percent in1983 and 2002, respectively.Another point emerges from

Table 5.2, namely thatallocations for capitalinvestments have declinedsteadily over this period, from45 to 20 percent. Given therelatively high cost ofspecialized buildings, workshopand equipment required forTEVT, this declining share ofcapital budgets has adverseimplications of delayedreplacement and growingobsolescence of trainingequipment.

Evolution of TEVT Institutionsand Training Policies

5.9. The TEVT institutionalframework and training policieshave evolved over the past threedecades. One policy objective,however, has remained constant- government provision of pre-employment technical andvocational training to addressthe problem of high youthunemployment. Over time,however, other policyconsiderations have emerged

Table 5.2 Total Expenditures by Major TEVT Institutions (Rs. Millions)

Sources: Annual Government Budget Estimates and Final Accounts of Statutory BodiesNotes: Recurr. = Recurrent budget ; Capital = Capital budget; SDP = Skills Development Project;

* Apportioned based on ADB project documents

Institution 1983 1993 2002Recurr. Capital Recurr. Capital Recurr. Capital

DTET 43.6 37.7 130.7 262.3 319.0 125.0Department of Labor 21.4 7.1 61.0 16.5 - -NAITA 57.4 55.5 152.3 33.7 217.0 20.0VTA - - - - 208.0 16.0NYSC 4.1 3.6 22.5 - 214.0 4.0TVEC - - 7.0 - 21.0 5.5SLIATE - - - - 101.5 19.3SDP Skills 136.0* 121.0*Total 126.5 103.9 373.5 312.5 1216.5 310.8

with changes in themacroeconomic environment,including meeting the skillneeds of industrial restructuringand competitiveness in a globaleconomy. This section describesthese major changes in thinkingabout the roles and objectives ofTEVT and the institutionalreform that accompanied thesepolicy changes, drawing upon anumber of key policy statementsand documents, as well asinterviews with responsiblegovernment agencies.

5.10. Demand sideconsiderations in promotingTEVT began with the post-1977pro-market reforms andcontinued through the 1980s.The 1977 policy reforms oftrade and industry, and adoptionof export-led growth strategiesled to expansion of thesecondary and tertiary sectors ofthe economy, creating additionaldemand for skilled, semi-skilledand unskilled labor. The stateTEVT sector did not have theresources or the institutionalflexibility to respond, whichcreated opportunities for privatesector providers to enter and fillthis unsatisfied demand fortraining. The private sectortraining role was explicitlyrecognized as part of a largerstrategy of promoting TVET fornational competitiveness in the1989 Industrialization Strategy

of the Ministry of IndustrialDevelopment (MID). Todevelop specific skills andexpertise required by an outwardlooking industrial sector, MIDmade upgrading of technicaltraining institutes andestablishment of a Vocationaland Tertiary EducationCommission (VTEC) prioritiesfor immediate action. It alsostressed the need to expand therole of private industry andNGOs in TEVT development,with industry providing moretraining facilities at lower levelsand voluntary organizationsoffering training in technologyand management skills. Fiscalincentives (for example, doublededuction of trainingexpenditures from income tax)were also considered as possiblepolicy initiatives to promoteindustry's participation intraining provision.21

5.11. The new government thatwas formed in 1994 continuedthis policy in developing theTEVT sector. The PolicyStatement issued by thePresident in January 1995 isinstructive; it stated that“despite its high levels ofliteracy, the Sri Lankanworkforce lacks the requisiteskills that are essential forindustrial upgrading anddiversification. The numerousshort-run skill development

programs and the generaleducation system are essentiallysupply-driven and thereforehave a poor record of providingindustry-relevant skills. Thevocational training system willbe extensively restructured so asto be demand-driven, incooperation with the privatesector who will be the eventualemployers.” Echoing this, the1995 Budget Speechemphasized the importance of“consolidating the variety ofdispersed and uncoordinatedfacilities for vocational trainingand directing them to createskills which are sorely neededby the growing economy”. Tothis end, the NewIndustrialization Strategy ofMID (1995) identified threemajor areas for TEVT reform:(i) restructuring tertiaryeducation and the vocationaltraining system; (ii) setting up aSkills Development Fund; and(iii) reorienting public sectortechnical institutes to meetmarket needs. In October, 1997a special Task Force wasestablished to study inter-ministerial barriers and to comeup with policy recommendationsfor TEVT sector developmentinvolving both the public andprivate sectors. Its findings ledto major changes in the TEVTsector in terms ofrationalization, recognition ofvocational training as a

98

21. References to double deduction tax incentives for training (DDIT) have appeared periodically in Budget speeches over theyears, but it remains unclear whether this fiscal incentive has ever been implemented, and if so, how wide-spread its use is,or whether it has had the desired impact of fostering training in industry. In Malaysia, an assessment of its DDIT indicatedthat the incentive was used principally by large companies and multi-national corporations who were already training; itwas not widely used among local small and medium-size enterprises (SMEs), most of whom provide little in-servicetraining. As such, the GOM eliminated DDIT in 1993 and, in its place, introduced the Human Resource Development Fund.

Ministerial Function, and co-ordination of TEVT activities atthe national level.

5.12. Since 2001, a largenumber of policy documentshave been published,highlighting the challengesfacing the TEVT sector,proposing further reforms to andrationalization of public traininginstitutions, introducingcompetency standards andaccreditation, and offering newfinancial incentives and policyinstruments for fosteringdemand-led training:

• “Vision 2010” (2001)identified three majorchallenges facing theTEVT sector - qualitativeand quantitativemismatches in certainareas of skills demand;external and internalinefficiencies in the sectorwith duplication ofcourses, outdatedequipment and curricula,shortage of good trainersand high dropout rates;and sub-optimal use ofpublic sector workshopsand laboratories. Inresponse to thesechallenges, policies in theTEVT sector should: (i)promote private sector-ledskills training; (ii) targetyouth with anentrepreneurial mind-set;(iii) link performance andbudgets and provide

institutions with increasedautonomy; (iv) fostermore skilled trainingthrough a system of skillaccreditation; (v) provideskills to compete in globallabor markets; and (vi)train to match industry'sskill needs.

• “Regaining Sri Lanka”(2001), published by thenew government, notedthat, “post secondaryskills training institutessuffer from a mismatchbetween the trainingoffered and the skillsrequired in a modern,market economy.Management deficiencies,outdated equipment andcurricula and shortage ofcapable trainers lead tohigh drop out rates andlow returns to training.” Itproposed that (i)performance standards forvocational training berevised along the lines ofc o m p e t e n c y - b a s e dtraining, based onstandards derived fromindustry; and (ii) a HigherInstitute of AppliedTechnology be establishedto lead the vocationaltraining process, andprovide a recognizedsystem of professionalcertification in thevocational trades.

• The “Draft NationalEmployment Policy”

(2002), put out by theMinistry of Employmentand Labor, recommendedsix major policy reformsfor the TEVT sector: (i)training systemsrestructured to meet futuredemands; (ii) promotionof vocational training forthe informal sector; (iii)fostering government-industry partnerships intraining; (iv) thegovernment to function asa facilitator, standardsetter and regulator oftraining; (v) financialincentives for trainingtargeting the corporatesector; and (vi) provisionto disadvantaged groupsof financial assistance fortraining. The BudgetSpeech for 2003 and 2004elaborated on theoperational details ofthese recommendations,including theestablishment of a HumanResources EndowmentFund (HREF) that wouldprovide training vouchersfor unemployed youth andskills upgrading foremployees, as well as softloans to upgrade trainingfacilities for providers.22

• The “National Policy andAction Plan for theDevelopment of TechnicalEducation” (2002)published by the NationalEducation Council (NEC)

99

22. The proposed Human Resources Endowment Fund (HREF) is to be funded from dormant social security accounts. The 2004Budget Speech proposed that funds from the ADB Skills Development Fund also be used to finance HREF. The fundremains to be implemented as the proposed new training initiative has yet to be presented to, and discussed by, parliament.

called for thereorganization andupgrading of the TEVTsector. Seven majorelements are to beconsidered in formulatingthe national policyframework for the TEVTsector: (i) skills ladderingto facilitate upwardmobility of skillsacquisition; (ii) betterlinkages of TEVT with theschool and universitysystems; (iii) privatesector participation intraining; (iv) improvedfunctioning of theTechnical and VocationalEducation Commission(TVEC) as the apex bodyfor TEVT; (v)accreditation of andquality monitoring ofpublic and private traininginstitutions; (vi)rationalization of thepublic sector TEVTsystem; and (vii) settingup a degree awardinginstitution for TEVT.

5.13. The TEVT sector'srestructuring, rationalization andreform process to implement themany proposals introduced sincethe mid-1990s is still a work-in-progress. Interviews withMTEVT, TEVC and individualTEVT institutions indicated thata number of initiatives areunderway. A partial list includes(i) pilots in several provinces to

simplify applications fortraining-a single application forall TEVT institutions-and tocoordinate and rationalizecourse offerings from differentTEVT institutions in the samelocal area; (ii) development ofconcrete action plans in theTVEC to implement broadpolicy directives, especially inthe area of competency basedtraining and skills laddering;(iii) course offerings on a feepaying basis for selected high-demand courses as a way topartially recover costs; and (iv)increased collaboration acrossTEVT institutions, especiallywith NAITA which has stronglinks with employers, to solicitopportunities for practicaltraining from private sectoremployers to enhance theindustrial relevance of training.

Monitoring and Evaluationof Training5.14. TVEC is charged withoverall implementation oftechnical and vocational trainingpolicies, and coordination oftraining in the public and privatesectors. As part of this mandate,it assembles and publishesstatistics on training, a directoryof public and private sectortraining providers, as well asselected labor market data onjob vacancies and job seekers.23

This labor market informationsystem (LMIS) is valuable butits usefulness is limited by its

current availability in hard-copyonly; it remains to be madeavailable electronically topolicymakers and the public.The TVEC data base's coverageof training opportunities in theprivate sector is also limited-registration with TEVC ismandatory but an unknownnumber of private traininginstitutions have reportedly yetto do so. Those that do registerprovide limited information ontheir course offerings, butalmost no hard data onenrollments and theiroperations. As such, monitoringof training incidence and trendsis necessarily limited to thetechnical and vocational trainingprovided by the public sector;virtually nothing is known abouthow widespread training is inthe private sector, either thatprovided by private and NGOsector training providers, or byemployers themselves in theform of in-service training.This paucity of training datalimits the ability of TVEC andpolicymakers to make informedevidence-based policyjudgments about the adequacyof training provided by thepublic and private TEVTsectors, or by employers.

