Female Labour Supply in Australia and Japan: The Effects ...€¦ · AQF qualificationsaligned as...

233AUSTRALIAN JOURNAL OF LABOUR ECONOMICS

Volume 17 • Number 3 • 2014 • pp 233 - 255

Female Labour Supply in Australia and Japan: The Effects of Education and Qualifications

Tomoko Kishi, NanzanUniversity

Abstract This paper compares the effects of educational attainment on women’s employment status in Australia and Japan from 2005 to 2009. Our data are from the Household Income and Labour Dynamics in Australia Survey and the Japanese Panel Survey on Consumers. Using both static and dynamic models to estimate female labour supply, we find robust positive effects of education on employment in Australia, but not in Japan, where the effects of higher education on permanent employment are weaker. The results of the dynamic estimation suggest that the effect of previous employment status is more significant than that of education in Japan, unlike in Australia. This suggests that Japanese firms value previous experience more than educational attainment. Vocational education has a significant effect in Australia, but not in Japan. This result suggests that the skill acquired from education, in particular, vocational education, is not fully utilised in Japan.

Keywords:Femalelabourforceparticipation,Vocationaleducation,Longitudinaldata

JELClassification:J210,J220,J240

1. Introduction TheobjectiveofthisstudyistocomparefemalelaboursupplyinAustraliawiththatinJapanusinglongitudinaldata.Inparticular,wefocusontherelationshipsbetweeneducation, vocational qualifications, and employment status of women in bothcountries.Asthetwocountriesdifferinvocationaltrainingandemploymentsystems,thecomparisonenablesus to identifyproblemsconcerningskill formationand jobcreationintheJapaneselabourmarket.

Address for correspondence: Professor Tomoko Kishi, Department of Economics, NanzanUniversity,18Yamazato-cho,Showa-ku,Nagoya466-8673,Japan.Email:[email protected]: The author thanks the Melbourne Institute of Applied Economic andSocialResearchandtheInstituteforResearchonHouseholdEconomicsforallowingaccesstothelongitudinaldata(theHouseholdIncomeandLabourDynamicsinAustraliaSurveyandtheJapanese Panel Data on Consumers, respectively). The author is grateful to Professor ShigekiKanoforhisusefulsuggestionsregardingeconometricmethods.TheauthorisalsogratefulforthecommentsofparticipantsinaseminarattheMelbourneInstituteofAppliedEconomicandSocialResearchon29April2014.Allremainingerrorsaretheauthor’s.©TheCentreforLabourMarketResearch,2014

234AUSTRALIAN JOURNAL OF LABOUR ECONOMICSVOLUME 17 • NUMBER 3 • 2014

Japan’slabourmarkethasbeencharacterisedbylowemployeemobilityandhighlyorganisedinternallabourmarkets,andemploymentpracticesarecloselyrelatedto the seniority-based wage system. Firms provide intense on-the-job training tonewly-hiredemployees,enablingthemtodeveloptheskillsneededtoperformawidevarietyofjobs.Therearesystematicjobrotationswithinfirms,allowingemployeestograduallyaccumulateknowledgeandskillsandbepromotedtorolesinvolvinghigherresponsibility.Jobtrainingwithinfirmsisregardedasmoreimportantthantrainingoutsidefirms(Koike,1991;KoikeandInoki,2003).

However, since theearly1990s, the Japanese labourmarkethasundergonefundamentalchangesamidaperiodofprolongedeconomicstagnation.Thechangesincludeincreasednumbersoffemaleemployeesandpart-timeorcontractemployees,growing competitiveness, and the emergence of a performance-based pay system.The portion of people who spend their entire career working for a single firmis declining.As a result, themajority ofworking people no longer experience thetraditionalprocessesofskillformationoperatingwithinJapanesefirms.Undersuchcircumstances, providing the full range of workers with opportunities for careeradvancementrequiresestablishingasystemforevaluatingvocationalskillsacquiredoutsidefirms(JILPT,2008).

The effect of vocational skills on individuals’ careers is becoming anincreasingly important research question. As such, this study analyses the effectsof formal vocational education on female employment in contemporary Japan.Weperformaninternationalcomparisontoevaluateourresultsobjectively.Australiaisconsideredtobeausefulcomparativecase,asithasawell-establishedqualificationsystem,includingtheAustralianQualificationsFramework(AQF)andtheAustralianApprenticeship.

Therestofthepapercompareslinkagesbetweeneducationandfemalelaboursupply in Australia and Japan. Section 2 provides an overview of female labourforce participation and education in both countries. Section 3 explains the currentsituation of vocational education in Japan and compares it with that of Australia.Section4reviewstherelevantliterature.Section5explainsthelongitudinaldata,theeconometricmodel,andthevariablesusedinthecomparativeanalysis.Sections6and7interprettheestimatedresultsandconcludethepaper.

2. An overview of labour force and education in Australia and Japan Figure1shows the trends inbothmaleand female labour forceparticipation ratesinAustraliaandJapansince1980.Thetrendsinmalelabourforceparticipationarelargelysimilarbetweenthetwocountries,showingconsistentdeclinessincetheearly1980s.Incontrast,thefemalelabourforceparticipationtrendsaremarkedlydifferent.InAustralia,femalelabourforceparticipationincreasedfromlessthan50percentin1980tonearly60percentin2010,whileinJapan,ithasremainednearlyconstantat50percent,withaslightdeclineinrecentyears.

235TOMOKO KISHI

Female Labour Supply in Austral ia and Japan: The Effects of Education and Qualif ications

Figure 1 - Australian and Japanese labour force participation rates for men and women

Sources: AustralianBureauofStatistics(2014);StatisticsBureauofJapan(2014).

3. Vocational education in Japan TheoutlineoftheeducationsystemofJapanasof2014isasshowninfigure2.Sixyearsof elementary schooling are followedby threeyearsof lower secondary andthree years of upper secondary education. Both elementary and lower secondaryeducation are compulsory. Post-secondary academic education is offered by juniorcolleges,whichprovidetwo-yeareducation,anduniversities,whichprovidefour-yeareducation. Junior colleges, usually female-dominated, teach both liberal arts andpracticaleducationbasedonliberalartswithinashorterperiodthandouniversities.Institutions for vocational education are shown by the shaded areas in figure 2. Itsuggests that the vocational education system is not well coordinated as multipleinstitutionscoexist,insteadofformingahierarchy.

The vocational education system of Japan has developed in a way quitedifferentfromthatofAustralia.BeforeWorldWarII,educationinJapanwasbasedona‘doubletrack’systemconsistingofthevocationaltrackandtheacademictrack,whichwerenotlinkedatall.AfterWorldWarII,thesystemwaschangedtoa‘singletrack’system,whereacademiceducationhadahigherstatusthanvocationaleducation(Goodman, Hatakenaka and Kim, 2009). A number of vocational schools were‘promoted’touniversitieswithmultipledisciplinessoonafterthewar.Thegovernmentlaidmuchstressonthedevelopmentofuniversityeducation,whilereducingsubsidiesforvocationalschools.Atthesametime,thedemandforhigheracademiceducationincreasedwiththeriseinpercapitaincomes.Asaresult,thenumberofjuniorcollegesanduniversitieswentupsignificantly.

The post-war educational policy was closely related with the Japanesemanagementsystem.Theexpansionofheavyindustrieshasbeenaccompaniedbythedevelopmentofinternallabourmarkets.JapanesefirmsgaveintensiveOJTstonewlyhiredemployeessothattheywouldbeabletoperformawiderangeofjobswithinafirm.Astheemployees’averageyearsofservicewerelongandwagedeterminationwasleft

90.0

80.0

30.0

20.0

60.0

40.0

1980

1982

1986

1988

1992

1994

1996

1998

2000

2002

2004

2012

70.0

50.0

1984

1990

2006

2008

2010

AUS_MaleAUS_Female

JPN_MaleJPN_Female


towagenegotiationsinindividualfirms,firmsvaluedtheexperiencethatemployeesgainedwithinthefirmmorethantheemployees’experiencepriortorecruitment.Firmsbelieved that ‘qualifications themselves are not significant signals of the quality oflabourforwhite-collaremployees’inJapan(KoikeandInoki,2003,Chapter3).