5.15. Compounding dataconstraints is weak public sectorcapacity in impact evaluations,so that training (and other labormarket) interventions to addressproblems of youth

100

23. Until recently, the main source of information was job vacancies culled from newspaper advertisements. JobsNet - set up withinthe past year to receive job applications from job seekers and job vacancies listed by employers, and provide a web-based jobmatching clearing house - is a promising new source of labor market information.

unemployment are rarelyevaluated. The exceptions arethree recent tracer studies ofgraduates from DTET, NAITAand VTA conducted withtechnical assistance and fundingfrom GTZ.24 All three tracerstudies used a unified approach-following graduates 1-3 yearsafter training and administeringa common (and thuscomparable) retrospectivesurvey. The broad outcomecriteria used to assess theexternal efficiency of theseTEVT programs includedemployability after training, andrelevance of training received asper industry's requirements.

5.16. The two outcomemeasures are summarized inTable 5.3 for each program.First, at the time of the survey,employment rates for graduateswere 71.8, 66.6 and 28.6 percentfor NAITA, DTET and VTA,

respectively. Employability ofVTA graduates was especiallylow. In other results not reportedin the table, duration of jobsearch also varied greatlybetween TEVT providers. In thecase of NAITA, for example, 30percent were able to findemployment immediately upontraining completion whileanother 54 percent found jobswithin six months. In contrast,only 2.7 per cent of VTAgraduates and 37.6 per cent ofDTET graduates were able tofind some employment withinthe first six months. Second, forthose finding jobs, graduatesrated highly the relevance oftraining provided by NAITA (73percent “very relevant”), and byVTA (65 percent), as comparedto 40 percent for DTET. At theother extreme, focusing on “Notrelevant at all” responses,NAITA fared best with just 11percent negative responses as

compared to 25-30 percentnegative responses for the othertwo TEVT providers.

5.17. These tracer studiesprovide useful insights but arenonetheless limited. First, asone-off evaluations, tracerstudies are of limited use unlessthey are repeated periodicallyand used to continuouslymonitor whether progress isbeing made to improve trainingoutcomes and remedy identifieddeficiencies in trainingrelevance. A second, and moreserious limitation is the absenceof a control group against whichperformance of graduates can becompared. A control group-comprising individuals withidentical characteristics asTEVT trainees but notparticipating in trainingprograms-is needed to addressthe counterfactual question ofhow trainees would have fared

101

Table 5.3 Principal Outcomes of Tracer Studies of TEVT Graduates

Sources: DTET(2002), NAITA (2001) and VTA (2001)

DTET NAITA VTA

Employment StatusWage employment 49.3 62.2 14.3Self employment 6.1 9.6 14.3Employed in other categories of work 10.7 - -Unemployed 33.9 28.2 62.7Not seeking a job - - 8.7Total 100.0 100.0 100.0

Relevance of Training to the JobVery relevant 39.8 72.7 65.4Relevant 30.4 15.8 8.9Not relevant at all 29.8 11.5 25.7Total 100.0 100.0 100.0

24. The DTET study (1999/2000) was based on a sample of 2,732 graduates taking DTET courses in 1995 and 1996, and drawnfrom across the country except for the Northern Province. The NAITA study (2000) was based on a sample of 706 apprenticesgraduating between 1996 and 1999, drawn from the Western, Central, Southern and North Western Provinces. Similarly, theVTA (2001) study was made up of a sample of 973 trainees in 48 trades drawn from the Central, Uva and Southern Provinces.

had they not participated intraining, and to measure the netcontribution of training to theoutcomes observed. Forexample, if 40 percent of thecontrol group of unemployedfound jobs within 12 months ascompared to 55 percent ofDTET graduates, then just 15percent (55 minus 40 percent),the observed difference inemployment rates, can beattributed to the trainingprovided in DTET. The netimpact of training is the relevantoutcome to consider in judgingthe efficacy and cost-efficiencyof training (and other labormarket) interventions to addressyouth unemployment.25

5.2 Overview of TVETTrends from Labour ForceSurveys5.18. With this overview ofTEVT institutions and policiesas background, the focus of thechapter turns next to an analysisof training trends and theireffects on labor marketoutcomes, especially as theypertain to unemployed youth.

5.19. This section paints a broadbrush overview of trends inpost-school vocational andtechnical training (henceforthreferred to simply as “training”)

as revealed in a time-series ofthe Sri Lanka Labor ForceSurvey (LFS). It addresses thefollowing questions: who getstraining, what types of training,are there wide regionaldifferences in training receipt,does training substitute forformal education or does it tendto complement (go togetherwith) education. It exploits theavailability of a long time-seriesof LFS to explore how trainingvaries over the work life fordifferent cohorts of individuals,focusing on training-experienceprofiles rather than training-ageprofiles as is more common inthe extant Sri Lanka literature,and on the potentially importantlink between training andunemployment.

The Labor Force Surveys

5.20. The LFS has elicitedinformation on trainingconsistently since 1992.Following the questions oneducational attainment,respondents are asked abouttraining: (1) Did he/she havevocational / technical training?(2) What type of training? (notcoded in several years), (3) Wastraining formal (providing adiploma or certificate) orinformal (not certificated)? and(4) What was the duration of

training (months)? Theavailability of training datacovering a 11-year periodbetween 1992-2002 is unique,especially coupled with the richinformation on other individualand household variables, and onlabor force status, employment,unemployment, job search,hours and weeks of work, andmonthly earnings for wage andsalary workers.

5.21. While invaluable for theproposed analysis, nonethelessthe training data have severallimitations. The LFS does notask when the training wasreceived - for some individuals,it may be recent; for others(older individuals) it could havebeen many years ago.26 Thetraining question implicitlyassumes just one training event,while some individuals mayactually have received multipletraining events.27 Neither does itidentify which traininginstitution provided the training,whether or not it was a public orprivate provider, nor whether thetraining was taken whileunemployed or as part ofemployer-provided or financedin-service training. This data setthus precludes an assessment ofthe quality of public sectorversus private sector training, as

102

25. A good review of rigorous evaluations of active labor market programs, including training, in industrialized and developingcountries may be found in Betcherman, Olivas and Dar (2004), “Impacts of Active Labor Market Programs: New Evidencefrom Evaluations with Particular Attention to Developing and Transition Countries”.

26. This is important since training is an investment, and the payoffs to training may occur with a lag and diminish over time.Thus, for recent trainees, the outcomes from training may not be manifested yet, while the depreciation of training taken along time ago may yield few measurable labor market benefits.

27. For example, see Lillard and Tan (1991), who find multiple training events common in their study of private sector trainingin the U.S. using longitudinal surveys of youth, adults and women. Data from a Sri Lanka sample of TEVT trainees in2001/2002 in the Colombo area also revealed that a quarter of them had received some form of training prior to enrollingin the current public TEVT program.

might be done by comparing thereturns to training from thesetwo sources, or an evaluation ofthe potentially importantproductivity and wage impactsof employer-sponsored in-service training reported by agrowing body of studies.28

5.22. What do the LFS time-series data say about trainingtrends? Figure 5.3 shows theweighted proportions of theworking age population thatreported having receivedvocational or technical training,separately for any training in thetop panel and for formal(certificated) training in thebottom panel, and within eachcategory of training by agegroup - youth aged 15-29 years,and adults aged 30-65 years.First, what is readily apparentfrom these data is the secularlyrising trend in training incidencebetween 1992 and 1999, a

stagnation and marked declinein 2001 with negative economicgrowth, and recovery in trainingincidence by 2002.

5.23. Second, training receivedis increasingly more formal overtime. The proportion of theworking-age populationreceiving any training rises from11 to 13 percent over this 1992-2002 period; the proportiongetting formal (i.e., certificated)training rises even moredramatically from 7 to 10percent. Finally, in each year ahigher proportion of youth aged15-29 years reported trainingthan did adults aged 30-65years, and over time, these age-related differences in trainingwidened. Recent entrants intothe labor market are more likelyto get training as compared toentrants from years past, whichmay either reflect an increasedsupply of training seats for

technical and vocationaltraining, an increased deriveddemand for skills fromemployers, or some combinationof the two factors.

Who Gets Training? Snapshotsfrom 1992, 1997 and 2002 LFS

5.24. Table 5.4 provides aprofile of who gets trainingusing three years of LFS data, atthe beginning, middle and endof the 1992-2002 period. Thetop panels of the table report theincidence of any training andformal training by level ofeducational attainment andgender. Several points areimmediately apparent. First,educational attainment andtraining are complements-theincidence of training generallyrises with education, from 0-2percent for those withoutschooling, to 20-40 percent forthose with a university degree.Second, females are less likelyto get training as compared tomales at all levels of education,a point raised in the previoussection. The gender gap intraining is least pronounced athigher levels of educationbeginning with GCE O-levels.Third, there is a rising trendbetween 1992 and 2002 in theincidence of training amongGCE A/L qualified individualsand graduates, for both malesand females. Fourth, among thetrained, the less educatedreceive training that is primarilyinformal (non-certificated)while the more educated tend to

103

Source: World Bank estimates, based on the Department of Census and Statistics, LaborForce Surveys 1992-2002.

.06

.08

.1

.12

.14

1992 1994 1996 1998 2000 2002

trn_all trn-ytrn-a ftrn_all

ftrn-y ftrn-a

Pro

port

ion

Tra

inin

g

year

Formal training

All agesAny training

Youth 15-29

Adults 30-65

28. See, for example, Tan and Batra (1997), Tan (2002). Such an analysis is possible using the forthcoming Sri LankaInvestment Climate Survey which elicited information on in-service training.

Figure 5.3 Proportion of Working Age Populationwith Technical and Vocational Training: 1992-2002

get formal training. Finally,echoing Figure 5.3, it appearsthat the trend over time istowards formal training,especially for males startingwith lower secondary educationand above, but not for females atall levels of education.

5.25. The bottom panel of Table5.4 reports the regionalincidence of vocational andtechnical training. Amongprovinces, the incidence of anytraining is highest in the WesternProvince, at 18 percent for males

in 1992, and in the SouthernProvince at 15 percent, andlowest in the North-CentralProvince at 9 percent and in UvaProvince at 12 percent. By2002, the Western Provincecontinued to enjoy the highestincidence of training. Thistraining gap between high andlow training provinces reducedfor men but widened for womenbetween 1992-2002. Alsonoteworthy is the rise in trainingincidence in the North-CentralProvince, the province with thelowest training incidence in

1992. This finding appears to beconsistent with the point madein Chapter Two about theCentral Government'sprogressive financing ofeducation and training, targetingresources disproportionatelytowards provinces with lowerper capita income.