That government stressed academic education does notmean that itmadelightofvocationaleducation;itestablishedcollegesoftechnologyin1962tonurtureengineerstomeetdemandfromindustries.Thesecollegeshaveprovidedstudentswhofinishedlowersecondaryeducationwithfive-yeareducationonspecialisedsubjectssuchasmechanicalengineering,electricalengineering,informationtechnology,andarchitecture,aswellasgeneralsubjects.However,thenumberofcollegesoftechnologyhasbeenmuchsmallerthanthatofuniversities.

The ‘vacuum’ invocational education led to anemergenceof anumberofmiscellaneousvocationalschools,mostofwhichwereprivateandnot regulatedbythegovernment.In1975,aftertherevisionoftheBasicEducationLaw,miscellaneousschools which fulfilled certain conditions were promoted to specialised trainingcolleges. In addition, from 1994 onwards, studentswho completed courses of twoyears ormore or 1,700 hours ormore were awarded a ‘Diploma’, and from 2006onwards,thosefinishedfouryearsofspecialisedcourseswithcertainconditionsweregrantedan‘AdvancedDiploma’.Thegovernmentalsocreatedalinkagebetweenthespecialisedtrainingcollegesanduniversities.Thisenabledgraduatesofspecialisedtraining colleges to transfer to universities, subject to fulfilling certain conditions.Nowadays, specialised training colleges can be categorised into three types byprograms–thoseassociatedwithgeneralprograms,uppersecondaryprograms,andspecialisedprograms.Generalprogramsdonothaveanypre-requisites.Ontheotherhand,uppersecondaryprogramsrequiregraduationfromlowersecondaryeducation,andspecialisedprogramsrequiregraduationfromuppersecondaryeducation.1Themain areas of education include engineering, agriculture, healthcare, hygiene suchas hairdressing, cooking and nutritional management, social welfare, commercialbusiness, home economics, and culture/general education. The curriculum withinspecialised training colleges emphasizes practical education.2 Some of them grantcandidacyforthestateexaminationsforprofessionssuchasnurses,physicaltherapists,occupational therapists, medical technologists, architects, registered surveyors, taxaccountants,andsoon.

The education system of Australia is quite different from that of Japan.Australia’s national system, the AQF, incorporates the qualifications from eacheducational and training sector into a single national qualification frameworkwith10 qualification levels. TAFE (Technical and Further Education) colleges awardAQF qualifications aligned asCertificate 1-IV,Diploma,GraduateCertificate, andGraduateDiplomaqualifications.Ontheotherhand,universitiesgrantbachelorsandhonours,masters,anddoctoraldegreesfollowingcompletionofacademiceducation.Eachqualificationatteststhattheindividualhasacertainlevelofskillorknowledge,beitacademicorvocational.1 In 2014, therewere 3,205 specialised training colleges in Japan, amongwhich 2,812 offeredspecialisedcourses,438haduppersecondarycourses,and178offeredgeneralcourses.2Practicaleducationisusuallyofferedwithinthecolleges.Usually,thisisdifferentfromOJT.

237TOMOKO KISHI


Figure 2 - The Japanese educational system

Source:MinistryofEducation,Culture,Sports,Science,andTechnologyinJapanwebsiteforspecialisedtrainingcolleges.http://www.mext.go.jp/component/a_menu/education/detail/__icsFiles/afieldfile/2013/10/11/1322361_2_1.pdf

Collegesoftechnology(5years)

GraduateschoolsDoctorateprogram(3years)Master’sprogram(2years)

Specialisedtrainingcolleges(generalprogram)

Miscellaneousschools

Juniorcolleges(2years)

Universities(4years)

Uppersecondaryschools(3years)

Lowersecondaryschools(3years)

Elementaryschools(6years)

Specialisedtrainingcolleges(specialisedprogram)

Specialisedtrainingcolleges(uppersecondaryprogram)


50

40

50

25

10

2001

2004

2005

2006

2007

2008

2009

2013

35

20

2010

2011

2012

UniversitiesJunior collegesSpecialised training colleges

2002

2003

45

30

15

%

Theabsenceofthefixedconceptofa‘qualification’inJapanmakesithardtomakeprecisecomparisonswith theAustraliancontext.However, therearesomesimilarities between the vocational education systems of Australia and Japan. Forbothcountries,universitiesgrantbachelor’sdegrees,master’sdegrees,anddoctoraldegrees. Both countries have post-secondary vocational education systems. In thefollowingsection,wecomparetheeffectsofhighereducationandqualificationsonlabourmarketoutcomesinbothcountries.

Statistics on post-secondary education in Australia and Japan Official data show that higher education enrolment rates have been increasing inbothAustralia and Japan.However, thedatapublishedpertaining to the indicatorsforeducationdiffer.TheJapanesegovernmentpublishesdataon theratioofnewlyenrolleduniversity,juniorcollege,orspecialisedtrainingcollegestudentstototalhighschoolgraduates,asshowninfigure3.1.Ontheotherhand,theAustraliangovernmentpublishesdataforstudentsenrolledinhighereducationasaportionofthepopulationbyagerange,asfigure3.2shows.ForJapan,theuniversityenrolmentrateincreasedfrom2001to2009buthasremainedrelativelystablesince2009.Theenrolmentrateforjuniorcollegeshasbeendeclining,whilethatforspecialisedtrainingcollegeshasbeenincreasingafterreachinganadirin2009.

ForAustralia, theenrolment rate forbachelor’sorhigherdegreeshasbeenincreasingrapidly,whereas that foradvanceddiplomasordiplomashasbeenfairlystable.TheenrolmentforCertificateIIIorIVhasincreasedfrom2.4percentto6.1percentintheperiod2001-2013(figure3.2)

Figure 3.1 Ratio of female students enrolled in higher education to female high school graduates, 2001-2013, Japan

Source:MinistryofEducation,Culture,Sports,Science,andTechnologyofJapan,SchoolBasicSurvey2001-2013.

239TOMOKO KISHI


Figure 3.2 - Enrolment in higher education for females aged 18-24, 2001-2013, Australia

Source:AustralianBureauofStatistics,DataCatalogue4125.0(2014).Note:ThedatashowfemalesenrolledinAdvancedDiplomas,Diplomas,andCertificatesIIIandIVasaportionofallfemalesaged18-24.

4. Previous research Previous research in Japan In Japan, a number of researchers have estimated female labour supply functions,usually focusingon theeffectsofchildbirthonemployment.Research that focuseson the relationship betweenwomen’s education and employment outcomes is rare.However, some researchers have estimated the effect of education on women’semploymentoutcomes,evenwherethiswasnottheprimaryconsideration.MatsuuraandShigeno(2001)employstaticestimation tofindapositiverelationshipbetweeneducation levels and employment. In contrast, Nawata and Ii (2004) conclude thathigher educationhas anegative effect on female employment in Japan.3Waldfogelet al. (1999), analysing the impactof childcare leave,find that the effect ofhighereducation on post-childbirth job retention of female employees isweaker in JapanthanintheUnitedStatesorBritain.Kohara(2010)usesadynamicmodeltoestimatelaboursupplyforwomenwhosehusbandsfacejoblossandfindsthathighereducationhasno significant impact compared to seniorhigh school education.Okamura andIslam (2011) estimate inter-temporal participation decisions ofwomenusing linearprobabilitymodelsthatassumestatedependenceandunobservedheterogeneity;theyfindthathighlyeducatedwomenarelesslikelythanless-educatedwomentore-enterthelabourmarketafterchildbirth.