Who Enrolls in Public SectorTVET and When? A Case Studyof Colombo Trainees

5.26. The LFS does notdistinguish training by public orprivate sector providers. This

104

Sources: World Bank estimates. Calculated from the Labor Force Surveys 1992, 1997 and 2002. Figures are for the population aged 15to 65 years, weighted using Department of Census and Statistics sampling weights. Training incidence in the North-EasternProvince is unavailable as the LFS could not cover the area.

Table 5.4 Overview of Training in Sri Lanka - 1992, 1997, 2002Percent Getting Training by Type of Training, Education, Sex and Province

Education Completed Percent Getting Any Training

No schooling 2.6 3.7 2.4 1.2 0.7 0.7Primary 8.0 6.2 5.5 2.0 1.2 1.5Lower secondary 11.5 10.5 9.9 3.5 2.7 2.3Upper secondary 15.9 14.8 15.8 8.7 6.8 6.8GCE O level 21.0 22.7 21.2 15.9 16.3 13.6GCE A level 29.0 34.0 37.3 27.8 29.3 32.6Graduate 29.9 33.0 39.6 21.9 24.1 31.4Post-graduate 57.5 53.3 46.9 41.8 48.9 46.7

Education Completed Percent Training That is Formal


Province Percent Getting Any Training

Western 17.9 19.7 22.1 9.1 12.1 14.1Central 12.5 14.3 12.4 7.5 8.3 6.8Southern 15.4 14.5 14.1 10.8 7.8 8.6North-West 13.4 11.8 12.6 7.3 5.8 6.2North-Central 9.2 11.3 13.6 4.2 4.5 7.4Uva 12.3 10.6 13.2 7.3 6.9 9.1Sabaragamuwa 13.7 12.8 13.6 8.6 7.9 9.7

Education / Province Males Females1992 1997 2002 1992 1997 2002

raises the following questions:Who are public TVETinstitutions serving? Are youthproceeding directly to TVETafter schooling completion, or isit more typical for youth to firsttest the labor market, that is,engage in job search oremployment, before enrolling inTVET? To gain insights into theattributes of the youthpopulation undertakingtechnical and vocational trainingin public TVET institutions, theapplication forms of a sample of846 trainees enrolled in theColombo-area TVET instituteswere computerized and analyzedas a case study.29

5.27. The sample of 846trainees was drawn from threeTVET institutions - theTechnical College at Maradana(235), NAITA Regional Officein Colombo (500), and VTADistrict Office in Colombo (111)- and together they covered abroad range of training - 52programs in total, 14 fromDTET, 28 from NAITA, and 10from VTA. The programs fromDTET and VTA representinstitution-based training whilethose from NAITA focus onenterprise-based apprenticeshipprograms.30 Consistentinformation was collected ongender, age, district of origin,educational attainment, numberof attempts at national GCE

O/A-level examinations, streamstudied in (science, arts orcommerce), year of schoolcompletion, prior training if anyand whether provided by publicor non-state providers, andcurrent course of study.

5.28. Tabulations of enrollmentrecords suggest the followinghighlights about the attributes oftrainees in Colombo-basedTVET institutions:

• The majority (96 percent)of trainees are youth lessthan 30 years old, andthey are predominantlymale (80 percent).

• Their educationalqualifications have thefollowing distribution - 21percent with grade 10 orless, 45 percent with GCEO-level, 14 percent thatfailed A-levels, and theremaining 20 percent withGCE A-levelqualifications. Among theA-level qualified, traineescoming from Commerceand Arts streams weremost common (9 and 8percent, respectively) andthen Sciences (2 percent).

• The number of repeatGCE exam takers (up to 3tries) is high - 7 percentamong the O-levelqualified, and a very high44 percent among those

seeking A-levelqualifications when A-level failures are included.The implication is thatrepeated exam taking toqualify for admission touniversity is delayinglabor market entry formany youth.

• Over a quarter (28percent) of the samplehave had some form oftraining prior to currentenrollment in the publicTVET institution.31 Alsosignificant, given thepolicy focus on publicTEVT, is that 55 percentof those with priortraining got their trainingfrom private sectorproviders, 36 percent frompublic institutions, and 9percent from unknownsources.

• Most enrollments (58percent) occur within thefirst 1-4 years afterschooling completion, 18percent after 5-6 years, 19percent after 7-9 years,and just 5 percent after 10or more years in the labormarket.

Training Over the Work Life

5.29. What do training profileslook like as individualscomplete formal schooling andacquire work experience in the

105

29. None of the training institutions had computerized individual records for their trainees; when they were computerized, itwas only in the form of aggregated totals.

30. Within the 52 training programs selected for study, information on all student enrollments is included to preclude anypotential selection biases (whether by age, sex, education, or work experience) in the choice of the trainee sample.

31. This further reinforces the point that the LFS is too restrictive in eliciting information on just one episode of training, whenmultiple training events-even among youth-are relatively common.

Sources: World Bank estimates. Calculated from the Labor Force Surveys 1992, 1997 and 2002. Figures are for the population aged 15to 65 years, weighted using Department of Census and Statistics sampling weights.

Education Completed Percent Getting Any Training


Education level Youth Age 15-29 Adults Age 30-651992 1997 2002 1992 1997 2002

Table 5.5. Percent Getting Training by Age Group and Education - 1992, 1997, 2002

labor market? The sample ofpublic TEVT trainees suggeststhat this profile is steeply risinginitially and then flattening outwith age. Table 5.5 reportstraining data from the LFS byeducational attainmentseparately for two broad agegroups-youth aged 15-29 years,and adults aged 30 years andabove-for insights into thesetraining-age profiles.

5.30. Two points stand out.First, among youth, there is anincrease in the incidence oftraining for those with GCE A-levels and above, but not forthose with GCE O-levels andbelow. For adults aged 30-65years, the only group to show arising trend in training areuniversity graduates. Similarage-related differences intraining - but across alleducation groups - were shownin Figure 5.3. Second, at each

level of education, a roughlyequal or higher proportion ofadults report having training ascompared to similarly educatedyouth among GCE A/L qualifiedindividuals and graduates. Thisis consistent with theprobabilistic interpretation thatpeople tend to acquire trainingas they become older (though ata slower pace), and thateducated individuals are morelikely to get training as they age,compared to the less educated.32

However, this pattern does notappear among postgraduateeducated individuals.

5.31. From this intuition, onemight contemplate using cross-sectional information onindividuals varying in age oryears of potential workexperience (defined as ageminus age at schoolingcompletion) to estimate trainingexperience profiles for different

education groups. This exercise,however, is potentiallymisleading if cohort effects areimportant. Essentially, these areeffects associated with whenindividuals entered the labormarket. For example, youthtoday face different trainingopportunities as compared tosimilarly educated youth(today's adults) who entered thelabor market decades earlierwhen post-school traininginfrastructure was lessdeveloped. The twodemographic groups would alsohave been exposed to differentsets of macroeconomicconditions and policies thatinfluence both the incentives aswell as the opportunities to gettraining. Training-experienceprofiles estimated from cross-sectional surveys potentiallyconfound the life-cycle profilesof interest with year-of-entrycohort effects.33

106

32. These training-age profiles can also be given a human capital interpretation. With a finite work-life, human capital theorywould predict that individuals would concentrate their training (and education) investments early in their lives so as tomaximize the present value of the future stream of returns to their investments, reducing (the probability of) investments intraining later in their work-life as the period to recoup returns shortens.

33. Indeed, given the rising training trend among higher educated youth in the 1990s, training-experience profiles computedfrom cross-sectional surveys (i.e. from different year entry cohorts) would be downward sloping instead of flat or upwardsloping as required in a cumulative function.

5.32. An alternative approach isto estimate training-experienceprofiles from pseudo-cohortsconstructed from the time-seriesof LFS spanning the 1992-2002period. The intuition of thepseudo-cohort approach isstraightforward. Consider agroup of individuals aged 18-19years with a GCE O-leveleducation in the 1992 LFS; onaverage, they might be expectedto have 2 years of potential workexperience. In the 1994 LFS,this cohort would be age 20-21with 4 years of potentialexperience, age 22-23 years inthe 1996 LFS with 6 years ofpotential experience, and so on.34

By grouping individuals andtracking similar groups ofindividuals over time using thetime-series LFS, data on thesepseudo-cohorts can be used toestimate the probability oftraining with potential workexperience, controlling forgender, level of educationalattainment, and year-of-entrycohort effects. The regressionestimates suggest the followingmain findings:

• On average, males have ahigher probability ofgetting training over theirwork life as compared tosimilarly educatedfemales. This greaterlikelihood varies by levelof educational attainment-a gender gap of about 17

percent for those withprimary education, risingto 41 percent for lowersecondary, and thendeclining to 30, 20 and 4percent for grades 9-10,O-levels and A-levels,before widening again to10 percent for universitygraduates.

• Tr a i n i n g - e x p e r i e n c eprofiles vary dramaticallyby level of educationalattainment.35 For thosewith primary education,there are no discernibleeffects of increased timein the labor market on theprobability of gettingtraining, that is, training-experience profiles areflat. For lower secondary,training profiles rise (butat a diminishing rate) withyears of potentialexperience, becomingincreasingly steeper athigher levels ofeducation-about 13-15percent greater probabilityof training per year ofpotential work experiencefor grades 9-10, GCE Oand A levels, and 10percent for universitygraduates.

Probability ofUnemployment, PotentialExperience and Training

5.33. These grouped cohort datacan also be used to describe therelationship between trainingand a number of labor marketoutcomes. One outcome ofparticular policy interest isyouth unemployment,specifically the issue of whetherreceipt of technical andvocational training facilitatesyouth transition from school towork, and from unemploymentto productive work. Just aspseudo-cohorts can be trackedover time to look at training-experience profiles, so too cantheir unemploymentprobabilities with years ofpotential work experience beanalyzed, free of theconfounding year-of-entrycohort effects.

5.34. Table 5.6 reports theregression estimates for theprobability of being unemployedlast week, and its relationship toyears of potential workexperience, separately for thosewith and without training. Inaddition to experience, theregression model includedcontrol variables for gender,level of education, and differentyear-of-entry cohorts. Severalfindings emerge:

• Males are significantlyless likely to experience aspell of unemploymentlast week as compared tofemales with similar

107

34. In essence, this pseudo-cohort approach involves collapsing the 1992-2002 LFS survey data into cells cross-classified byyear, sex, level of educational attainment, 2-year age intervals, and training receipt, with information on the weighted countof individuals in each cell.

35. This finding is based on training probability models estimated on the grouped data separately by level of education. Theregression results are not reported here but are available from the World Bank.

educational, experienceand year-of-entry cohortattributes. This gendergap in unemploymentprobability is 44 percentfor the group withtraining, and about 37percent for those withouttraining.

• For the group withtraining, the level ofeducational attainment isnot significantly related tounemployment, asurprising finding but onethat may be attributed tothe strong association

between training andeducation (onedeterminant of training),an issue analyzed ingreater depth in the nextsection. In contrast, forthe group withouttraining, level ofeducation matters and theprobability ofunemployment rises witheducation level to a peakfor those with GCEA-levels.

• The probability ofexperiencing anunemployment spell

declines (at a decreasingrate) with years ofpotential experience.Especially significant isthe markedly steeperunemployment -experience profile of thegroup with training - eachadditional year ofpotential experiencereduces their probabilityof an unemployment spellby over 19 percent ascompared to just under 15percent for the groupwithout training.

• This steeper decline in

108

Table 5.6 Training and Probability of Unemployment Last WeekUsing Pseudo-Cohort Data LFS 1992-2002

Source: World Bank estimates. Calculated from the Labor Force Surveys 1992, 1997 and 2002. Figures are for the population aged 15to 65 years, weighted using Department of Census and Statistics sampling weights.

Male -0.4415 -17.22 -0.3728 -27.17

Level of educationPrimary 0.1124 0.16 -0.1113 -1.73Lower secondary -0.2166 -0.32 0.0760 1.20Upper secondary -0.1498 -0.22 0.3526 5.68GCE O-Level -0.0904 -0.13 0.4440 7.09GCE A-Level -0.1443 -0.21 0.4665 7.28Degree -0.1434 -0.21 0.2852 3.11Post-graduate 0.0979 0.10 0.8098 1.41

Potential experience (pexp) -0.1935 -12.06 -0.1489 -21.31

Pexp-2 0.0050 4.75 0.0019 4.45

Pexp-3 -0.00004 -2.15 0.000003 0.40

Year of entry cohort1950s -0.6845 -1.19 0.0532 0.331960s -0.9704 -1.51 -0.1919 -1.031970s -1.1874 -1.79 -0.4279 -2.171980s -1.5056 -2.28 -0.7637 -3.861990-94 -1.7345 -2.63 -0.9784 -4.921995-99 -1.7564 -2.64 -1.1836 -5.912000-02 -1.8134 -2.66 -1.3189 -6.25

Constant 2.6137 2.73 1.0518 5.04

Number of observations 1081 2009

Adjusted R-squared 0.5963 0.8007

Dependent variable: Training Status

Unemployed last week (1,0) With training No training

Coef. t-stat Coef. t-stat

u n e m p l o y m e n t -experience profiles for thegroup with training ascompared to the non-training group continuesto hold within educationgroups. This comparisonis confirmed by regressionmodels estimatedseparately by level ofeducation (not reportedhere but available fromthe author).

• Year-of-entry cohorteffects are negative andincreasingly morenegative for recent entrycohorts, which isconsistent with theinterpretation that 1990sentry cohorts experienceda more favorable set ofmacroeconomic factorsand policies than theirpeers in the entry cohortsof the 1960s and 1970s.

5.3 Determinants ofTraining5.35. With this broad overviewof training trends and thetraining-unemployment nexusas background, this section turnsfirst to a microeconomicanalysis of the determinants oftraining, with subsequentsections focusing on the labormarket outcomes of training.The analysis - based on pooledcross-section time-series LFSdata - does not account forcohort effects, but makes up forit by exploiting rich information

in the LFS to gain insight intothe important individual,household and aggregate factorsthat shape decisions ofindividuals to invest in post-school training, and their labormarket outcomes.

5.36. The individual decision toinvest in training is modeledwithin a human capitalframework and is thought todepend on several factors:

a. level of schoolingattainment - training andeducation are eithersubstitutes for each other(because trainingcompensates forinadequate quantity orquality of schooling), orare complements (themore educated train morebecause they are alsomore efficient learnersand gain more than theless educated frominvesting in training);

b. age - individuals are morelikely to invest early intraining, rather than laterin the life-cycle, so thatthe returns to training canbe realized over a longerperiod of time;

c. marital status - familialresponsibilities reflectedin an individual's maritalstatus may influencetraining decisions;

d. household factors - just asparental education affectsschooling of children,

they may affect trainingdecisions too eitherthrough sociologicalinfluences (taste foreducation), or throughincome effects associatedwith higher educationalattainment;

e. location - provinces differin their levels of per capitaincomes and employmentopportunities, and thesefactors can shapeindividual trainingdecisions;

f. access to training -training may also dependon the supply of TVETopportunities, and thisvaries across provincesand over time; a proxy fortraining supply is thenumber of public TVETenrollments per 1,000 ofthe provincial youthpopulation.36

5.37. Table 5.7 reports theresults of estimating theprobability of training - anytraining, and formal(certificated) training - on thisset of individual, household,location and training supplyvariables. Probit models areestimated separately for thesample of males (about 190,000)and females (197,000). Theregression results are notsurprising, and are broadlyconsistent with the tabularinformation presented earlier inSection 5.1 of this chapter.

109

36. Training seats by province and year are calculated from TEVT enrollments in Table 5.1 and from counts of youth aged 15-29 years by province computed from the time-series of LFS surveys.

5.38. First, the probability ofany training, and especially offormal training, is stronglycorrelated with level ofschooling attainment. Togetherwith the effects of age, thiseducation - trainingcomplementarity would explainwhy training-profiles tend to bemore steeply inclined upwardsfor the more educatedindividuals and relatively flat forthose with little formal

education. Second, theprobability of getting training inthe Western Province (theomitted category) is higher thanin all other provinces, with theNorth-Central Province havingthe lowest probability of trainingonce all other control variablesare included. Third, the effectsof training supply are negative,which suggests counter-intuitively that the probability oftraining is lower the more

“training seats” are available inthe province. One possibleexplanation is that this resultmay simply reflect the centralgovernment's targeting of TEVTresources at provinces with lowtraining incidence. Fourth,parental schooling positivelyinfluences children's trainingdecisions; interestingly,compared to father's education,mother's education is morestrongly associated with training

110

Dependent variable: Males Females

Training Any training Formal training Any training Formal training

Coef. z-stat Coef. z-stat Coef. z-stat Coef. z-stat

Table 5.7 Probability of any Training and Formal Training: 1992-2002

Notes: [1] Training supply proxy = total number of TVET graduates by province divided by total youth population (x1000) aged 15-29years. [2] Household variables: dummy variables included for missing father’s and mother's years of schooling.

Source: World Bank estimates. Calculated from the Labor Force Surveys 1992, 1997 and 2002. Figures are for the population aged 15to 65 years, weighted using Department of Census and Statistics sampling weights.

Marital status -0.0364 -3.8 -0.0810 -7.1 -0.1019 -10.9 -0.1331 -12.4

Education levelPrimary 0.3664 10.7 0.3019 5.5 0.1825 4.8 0.5813 7.1Lower secondary 0.6921 20.6 0.7598 14.2 0.5000 13.7 0.9364 11.7Upper secondary 0.8990 27.0 1.1264 21.3 0.9527 27.2 1.5026 19.2GCE O-Level 1.0903 32.6 1.4598 27.5 1.3191 37.7 1.9384 24.7GCE A-Level 1.3315 38.6 1.7597 32.8 1.7365 48.5 2.4000 30.5Degree 1.3265 31.4 1.8274 31.0 1.6091 35.9 2.3059 27.7Post-graduate 1.8490 31.4 2.3629 32.9 2.0228 29.4 2.7207 27.7

Age 0.0909 46.5 0.0865 37.5 0.0703 30.2 0.0678 26.2

Age-squared -0.0011 -46.6 -0.0011 -37.3 -0.0009 -28.9 -0.0008 -24.4

ProvinceCentral -0.1328 -8.1 -0.0812 -4.3 -0.0358 -1.9 0.0298 1.4South -0.0231 -1.6 -0.0615 -3.6 0.0180 1.1 -0.0030 -0.2North-West -0.1703 -11.9 -0.1786 -10.7 -0.1816 -10.3 -0.1419 -7.4North-Central -0.2499 -11.9 -0.2946 -11.8 -0.2096 -8.1 -0.1794 -6.3Uva -0.1116 -7.0 -0.0737 -4.0 0.0970 5.3 0.1110 5.4Sabaragamuwa -0.0832 -6.4 -0.0640 -4.3 0.0235 1.5 0.0574 3.4Training seats [1] -0.0175 -5.5 -0.0157 -4.3 -0.0148 -3.9 -0.0143 -3.5

Household [2]Father’s schooling 0.0185 8.8 0.0199 8.5 0.0112 4.4 0.0128 4.7Mother’s schooling 0.0281 10.1 0.0300 9.4 0.0110 3.3 0.0139 3.8

Year (linear) 0.0060 2.7 0.0133 5.2 -0.0038 -1.4 -0.0011 -0.4

Constant -15.1954 -3.5 -30.3148 -6.0 4.6327 0.9 -1.3909 -0.2

Number of observations 190,148 190,148 197,541 197,541

Pseudo R-square 0.0737 0.1284 0.1548 0.1905

of sons, but neutral with respectto the training of daughters.37

Finally, consistent with thetrends noted earlier, there is asecularly rising trend over timein the training of males,especially formal training, butno apparent time trend intraining for females.

5.4 Training and LaborMarket Outcomes5.39. What are the labor marketoutcomes of these investmentsin post-school training? Chiefamong the labor marketoutcomes of policy interest areunemployment, job search andearnings. Specifically,

policymakers are keenlyinterested in the effects oftraining on (i) the probability ofexperiencing an unemploymentspell last week; (ii) length oftime spent in job search, or theschool-to-work transition; and(iii) monthly earnings, and howthe returns to training compareto those for formal schooling.

111

37. To see this, note that the marginal impact of mother's education on training of males (sons) is 0.028 while that of father'seducation is 0.018; for females (daughters) , the corresponding marginal impact is 0.011, identical for mother's and father'seducation. For some individuals-renters, relatives, households without parents-education of parents is not known, andmissing value indicators are included to take this into account.

Source: World Bank estimates. Calculated from the Labor Force Surveys 1992, 1997 and 2002.Note: Model estimated by probit regressions, including dummy variables for LFS years.