35.0

5.0

0.0

25.0

10.0

2001

2004

2005

2006

2007

2008

2009

2013

20.0

2010

2011

2012

Bachelor degreeAdvanced diploma, diploma

2002

2003

30.0

15.0%

Certificate III/IV

3Asidefromtheanalysesoftheeffectsofeducation,someresearchers,includingEdwardsandPasquale(2003),pointoutthatinJapan,highereducationforwomenisnotcareeroriented.


Previous research in Australia Australiaalsohasarichliteratureonfemalelaboursupply.Anumberofresearchershavefocusedonspousalincomesorwomen’sownwagesandfemalelaboursupply,asreviewedinBirch(2005),orontheeffectsofchildcareonlaboursupply(Breuniget al., 2012; andMoschion, 2013). Recent analyses using panel data indicate thathighereducationalattainment,includingvocationaleducation,raisestheprobabilityofobtainingajob.Cai(2010)showsthatcompletingmorethan11yearsofeducationhasasignificantpositiveeffectonfemalelaboursupply.TheresultsofBuddelmeyeret al.(2006)alsoindicatethatdegreesanddiplomashavesignificantpositiveeffectsonfemaleemployment,evenaftercontrollingforstatedependence.MitchellandWelters(2008), using aCoxproportional hazard analysis of labourmarket transitions,findthatthehazardratesoutofcasualemploymentandintopermanentemploymentarehigher for thosewithbachelor’sdegreesorpost-bachelor’s education.On theotherhand, some researchers, suchasMavromaraset al. (2009),have recentlynoted theexistenceofoverlyeducatedandoverlyskilledjob-seekersinAustralia.

Previous comparative research on Australia and Japan ComparisonsoffemalelaboursupplybetweenAustraliaandJapanarestillrare.Daly,Meng,Kawaguchi,andMunford(2006)estimatedthegenderwagegapinfourcountriesincludingAustraliaandJapan.TheyfoundthatsecondaryeducationaswellastertiaryeducationhadsignificantpositiveeffectsonfemalewagesinbothAustraliaandJapan.Kishi (2014) found that part-time employment in Japan was different from that inAustraliabothintermsoftherelationshipwitheducationandintermsofjobduration.

5. Data Two longitudinal datasets Weusepaneldata fromtheHousehold IncomeandLabourDynamics inAustralia(HILDA) Survey and the Japanese Panel Survey on Consumers (JPSC). HILDASurvey is a household-based panel dataset collected by theMelbourne Institute ofAppliedEconomic andSocialResearch,which has been gathering information oneconomicandsubjectivewell-being,labourmarketdynamics,andfamilystructuressince2001byfollowingrespondentsovertime.4TheJPSCcollectsdataonthesimilartopics and has been conducted by the Institute onHousehold Economics in Japansince1993.ThefeaturesofboththeHILDASurveyandtheJPSCaredescribedinAppendix A2. There are some differences between the two longitudinal datasets;for theHILDASurvey, interviews are conducted annually,while for the JPSC, nointerviewsareconducted;questionnairesaresenttorespondentsandreturnedtotheinstitutebypost.TheHILDASurveycoversbothmenandwomenof15yearsofageorolder,whiletheJPSCcoversonlywomenof24yearsofageorolder.

4DetailedinformationonHILDASurveyisavailablefromhttp://www.melbourneinstitute.com/hilda/.

241TOMOKO KISHI


For theHILDASurvey,we use only data forwomen in order tomaintaincomparabilitywiththeJPSC.Thestudyperiodis2005-2009.FortheHILDASurvey,thiscorrespondstowaves5-9,whilefortheJPSC,itcorrespondstowaves13-17.Weconfineouranalysistorespondentswhofulfilthefollowingconditions:

(1) Femalerespondentsaged25-40years,(2) Respondents who provided answers to all questions pertaining to

education,qualifications,employmentstatus,andincomesforeachyearfrom2005through2009,and

(3) Respondentswhoseemploymentstatusisclear.