Table 5.8 Probability of Unemployment Last Week and Usually Unemployed Last Year, 1992-2002

Male -0.366 -33.0 -0.365 -32.9 -0.379 -34.1 -0.378 -34.0

Education levelPrimary 0.137 2.5 0.138 2.6 0.036 0.8 0.038 0.9Lower secondary 0.246 4.6 0.253 4.8 0.038 0.9 0.043 1.0Upper secondary 0.328 6.2 0.346 6.6 0.105 2.5 0.118 2.8GCE O-Level 0.463 8.7 0.495 9.4 0.187 4.3 0.213 4.9GCE A-Level 0.470 8.8 0.518 9.7 0.208 4.7 0.250 5.7Degree 0.186 2.9 0.223 3.5 -0.124 -2.1 -0.084 -1.5Post-graduate -0.453 -3.3 -0.396 -2.9 -0.844 -5.8 -0.785 -5.4

Potential experience -0.116 -33.9 -0.113 -33.3 -0.136 -39.8 -0.134 -39.3

Potexp-squared 0.003 13.7 0.002 13.2 0.003 19.3 0.003 18.8

Potexp-cubed 0.000 -7.2 0.000 -6.8 0.000 -11.7 0.000 -11.3

Formal training 0.146 10.0 0.091 6.0

Informal training -0.012 -0.4 -0.060 -2.2

For. training duration 0.007 0.9 -0.020 -2.2

Infor. training duration -0.040 -1.8 -0.061 -2.8

Urban -0.024 -1.4 -0.024 -1.5 -0.021 -1.2 -0.020 -1.2

ProvinceCentral 0.161 9.0 0.159 8.9 0.115 6.4 0.114 6.3Southern 0.293 16.6 0.287 16.2 0.273 15.3 0.268 15.0North-West -0.034 -1.6 -0.042 -2.0 -0.004 -0.2 -0.010 -0.5North-Central -0.112 -4.4 -0.120 -4.7 0.043 1.8 0.035 1.5Uva -0.108 -3.9 -0.109 -4.0 -0.116 -4.2 -0.117 -4.3Sabaragamuwa 0.156 7.3 0.152 7.2 0.126 5.9 0.123 5.7Share of public jobs 0.990 5.6 0.997 5.6 0.780 4.4 0.785 4.5Constant -0.011 -0.2 -0.012 -0.2 0.406 6.9 0.401 6.8

Sample size 151,763 151,763 145,110 145,110

Pseudo R-squared 0.2623 0.2613 0.257 0.256

Dependent variable : Unemployed Last Week Usually Unemployed Past YearProb(Unemployment) Coef. z Coef. z Coef. Z Coef. Z

1992 2002 1992 2002

Also of interest in each of theseareas is whether the labormarket effects of training varyby type of training-formal orinformal-and duration oftraining.

Impact of Training onProbability of Unemployment

5.40. The analysis of theprobability of unemploymentuses two alternative measures ofunemployment defined by theLFS. The first one-termed“unemployed”-is based onwhether individuals areavailable and/or actively lookingfor work, and who did no workduring the past week. Thesecond measure-termed “usuallyunemployed”-refers to a personwho was looking or available forwork for the major part of thepast year. An enumeration ofwork and availability forwork/job search for each of thepast 12 months is used toclassify individuals as usuallyemployed, usually unemployedor usually not economicallyactive. The second measure-byaveraging over a 12 monthperiod-may provide a morerobust characterization of workstatus as compared to theconventional unemploymentmeasure, which is a snapshot intime.

5.41. A probit model is used torelate unemploymentprobability to a set of individualand location attributes,including level of educational

attainment, a flexible (cubic)measure of years of potentialwork experience since schoolingcompletion, receipt of trainingmeasured either by formal andinformal training indicatorvariables or by training duration,location in an urban area and byprovince, and year dummyvariables to capture seculartrends in unemployment levels.In addition, the model includes avariable representing theproportion of all jobs that are inthe public sector, which variesboth across provinces and overtime.38 This measure allows us toaddress the hypothesis thatunemployment rates are highbecause individuals queue up forpublic sector employment whichpays not only higher wages butalso better benefits. Theregression results are reported inTable 5.8 above, separately foreach of the two measures ofunemployment.

5.42. The results are broadlysimilar irrespective of whetherthe reference period is last weekor last year. First, males aresignificantly less likely (byabout 37 percent) to experienceunemployment as compared tofemales. Second, compared tothose without schooling (theomitted group), the probabilityof unemployment appears to risewith level of schoolingattainment, peaking at GCE A-levels, and falling for degreegraduates and post-graduates.Third, probabilities of

unemployment decline (at adecreasing rate) with years ofpotential work experience,which might be expected iflonger search time yields betterinformation about jobopportunities and an exit fromunemployment. Fourth, urbanlocation is not statisticallyimportant but there are largedifferences across provinces.Compared to the WesternProvince, unemploymentprobabilities are higher in theCentral, Sabaragamuwa andSouthern provinces, and lowerin the North-Central, North-Western and Uva Provinces.Finally, there is a strong positiveassociation betweenunemployment probabilities andthe share of public sector jobs inthe province. It is unclear,however, which way thecausality runs.

5.43. The effects of post-schooltraining on unemploymentprobabilities are mixed. When(0,1) indicator variables fortraining are included, the effectsof formal training onunemployment are invariablypositive and statisticallysignificant, while those forinformal training are negative,though significant only in the“usually unemployed”regressions. When training ismeasured using duration oftraining, which distinguishesbetween incidental andsubstantive kinds of training,both formal and informal

112

38. The LFS identifies jobs as being public or private sector. The share of public sector jobs can thus be computed by dividingpublic sector jobs by job totals, separately by province and year.

training are negative andstatistically significant in the“usually unemployed”regressions. These mixed resultsmay arise because the modelconstrains the unemployment-experience profiles to be thesame for everyone, so thattraining can only affect overalllevels of unemployment but notunemployment experienceprofiles. This is clearly not thecase, as was suggested by theearlier analysis using pseudo-cohort data. There, Table 5.6showed that training isassociated with steeperunemployment-exper ienceprofiles, implying thatunemployment probabilities fallmore rapidly with potentialexperience for individuals withtraining than for those withouttraining.

Impact of Training on School toWork Transition

5.44. A second area of policyconcern is the long time manyyouth spend in job searchbetween leaving school andfinding employment. Forexample, according to the 2002LFS, almost 85 percent of youthaged 15-29 years that arecurrently unemployed report notever having a job. This figurerises from about 75 percent forlower secondary passes to

almost 95 percent for universitygraduates. While these figureshighlight the seriousness of thisissue, they can be misleading:they mix more and less educatedyouth with different years ofpotential work experience andthus time spent in job search,and they ignore the fact thatother youth are employed, andmay have found employmentsoon after schooling completion.

5.45. In this section,information from differentquestions in the LFS areassembled, and the pooled LFSdata from 1996-2002 reshaped,to provide more balancedinsights into the school-to-worktransitions of youth withdifferent levels of schoolingattainment, and to address thequestion of whether technicaland vocational trainingfacilitates this transition. TheLFS, in its current form, is notwell designed to study school towork transitions. So thechallenge here is to determinethe date of first recordedemployment39 for eachindividual with a given level ofeducation, from which the timetaken from schoolingcompletion to first employmentcan be calculated. Beginning in1996, the LFS asked theemployed how long they have

been on their current job, so thestart date of that job can beascertained.40 For theunemployed, the LFS askedwhether they have ever had a joband, if so, how long it had beensince the previous job.41 If theprior job is assumed to be ofsimilar duration as those held bytheir currently employed peers(about 2 years), then thisinformation and the interveningunemployment spell can be usedto determine the start date of theprevious job. For those whohave never had a job, theduration of search for a first jobis still ongoing (or censored).Finally, an adjustment is made tosearch time for those withtechnical and vocationaltraining-time spent in training issubtracted to reflect theirwithdrawal from active jobsearch while undergoingtraining.

5.46. These calculation weredone for a sample of 39,000individuals from the 1996-2002 LFS, restricted to thosewith some schooling up touniversity graduates, and with0-10 years of potential workexperience so as to keep thefocus on youth. Figure 5.4presents graphically theresulting distributions of time-to-employment for different

113

39. We caution that the first “recorded” employment is not necessarily the first job; some individuals may have had several jobsprior to the recorded job, so time-to-first “recorded” job may overstate duration of job search. But it is the only employmentspell for which information is available. It may be desirable to revise the LFS to elicit explicitly year of labor market entryand year of first job.

40. The 1996 LFS also started eliciting detailed information on years of schooling from which more precise schoolingcompletion dates can be calculated.

41. The intervening unemployment spell is reported in several intervals, ranging from several months to an open ended 5 ormore years. Some assumptions are needed to impute duration (in years) of unemployment to these categories.

levels of schooling.42 Severalpoints emerge from this figure.First, those with the lowestlevels of schooling attainment-primary and lower secondarypasses-are more likely to faceprotracted job search beforesecuring their first employment.Their distributions of time-to-first-employment areconcentrated around 4-7 yearsafter schooling completion.Second, most of those withupper secondary schooling andwith GCE O-level or A-levelqualifications find their first jobfairly soon after schoolingcompletion. Their distributionsof time-to-first-job areconcentrated around 0-4 years,

tapering off with time in thelabor market. Finally, theschool-to-work transition ofdegree graduates appears toresemble more that of youthwith lower secondary schoolingthan that of the GCE A-levelqualified. The distribution oftime-to-employment is bi-modal- some find a job within the firstyear, while many others appearto take longer, about 3-5 yearsafter graduation fromuniversity.43

5.47. These figures do notcontrol for other factors thatmay also shape school-to-worktransitions, such as gender,household characteristics,location and receipt of post-

school technical and vocationaltraining. The joint effects ofschooling attainment and theseother factors on time-to-employment can be studiedwithin a regression frameworkthat accounts explicitly for thefact that one part of the sampleis still actively searching for thefirst job, that is, with incomplete(censored) time-to-employment.44 Table 5.9 reportsthe results of estimating thisregression model for the sampleof youth as a whole, andseparately by training status, soas to investigate how receipt oftechnical and vocational trainingaffects school-to-worktransitions.

5.48. Table 5.9 makes thefollowing points. Compared toyouth with primary schooling,more educated groups findemployment after schoolcompletion much faster though,as suggested by Figure 5.4,degree graduates are more likethose with lower secondary than(say) those with GCE O or A-level qualifications. Genderdifferences are important, andmales find employment fasterthan females; a contributingfactor to this gender gap may bemarital status, since marriage is

114

42. Note that these graphs understate time-to-first-job because they include ongoing (censored) job search for the sample ofunemployed youth that have still not found employment at the time of LFS enumeration.

43. These distributions of time-to-employment appear to tell a consistent story about how low levels of schooling attainmentdisadvantage youth in their job search while higher level school qualifications facilitate the school-to-work transition. But the jobsearch of degree graduates stands apart - many experience fairly long job search before finding employment. Are the durations ofjob search reasonable? The mean time-to-employment is fairly long, and may suggest that individuals are not reporting prioremployment accurately, for example, ignoring casual work and reporting only formal jobs; another contributing factor is theassumption that the first “recorded” employment is also the first job, which would tend to overstate time-to-first job.

44. Survival models are ideally suited for studying the determinants of time to a failure event, in this case time taken to find a job afterschooling completion, and accommodating censored spells of job search. Such models may be fit using alternative distributionalassumptions about the underlying process, but the one used here is the lognormal distribution.

Graph by Edlvl

0

5

10

15

20

0

5

10

15

20

0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10

Primary LowerSec UpperSec

GCE O/L GCE A/L Degree

Per

cent

Fin

ding

Firs

t Job

Years to First Job

Figure 5.4 Time-To-First-Job by Level of Schooling

associated with delayed time-to-employment. Location alsomatters: job search is longer inurban areas, and relative to theWestern Province, also longer inthe Central, South andSabaragamuwa Provinces.Finally, year dummy variables(not reported here) indicate thatoverall length of job search hasdeclined over time in parallelwith falling unemploymentrates.

5.49. What about the effects oftraining? The first column ofTable 5.9 indicates that formaland informal training are bothassociated with shorter job

search time, with informaltraining appearing to have alarger impact (-0.10) than formalor certificated training (-0.07).The second and third columns,reporting results estimatedseparately by training status,make the additional point thatwhile having more educationreduces time to employment forboth groups, the impact ofeducation is more pronouncedfor the group with training. Tosee this, compare the relativecontributions of different levelsof education to shortening time-to-employment - in the groupwithout training, this peaks withthe upper secondary group; in

contrast, in the group withtraining the contribution riseslinearly with level of educationto a peak at degree level. Inother words, education andtraining interact positively toreduce time spent in job search.

Impact of Training on MonthlyWages

5.50. The final outcomeconsidered is the impact oftechnical and vocational trainingon monthly wages and salaries.The LFS only elicits incomeinformation for the sample ofindividuals in wage and salaryemployment, so this analysis isrestricted to the sample of about

115

Source: World Bank estimates. Based on the Department of Census and Statistics, Labor Force Survey.Notes : The regressions are estimated by maximum likelihood using a parametric survival-time model fit with a lognormal distribution.

About one-quarter of the sample were censored. The regression model included control variables for parental education and forLFS years.

Table 5.9 School to Work Transitions with and without Training

Dependent variable: All Youth Without Training With TrainingTime-to-Employment Coefficient z-stat Coefficient z-stat Coefficient z-statLower secondary -0.329 -15.6 -0.340 -15.8 -0.166 -1.7

Upper secondary -0.471 -24.2 -0.492 -24.7 -0.294 -3.1

GCE O-Level -0.434 -21.0 -0.448 -20.9 -0.284 -3.0

GCE A-Level -0.454 -20.8 -0.445 -19.4 -0.350 -3.6

Degree -0.340 -10.8 -0.276 -8.1 -0.459 -4.2

Formal training -0.069 -6.0

Informal training -0.106 -5.2

Male -0.070 -8.1 -0.069 -7.3 -0.077 -4.0

Married 0.113 9.6 0.136 10.4 0.028 1.0

Urban 0.030 2.9 0.049 4.1 -0.040 -1.8

ProvinceCentral 0.027 2.1 0.017 1.2 0.059 2.1South 0.177 13.1 0.169 11.1 0.190 6.5North-West -0.029 -2.0 -0.032 -2.1 -0.036 -1.0North-Central -0.072 -4.4 -0.098 -5.5 0.026 0.7Uva -0.048 -2.9 -0.086 -4.6 0.111 3.0Sabaragamuwa 0.114 7.9 0.099 6.1 0.161 5.1

Constant 1.964 54.0 1.979 50.2 1.771 14.7

Sample size 33,206 26,274 6,932

Number finding jobs 24,605 19,678 4,927

29,000 employees in the pooled1992-2002 LFS time-series.45

Monthly wages are deflated bythe consumer price index into2002 real wages, and expressedin logarithmic terms as iscommon in the earningsfunction literature.46

5.51. The wage analysisaddresses several questions:What are the returns to technicaland vocational training, and howdo they compare to the returns togeneral education? For thiscomparison, training duration inmonths is converted into year-equivalents like years ofschooling.47 Are the returns toformal certificated trainingsimilar to or greater than thoseto informal training? Thisquestion is particularly pertinentgiven the trend towards moreformalized forms of training inrecent years. And finally, howare the different kinds oftraining rewarded in the publicsector versus the private sector?The private sector may demand(reward more highly) differentkinds of technical andvocational skills, includingspecific skills, that may notnecessarily be certificated, atleast by public sector providers.To address this latter question,wage regressions are estimatedseparately for public and privatesector employees.

5.52. Table 5.10 reports theresults of wage regressionsincluding control variables forgender, a quadratic measure ofyears of potential workexperience, principal industry ofemployment, public sector job(in some specifications), urbanlocation, province, and yeardummy variables to capture timetrends in overall real wageincreases.

The results in Table 5.10 suggestthe following findings:

• The returns to anadditional year of generalschooling (9.0 percent)are higher than a year ofany technical orvocational training (6.6percent). One caveat isthat participation intraining programs(especially informaltraining) may not be on afull-time basis, whichwould tend to under-statethe returns to training.

• Overall, the returns toformal certificatedtraining (7.0 percent) areusually higher than thoseto informal training (4.4percent). This maypartially reflect the factthat most informaltraining is taken by thosewith the lowest levels ofschooling attainment,

while the more educatedtend to favor formaltraining.

• However, the rewards toformal and informaltraining variesdramatically acrosssectors. In public sectoremployment, formaltraining is highly valued(returns of 7.6 percent)but informal training isactually penalized (-38percent), perhaps becauseit is associated withmenial jobs by those withlow levels of education.In contrast, the privatesector appears to valueformal and informaltraining equally.Comparing returns,informal kinds of trainingmay actually have theedge (7.1 percent) overformal training (6.7percent), though thedifferences may not bestatistically significant.

• Several other findings arenoteworthy. First, thereare wage premiumsassociated withemployment in urbanareas (15 percent) and inthe public sector (9percent). Second, wagelevels are higher in theWestern Province (theomitted category) than in

116

45. The wage and salary information used in this analysis is not adjusted for weeks or hours worked, and does not include smallpayments in kind.

46. Within a human capital model framework, the log (wage) specification allows the coefficient estimates of years of educationand training to be interpreted as the rate of return to investments.

47. Unlike general education, which is full-time, some training programs may be taken on a part-time basis, which would tendto overstate the time spent in training and thus under-estimate the returns to training.

the rest of the country.Finally, though notreported in the table, thereis a secular rise in realwages over time between1992 and 1998, with someslippage in real wagegains thereafter.

5.5 Conclusions andImplications for the Future5.53. The general education andTEVT sectors are intimatelylinked with respect to the labormarket. The TEVT sectorintermediates between the skillsthat the general educationalsector provides school leavers

and the skills that the labormarket demands, providingtechnical and vocational trainingas needed to facilitate the schoolto work transition for youth. Asthe preceding analysis showed,the TEVT sector has not alwaysaddressed this skills demand-supply gap well or in acoordinated fashion. Butpolicymakers have recognizedthe problem, and have over timeproposed numerous initiatives torationalize the TEVT sector,address regional inequities intraining, and better coordinatethe TEVT policy framework;this reform process is still awork-in-progress. Monitoring

and evaluation of traininginstitutions and their courseofferings is nascent, and limitedby incomplete and poor data,especially about training byprivate sector training providersand employers, and by theabsence of an evaluation culture.The sector's limited interactionswith the general educationsector also precludes it fromhaving a more active role inproviding schools and guidancecounselors with informationabout training and labor marketemployment opportunities.

5.54. The preceding analysisalso highlighted the potential ofthe Sri Lanka Labor Force

117

Note: Regression models are estimated by ordinary least squares (OLS) methods. They include dummy variables for principal industrygroups, and year dummy variables. The full set of coefficient estimates are available from the author.

Table 5.10 Returns to Training For Wage and Salary Employees by Sector

Dependent variable: All Sectors Public Sector Private SectorLog(monthly wage) Coef. t-stat Coef. t-stat Coef. t-stat Coef. t-stat

Training

Years-Any training 0.0664 6.8 n.a. n.a. n.a.

Years-Formal training 0.0704 6.7 0.0766 4.7 0.0668 5.0

Years-Informal training 0.0444 1.9 -0.3770 -3.6 0.0707 2.9

Other attributes

Years of schooling 0.0904 26.4 0.0901 26.2 0.0832 11.4 0.0923 23.4

Male 0.0993 6.3 0.1000 6.3 0.1983 6.1 0.0610 3.3

Potential Experience 0.1059 9.5 0.1058 9.5 0.1597 5.4 0.1027 8.4

Experience-squared -0.0043 -4.6 -0.0043 -4.6 -0.0075 -3.3 -0.0042 -4.0

Province

Central -0.1367 -6.1 -0.1365 -6.1 -0.1264 -2.8 -0.1317 -5.2South -0.1826 -7.6 -0.1821 -7.6 -0.1125 -2.3 -0.1972 -7.1North-West -0.0656 -3.0 -0.0652 -2.9 0.0226 0.5 -0.0886 -3.5North-Central -0.0816 -2.4 -0.0806 -2.4 0.0322 0.6 -0.1378 -3.2Uva -0.2079 -6.1 -0.2073 -6.0 -0.2006 -3.2 -0.2135 -5.2Sabaragamuwa -0.2487 -10.1 -0.2482 -10.1 -0.1286 -2.5 -0.2833 -10.1

Urban location 0.1515 8.0 0.1513 8.0 0.0509 1.3 0.1768 8.3

Public sector job 0.0909 3.2 0.0903 3.2 n.a. n.a.

Constant 6.3251 110.1 6.3275 110.1 6.1459 41.9 6.3439 98.3

Number observations 29102 29102 6,742 2,360

Adjusted R-square 0.1001 0.1001 0.0755 0.0910

Survey to provide policy-relevant insights into trainingtrends, and the links betweengeneral education, post-schooltraining, and labor marketoutcomes. It showed thatschooling and training arecomplements, namely, that moreeducated school leavers receivemore technical and vocationaltraining. It highlighted thesignificant impact of training onreducing the probability ofunemployment, and infacilitating the school-to-worktransition. And it demonstratedthat investments in botheducation and training yieldrelatively high returns in theform of higher earnings. Butmany issues of policy interestcould not be addressed, oraddressed well, with the existingLFS, including the relativebenefits of training by publicversus private sector providers,the adequacy of in-service skillsdevelopment by employers, orrigorous panel analysis of labormarket behavior around youthunemployment.