Thesecondconditionisimposedbecausetheeconometricmodel(explainedinSubsection5.1)requiresthatthepaneldatahavenoattritionandthatthedatacoverthe samenumber of timeperiods for all respondents.Tomeet the third condition,we exclude respondents who did not provide answers to questions regarding theiremployment status.We also exclude self-employed people and familymembers ofJPSCemployees.Aftermakingtheseexclusions,thenumberofobservationsweuseinthedynamicestimationis912x5=4,560fortheHILDASurveyand764x5=3,820fortheJPSC.Thenumberofobservationsusedinthestaticestimationsislowerstill,becauseobservationswithmissingvaluesonfamilyvariablesaredeleted.

The effects of past employment status We classify employment status into three types. Within the HILDA sample, wedistinguishthreecategoriesofemploymentstatus:permanent(on-going)employment,casualemployment,andnotworking.Casualemploymentmeansemploymentwithoutpaidleave(FairWorkOmbudsman,2012)andischaracterisedbyalackofjobsecurity.For the JPSC, we also classify observations into three categories for employmentstatus:employmentonapermanentoron-goingbasis5,other typesofemployment,andnotworking.Amongtheemploymentsequences,thetenmostfrequentpatternsforthetwosamplesarelistedinAppendixB.IntheHILDASurvey,40percentofrespondentsretainthesameemploymentstatusforallfiveconsecutiveyears,whilethecorrespondingproportionisover60percentfortheJPSC.ThissuggeststhatJapanesepeoplechange theiremploymentstatusonly infrequently.This indicatespotentiallystrongstatedependenceofemploymentstatus,particularlyinJapan.

The model Weusetwokindsofmodelsforoureconometricanalysis;oneisstaticandtheotheris dynamic. The reasonwhywe use the lattermodel is that previous research hasfoundthatfemalelaboursupplyischaracterisedbystatedependence,asdetailedinHeckman(1981).Thatis,weassumethatworkexperienceraisestheprobabilitythatawomanwillworkinthefuture,evenifinitialentryintotheworkforceisdeterminedbyarandomprocess.Inordertocapturestatedependence,weusearandom-effectsdynamicprobitmodel;thebasicframeworkforthismodelisgivenbyStewart(2006).

5InJapan,employeesareclassifiedintoregularandnon-regularemployeesintheofficialstatistics.Regularemployeesroughlycorrespondtoemployeeswhoworkonapermanentorongoingbasis.On theotherhand,non-regularemployeesdonotcorrespond tocasualemployees inAustralia.Forthisreason,weconfinedouranalysistotheestimationsofemploymentandemploymentonapermanentbasis.Themeaningof‘non-regularemployees’isdetailedinKambayashi(2013).


Model 1 Supposethattheemploymentstatusforpersoniinperiodtisyitandthatthelatentvariableforyitisy

*it.

1 ifyit*3>0yit=(1) 0 ifyit*<0yit

*=x8it b + ai + eit

In equation (1), xit is a vector of exogenous explanatory variables; ai denotes anindividualspecific(time-invariant)errorterm;eitistheerrorterm;and

eit ~N(0,se2)

Equation(1)isestimatedusingbotharandom-effectprobitmodelandaprobitmodelclusteredforindividuals.

Model 2

1 ifyit*3>0yit=(2) 0 ifyit*<0yit

*=gyit –1 + x8it b + ai + uit

Inequation(2),xitisavectorofexogenousexplanatoryvariables;aiisanindividual-specific(timeinvariant)errorterm;uitisanerrorterm;and

uit~N(0,su2)

Equation(2)isestimatedusingtherandom-effectdynamicprobitmodelbasedontheStewartmethod(Stewart,2006).Thismethodassumesthattheinitialconditionfortheemploymentstatusisspecifiedasinequation(3).

yi1

*=zi18 p + hi (3)

whereZi1isavectorofexogenousinstrumentsandhiisanerrortermcorrelatedwithai butuncorrelatedwithuitfort3> 2.

hi=qai + ui1 (4)

Substitutingequation(4)forhiinEquation(3),weobtainequation(5).

yi1*=zi18 p + qai + ui1 (5)

243TOMOKO KISHI


We assume that the error term follows an auto-regressive process, as shown inequation(6).

uit=ruit–1 + eit(6)

Equations(3)and(5)areestimatedsimultaneously.6Whenestimatingequation(5),weusedatafortheyear2005.Thesemodelsareestimatedusingvariablesdrawnfromthetwodatasets,asdetailedinthenextsubsection.