5.55. The ability ofpolicymakers to monitor andevaluate the performance of theTEVT sector, or more broadly,to pilot and rigorously evaluate arange of labor marketinterventions (includingtraining) to address the problemsof youth unemployment andnational competitiveness, willrequire improvements to its datacollection system. Somesuggestions follow:

• Computerize all studentenrollment and programrecords for all TEVTinstitutions to facilitatetracking and evaluatingthe performance oftrainees, for use in tracerstudies as part of acontinuous M&E system.

• Enforce the mandatoryreporting requirements ofall private sector andNGO training providers togive the TVEC andpolicymakers a morecomplete view of the totalTEVT sector.

• Include questions onemployer-provided ororganized training in theinformation to beenumerated in allestablishment surveys andcensus conducted by theDepartment of Census andStatistics, as many otherindustrialized anddeveloping countrystatistical agencies havedone.

• Improve the trainingquestions in the LFSinstrument bydistinguishing furtherbetween public sector andprivate sector provision oftraining, whether trainedwhile unemployed or in-service, allowing for morethan one training event,and dating receipt oftraining event(s). Otherrelatively simple changesto date schoolingcompletion, labor market

entry and key eventswould greatly improveanalysis of school-to-work transitions.

• Employ a rotating panelsampling approach in theLFS so individuals can betracked over 12 months tostudy how behavior andoutcomes change overtime. This approach isincreasingly being used inboth industrialized anddeveloping countries, andcan readily beaccommodated within theexisting LFS samplingframework. This wouldfacilitate study of school-to-work transitions andtraining outcomes.Furthermore, the rotatingpanel feature will allowresearchers andpolicymakers to createcomparison groups for useas control groups in theevaluation of trainingprograms and other labormarket interventions.

• Improve LFS survey &training questions.

118

119

Abhayadeva, C. (2003), Early Childhood Care and Development in Sri Lanka, National EducationCommission (NEC), Sri Lanka.

Alailima, Patricia (1997), Social Policy in Sri Lanka, in Lakshman, W.D. (ed), Dilemmas ofDevelopment, Fifty Years of Economic Change in Sri Lanka, Sri Lanka Association of Economists,Colombo.

Alailima, Patricia (2002), Employment and Unemployment of Youth in Sri Lanka, Country Study,Department of National Planning, Colombo, Sri Lanka.

ANTRIEP (2000), Improving School Efficiency: The Asian Experience. Paris: UNESCO/IIEP.

Aturupane, Harsha (1993), Is Education Beneficial ? A Microeconomic Analysis of the Impact ofEducation on the Economic Welfare of a Developing Country, Sri Lanka. Ph.D thesis (unpublished),University of Cambridge.

Aturupane, Harsha (1994), Poverty, Human Development and Growth: An Emerging Consensus ?American Economic Review, May 1994 (Papers and Proceedings), 84(2), pp. 244-249.

Aturupane, Harsha (1996), "Is Education More Profitable for Women ? An Economic Analysis of theImpact of Schooling on the Earnings of Men and Women in Sri Lanka". Sri Lanka Journal of SocialSciences, Vol:19(1), pp. 27-46.

Aturupane, Harsha (1997), Unemployment Among Educated Women, Ministry of Finance andPlanning, Colombo.

Aturupane, Harsha (1998), Dimensions of Poverty: The Magnitude of Poverty in the Poorer Areas ofSri Lanka, with Special Emphasis on Female Headed Households, in Upanathi, Vol. 9.

Aturupane, Harsha (1999a), Econometric Analysis and Planning in the Social Sectors: An Applicationto Education in Sri Lanka, in Quantitative Tools in Economic Planning: Applications and Issues inAsia. United Nations Centre for Regional Development, Nagoya, Japan.

Aturupane, Harsha (1999b), Poverty in Sri Lanka: Achievements, Issues and Challenges. SynthesisReport for the Sri Lanka Poverty Reduction Framework.

Aturupane, Harsha and Abeygunewardene, B. (2000), The Cost and Financing of Primary Education,in Angela W. Little (ed) Primary Education Reforms in Sri Lanka, EPD, MEHE, Colombo, Sri Lanka.

Betcherman, Gordon, Karina Olivas and Amit Dar (2004), "Impacts of Active Labor MarketPrograms: New Evidence from Evaluations with Particular Attention to Developing and TransitionCountries", The World Bank, Washington, DC.

Bray, Mark (1987), "School Clusters in the Third World: Making Them Work". Digest 23.Paris:UNICEF.

Bibliography

120

Cambridge Education Consultants (2002), Evaluating Teacher Responsiveness to In-Service Trainingin Primary Mathematics, DFID/CEC, Colombo, Sri Lanka.

Carron, Gabriel and Ngoc Chau, Ta (1996), The Quality of Primary Schools in Different DevelopmentContexts. Paris:UNESCO Publishing/IIEP.

Central Bank of Sri Lanka, Annual Reports, various issues, Colombo, Sri Lanka.

Craig, Helen and Perraton, Hilary (2003), "Open and Distance Education for Teacher's ContinuingProfessional Development". In Teacher Education Through Open and Distance Learning. Ed.Bernadette Robinson and Colin Latchem. London: Routledge/Farmer.

Craig, Helen; Kraft, Richard and du Plessis, Joy (1997), Teacher Development: Making an Impact.Washington DC: USAID/World Bank.

Datt, Gaurav and Gunewardena, Dileni (1997), Some Aspects of Poverty in Sri Lanka, World BankPolicy Research Working Paper No. 1738, March 1997.

David, B.A. (2003), Implementation of Continuing Teacher Education Programmes, NationalEducation Commission, Colombo, Sri Lanka.

De Silva, Amala (2003a), Child Malnutrition and Water Borne Diseases in Sri Lanka. Processed.Colombo, Sri Lanka.

De Silva, C. (2003b), School Based Management, National Education Commission, Colombo, SriLanka.

De Silva, E.J. (2003b), Management of Education, National Education Commission, Colombo, SriLanka.

De Silva, W.A. (2003b), The Implementation of Teacher Education Programmes, National EducationCommission, Colombo, Sri Lanka.

Department of Census and Statistics, Demographic and Health Surveys, various years, Colombo.

Department of Census and Statistics, Household Income and Expenditure Surveys, various years,Colombo.

Department of Census and Statistics, Sri Lanka Labor Force Survey, various years, Colombo.

Department of National Planning (2001), Vision 2010, Colombo.

Department of Technical Education and Training, DTET (2002), "Tracer Study of Technical CollegeGraduates 1995-1996", conducted in collaboration with the Department of Census and Statistics andNational Education Research and Evaluation Centre, Faculty of Education, University of Colombo.

Dharmawardana, K. et. al. (1997), A Study on the Influence of Socio-Economic and Other Factors onPupils Who Failed in All Four Subjects at the GCE AL Examination Held in 1995.

Dharmawardane, K. et. al. (1999), The Effects of the Implementation of the Competency BasedPrimary Curriculum Introduced to Grade 1 Classes in the Gampaha District, NEC/NIE, Colombo, SriLanka.

121

Dickens, William and Lang, Kevin (1996), "An Analysis of the Nature of Unemployment in SriLanka", Journal of Development Studies, 31(4), pp. 620-36, April.

Evans, David and Liyanage, Sisira (1995), Managerial Behaviour and School Effectiveness, AnInternational Perspective, NIE, Sri Lanka.

Fernando, R. (2003), Study of the Implementation of Activity Based Oral English in Primary Grades1 & 2, National Education Commission, Colombo, Sri Lanka.

Finance Commission (2002), An Evaluation of the Implementation of the Norm Based Unit CostResource Allocation Mechanism. Colombo, Sri Lanka.

Fiske, Edward B. (1996), Decentralization of Education, Politics and Consensus, The World Bank,Washington, D.C.

Glewwe, P. (1985), An Analysis of Income Distribution and Labor Markets in Sri Lanka. Ph.D Thesis(Unpublished), Stanford University.

Glewwe, P. (1999), The Economics of School Quality Investments in Developing Countries. St.Martins Press, New York.

Glewwe, P. (2002), "Schools and Skills in Developing Countries: Education Policies andSocioeconomic Outcomes", Journal of Economic Literature, Vol. XL, pp. 436 - 482.

Glewwe, Paul (1987), "Unemployment in Developing Countries: Economist's Models in Light ofEvidence from Sri Lanka", International Economic Journal, 1(4), pp. 1-17, winter.

Government of Sri Lanka - Asian Development Bank (1996), Social Sector Financing in Sri Lanka.

Government of Sri Lanka - Asian Development Bank (2000a), Advancing Knowledge and Skills forDevelopment and Competitiveness: Tertiary Education Strategy. Colombo, Sri Lanka.

Government of Sri Lanka - Asian Development Bank, (2000), Improving Education Planning.Colombo, Sri Lanka.

Government of Sri Lanka, Budget Estimates, various years.

Greaney, V. and Kellaghan, Thomas (1996), Monitoring the Learning Outcomes of EducationSystems. The World Bank, Washington D.C, U.S.A.

Gunawardhane, R. (2002), Evaluation of the Implementation of School Based Assessment in PostPrimary Classes, National Education Research and Evaluation Centre, University of Colombo, SriLanka.

Gunawardhane, R. (2003), A Study on the Grade V Scholarship Examination, National EducationCommission, Colombo, Sri Lanka.

Gunewardena, C. (2002), Evaluation of the Effectiveness of the Implementation of Reforms inPrimary Education, National Education Research and Evaluation Centre, University of Colombo, SriLanka.

Gunewardena, C. and Lekamge, D. (2003), An Evaluation of the Implementation of the JuniorSecondary School Curriculum, National Education Commission, Colombo, Sri Lanka.

122

Gunewardena, Dileni (2000), Consumption Poverty in Sri Lanka, 1985-1996: A Profile of PovertyBased on Household Survey Data, Processed, University of Peradeniya.

Heneveld, Ward and Craig, Helen (1996), Schools Count: World Bank Project Designs and the Qualityof Primary Education in Sub-Saharan Africa. Technical Paper No. 303. Africa Technical Department.Washington DC: World Bank.

International Labor Organization (2001), Towards an Employment Strategy Framework, CountryProfile: Sri Lanka, Geneva.

Jayasena, A. (2003), The Teacher Education Programmes, National Colleges of Education, Sri Lanka,National Education Research and Evaluation Centre, University of Colombo, Sri Lanka.

Jayasuriya, J.E. (1969), Education in Ceylon, Before and After Independence, Wesley Press, Colombo.

Jayaweera, S. (1998), "Education in Sri Lanka- Fifty Years Since Independence", in A.D.V. de S.Indraratne (ed.) Fifty Years of Sri Lanka's Independence, A Socio Economic Review, Sri LankaInstitute of Social and Economic Studies, Colombo, Sri Lanka.

Jayaweera, S. and Sanmugam (2002), Graduate Employment in Sri Lanka in the 1990's. Centre forWomen's Research, Colombo.

Kakwani, N. (1993), Performance in Living Standards: an International Comparison, Journal ofDevelopment Economics, Vol.41(2), pp. 287-319.

Karunaratne, S. (2003), Impact Evaluation of 1997 Education Reforms in Teaching Science at theGCE Advance level, NEC, National Education Research and Evaluation Centre, University ofColombo, Sri Lanka.

Kellaghan, T. and Greaney, Vincent (2001), Using Assessment to Improve the Quality of Education.UNECSO, Paris, France.

Knamiller, Gary and Fairhurst, G. (Eds.) (1998) The Effectiveness of Teacher Resource CentreStrategy. Department of International Development, UK: University of Leeds.

Korale, R.B.M. (2003), Provision and Participation in Primary and Secondary Education, NationalEducation Research and Evaluation Centre, University of Colombo, Sri Lanka.

Kotelawala, E. (2003), Teacher Education Programmes in Universities. National EducationCommission, Colombo, Sri Lanka.

Kularatne, N.G. (2003), An Evaluation of the Divisional Schools Development Project and theProgramme to Revamp Central Schools, National Education Commission, Colombo, Sri Lanka.

LeBlanc, Wendy (1997), "Outreach Programs for Learners and Teachers of English In-service: AnEgyptian Model". Cairo.

Leelaratne, P.M. (2001), "The Relationship between Education, Training and Employment", Colombo,mimeo.

Lillard, Lee and Tan, Hong (1991), "Private Sector Training: Who Gets Its and What are its Effects?",in Ronald Ehrenberg (ed), Research in Labor Economics, JAI Press, Volume 13.

123

Liyanage, K. (2003), Educational Technology in the Next Decade, National Education Commission,Colombo, Sri Lanka.

Lockheed, Marlaine and Verspoor, Adriaan and Associates (1991), Improving Primary Education inDeveloping Countries. UK:Oxford University Press.

Marga Institute (1991), Education-Employment Linkages, Colombo, Sri Lanka.

Mendis, P. (2003), Children Who Have Disability (Special/Inclusive Education), National EducationCommission, Colombo, Sri Lanka.

Mettananda, D.S. (2002), Evaluation of Continuing Teacher Education Programmes Conducted inTeachers' Centres, Teacher Education and Teacher Deployment Project, Colombo, Sri Lanka.

Ministry of Economic Policy Reforms (2002), Regaining Sri Lanka: The Future, Colombo.

Ministry of Education, National Evaluation and Testing Service, Statistical Bulletin, various years.

Ministry of Education, School Census, various years.

Ministry of Employment and Labor (2002), Draft National Employment Policy for Sri Lanka: Fromskills-employment-empowerment-and beyond, Colombo.

Ministry of Planning and Employment (1971), The Five Year Plan 1972-1976, Government of Ceylon.

Ministry of Social Welfare (2003), Draft National Policy on early Childhood Development.

Murphy, P., Vincent Greaney, Marlaine Lockheed and Carlos Rojas (1996), National Assessments:Testing the System. The World Bank, Washington D.C., U.S.A.

National Apprentice and Industrial Training Authority (2001), Report on Tracer Study Survey Resultsfor Monitoring and Evaluation of Vocational Training, Planning and Information Division, NAITA,Colombo.

National Education Commission (1997), Reforms in General Education. Colombo, Sri Lanka.

National Education Commission (2002), National Policy and Action Plan for the Development ofTechnical Education in Sri Lanka, September, Colombo.

National Education Commission (2003), Envisioning Education for Human Development, Proposalsfor a National Framework on General Education in Sri Lanka. Colombo, Sri Lanka.

National Education Research and Evaluation Center (NEREC) (2002), Evaluation of the Effectivenessof the Implementation of Reforms in Primary Education. Colombo, Sri Lanka.

National Education Research and Evaluation Center (NEREC) (2003), National Assessment of Grade4 Learning Outcomes. Colombo, Sri Lanka.

National Education Research and Evaluation Center (NEREC) (2004), A Qualitative Study of SchoolManagement Practices. Colombo, Sri Lanka.

National Education Research and Evaluation Center (NEREC) (2004), A Qualitative Study of CivicCompetencies of Children in Sri Lanka. Colombo, Sri Lanka.

124

National Institute of Education, Sri Lanka (1997), School Based Management and ProfessionalAccountability, A Workshop Report. Colombo, Sri Lanka.

O'Grady, Barbara (1996), District Resource Teachers Program Lesotho: An interview with GerardMathot, Former Coordinator. Washington DC: Academy for Educational Development.

Perera, D.A. (2003), Implementation of Non-Formal Education in Sri Lanka, National EducationCommission, Colombo, Sri Lanka.

Perera L., Wijetunga, S. and Balasooriya, A.S. (2003) (MIMEO), Curricular Change and SocialCohesion in Conflict Affected Societies: Case of Sri Lanka.

Perera, L.P., Gunewardena, C. and Wijetunga, S. (2003), Evaluation of the Effectiveness of theImplementation of Education Reforms at Secondary School Level,(Grades 6-11). National EducationResearch and Evaluation Centre, University of Colombo, Sri Lanka.

Perera, S. (2003b), Sri Lanka Education System as a Reflection of Society's Ruptures: A Conceptualand Theoretical Exploration. National Education Commission, Colombo, Sri Lanka.

Perera, W.J. (1999), The Status of Head Teachers - Sri Lanka Case. NIE, Colombo, Sri Lanka.

Perera, W.S. (2003a), Implementation of Compulsory Education, National Education Commission,Colombo, Sri Lanka.

Presidential Task Force (1998), Reforms in University Education. Colombo, Sri Lanka.

Presidential Task Force on General Education (1997), General Education Reforms, Sri Lanka.Colombo Sri Lanka.

Presidential Task Force on TEVT Reforms, Ministry of Vocational Training and Rural Industries(1997), Technical Education and Vocational Training Reforms, Policies, Strategies and ActionProgram, Colombo, Sri Lanka.

Rajapaksa, Lalani (2003), Child Under-Nutrition in Sri Lanka: A Literature Survey. Processed.Colombo, Sri Lanka.

Rama, Martin (1994), "Flexibility in Sri Lanka's Labor Market", Policy Research Working Paper,1262, The World Bank, Washington, DC.

Rodrigo, C.; Korale, R.B.M. and Aturupane, D.H.C. (1987), "Employment and Manpower Planningin Sri Lanka", in Amjad, Rashid (ed.) Human Resource Planning: The Asian Experience; New Delhi:ILO, 257-284.

Rosenholtz, Susan (1991), Teachers' Workplace: The Social Organization of Schools. New York:Teachers College Press.

Rupasinghe, S. (2002), Deployment of Teachers in Difficult Areas, National Education Research andEvaluation Centre, University of Colombo, Sri Lanka.

Sahn, D.E. and Alderman, H. (1987), " The Effects of Human Capital on Wages, and the Determinantsof Labor Supply in a Developing Country " Journal of Development Economics, Vol.29(2), pp 157-183.

Scheerens, J. (2000), Improving School Effectiveness. UNESCO, Paris, France.

Schiefelbein, Ernesto (1992), Redefining Basic Education in Latin America: Lessons to be Learnedfrom the Colombian Escuela Nueva. Paris: IIEP.

Sri Lanka Association for the Advancement of Education (SLAAED) (2000), Education Reforms inSri Lanka. Colombo, Sri Lanka.

Sumathipala, K.H.M. (1968), History of Education in Ceylon: 1796-1965, Ceylon Historical Journal,Vol. 13. Colombo, Sri Lanka.

Tan, Hong (2002), Malaysia: Technological Change and Skill Needs in Industry, Report to theEconomic Planning Unit, PM's Department, Malaysia, by The World Bank, Washington, DC.

Tan, Hong and Batra, Geeta (1997), Enterprise Training in Developing Countries: Incidence,Productivity Effects and Policy Implications, PSD Department Monograph, The World Bank,Washington, DC.

UNDP (1993), Human Development Report. New York, U.S.A.

UNDP (1998), Sri Lanka: National Human Development Report, UNDP, Colombo, Sri Lanka.

UNDP (2003), Human Development Report. New York, U.S.A.

UNICEF (2003), Pre-School Education in Sri Lanka. Colombo, Sri Lanka.

UNICEF-World Bank (2003), Needs Assessment of Education Reconstruction in the Conflict AffectedAreas of Sri Lanka. Colombo, Sri Lanka.

University Grants Commission (2000), Sri Lanka Universities Year Book: 1999/2000, Sri Lanka.

Vocational Training Authority (2001), "Report on Tracer Study Survey Results for Monitoring andEvaluation of Vocational Training", VTA, Colombo.

Wijeratne, K., Cumaranatunga, L. and Perera, I. (2003), Evaluation of GCE A/L, General English,National Education Commission, Colombo, Sri Lanka.

Wijesuriya, G. (2003), The General Certificate of Education, Ordinary & Advance Level Examination,National Education Commission, Colombo, Sri Lanka.

Wijetunga, S. and Rupasinghe, S. (2003), Situational Analysis of General Education, SeniorSecondary School Curriculum (Grades 10-13), National Education Commission, Colombo, Sri Lanka.

Wijetunga, S. and Wickremaratne, V. (2003), Pre-School Survey - Ministry of Social Welfare andUNICEF. Colombo, Sri Lanka.

World Bank (1999), Sri Lanka: A Fresh Look at Unemployment, Report No. 19609-CE, PREM Unit,South Asia Region, August, Washington, DC.

World Bank (2000), Higher Education in Developing Countries, Peril and Promise. Washington D.C.,U.S.A.

World Bank (2003), World Development Indicators. Washington D.C., U.S.A.

World Bank (2004), Making Services Work for Poor People, World Development Report, WashingtonD.C., U.S.A.

125

TRAINING AND LINKS TO THE LABOR MARKETsiteresources.worldbank.org/INTSRILANKA/Resources/... ·...

Documents

Transcript of TRAINING AND LINKS TO THE LABOR MARKETsiteresources.worldbank.org/INTSRILANKA/Resources/... ·...