Usingbothmodel1andmodel2,weperformtwokindsofestimations. Inthefirstestimation,yit=1iftherespondentworksonapermanentbasisandyit=0otherwise;inthesecondestimation,yit=1iftherespondentworksonanybasis,andyit=0otherwise.

The variables ThevariablesusedfortheanalysesoftheHILDASurveyandtheJPSCareasshownintable1.

Table 1 - Variables used in the estimations

HILDA Survey JPSCDependent variable 1)Dummy=1iftherespondentis 1)Dummy=1iftherespondent employedinthetthperiod. isemployedinthetthperiod. 2)Dummy=1iftherespondentis 2)Dummy=1iftherespondent employedonapermanentbasisin isemployedonapermanent thetthperiod basisinthetthperiodIndependent variablesAge Respondent’sageasofJune30 Respondent’sageasofOctober1Children0_4 Numberofchildrenaged0-4years Numberofchildrenaged0-4yearsChildren5_14 Numberofchildrenaged5-14years Numberofchildrenaged5-14yearsSpousalincome SpousalincomeinAUD1,000 SpousalincomeinJPYmillionDummy variables for educationReferencegroup Respondentswhosehighest Respondentswhosehighest educationisyear12orless,or educationissecondaryschoolor certificatesI/II specialisedtrainingcollegewith lowersecondaryprogramsor generalprogramsPostgraduate Dummy=1iftherespondenthasa Dummy=1iftherespondenthas postgraduatedegree finishedapostgraduatecourseBachelor Dummy=1iftherespondenthasa Dummy=1iftherespondenthasa bachelor’sorhonoursdegree bachelor’sdegreeAdvanceddiploma/ Dummy=1iftherespondenthasa Dummy=1iftherespondenthasdiploma(HILDA diplomaoradvanceddiploma finishedajuniorcollegeoracollegeSurvey)Junior oftechnologycollege/collegeoftechnology(JPSC)

6 We apply a ‘random effects dynamic probit model with auto-correlated errors’ (redpace)commandinStata.


Table 1 - Variables used in the estimations (continued)

HILDA Survey JPSCCertificateIII/IV Dummy=1iftherespondenthasa Dummy=1iftherespondenthas(HILDASurvey) certificateIIIorIV finishedaspecialisedcourseofSpecialised specialisedtechnicalcollegetechnicalcollege(JPSC)lag1(Employed) Dummy=1iftherespondentis Dummy=1iftherespondentis employedinthetthperiod. employedinthetthperiod.lag1(Permanent) Dummy=1iftherespondentis Dummy=1iftherespondentis employedonapermanentbasisin employedonapermanentbasisin thetthperiod thetthperiod

Someexplanationisrequiredpertainingtothedummyvariableforeducation.

The variable ‘postgraduate’ represents postgraduate degrees for both the HILDASurvey and the JPSC. It represents a graduate certificate, a graduate diploma, amaster’s degree, or a doctoral degree for theHILDASurvey,while it covers onlymaster’s degrees and doctoral degrees for the JPSC, as graduate certificates andgraduate diplomas are not awarded in Japan.The variable ‘bachelor’ pertains to abachelor’sdegreeorhonoursfortheHILDASurveyandonlytoabachelor’sdegreefortheJPSC,asanhonoursdegreeisnotadistinctformalqualificationinJapan.Theeducation dummies for the qualifications other than ‘postgraduate’ and ‘bachelors’differbetweenthetwocountries.FortheHILDASurvey,weusethedummyvariable‘Advanceddiploma/diploma’and‘CertificateIII/IV’.Thereferencegroupconsistsofthe respondentswhose highest education isCertificate I/II orYear 12 or less. Forthe JPSC,we use the dummy variables ‘Junior college/college of technology’ and‘specialisedtrainingcollege’.7Asexplainedintheprevioussection,thereareseveraltypes of specialised training colleges in Japan.We distinguish specialised coursesoffered by specialised training colleges from other courses. The dummy variable‘specialisedtrainingcollege’is1iftherespondent’shighesteducationisaspecialisedcourse offered by a specialised training college. The reference group consists ofrespondentswhosehighesteducationisseniorhighschool,orageneralcourseofferedbyaspecialisedtrainingcollege,oramiscellaneousschool.ThedescriptivestatisticsfortheexplanatoryvariablesfortheHILDASurveyandtheJPSCarelistedintable2.

Wecalculatedtheproportionofemploymentandemploymentonapermanentbasisforeachlevelofeducationorqualification,forboththeHILDASurveyandtheJPSC samples (table3).This suggests that theproportion employed increaseswitheducationorqualificationfortheHILDASurveybutnotfortheJPSC.Ontheotherhand,theproportionofthoseemployedonapermanentbasishasapositiverelationshipwitheducationforbothsamples.

7IntheJPSC,juniorcollegegraduatesarenotseparatedfromthosewhograduatedfromcollegesoftechnology.Thismakesitdifficulttoestimatetheeffectofcollegesoftechnology.

245TOMOKO KISHI


Table 2 - Descriptive statistics

HILDA Survey JPSC Standard StandardVariable Mean deviation Variable Mean deviationEmployed 0.718 0.45 Employed 0.495 0.5Employed(t-1) 0.719 0.45 Employed(t-1) 0.493 0.5Permanentemployment 0.485 0.5 Permanentemployment 0.212 0.409Permanentemployment(t-1) 0.491 0.5 Permanentemployment(t-1) 0.226 0.419Children0_4 0.719 0.801 Children0_4 0.623 0.698Children5_14 0.765 1.03 Children5_14 0.848 0.848Spousalincome(AUD1,000) 59.694 46.554 Spousalincome(JPY1,000 503.31 255.16 thousand)Postgraduate 0.131 0.337 Postgraduate 0.011 0.104Bachelor 0.273 0.446 Bachelor 0.164 0.37Advanceddiploma,diploma 0.105 0.306 Juniorcollegeorcollege 0.264 0.441 oftechnologyCertificateIIIorIV 0.11 0.312 Specialisedtrainingcollege 0.171 0.376Nofobservations 3,247 2,359

Note:Author’scalculationsbasedontheHILDASurvey,waves5-9andtheJPSC,waves13-17.

Table 3 - Labour market outcomes by education/qualifications, HILDA Survey and the JPSC

Proportion Proportion employed employed Proportion on a Proportion on a employed permanent employed permanentHILDA Survey (%) basis (%) JPSC (%) basis (%)Postgraduate 83.29 62.35 Postgraduate 57.69 53.85Bachelor 80.59 64.79 Bachelor 48.32 33.59Advanceddiploma, 76.97 52.25 Juniorcollegeorcollege 45.10 21.03diploma oftechnologyCertificateIII/IV 63.95 36.05 Specialisedtraining 56.82 25.56 collegeCertificateI/IIorlower 62.04 34.30 Seniorhighschool, 49.57 13.26(reference) specialisedtraining college(notspecialised courses)

Note:Author’scalculationsbasedontheHILDASurvey,waves5-9andtheJPSC,waves13-17.


6. Econometric results Asexplainedintheprevioussubsection,ouranalysesaremainlycomposedoffourkindsofestimations,asshowninthefollowingtable.

Model Type Description of the type DataModel(Static) 1 y=1ifemployedonapermanentbasis;y=0otherwise HILDA JSPC 2 y=1ifemployedonanybasis;y=0otherwise HILDA JSPCModel(Dynamic) 1 y=1ifemployedonapermanentbasis;y=0otherwise HILDA JSPC 2 y=1ifemployedonanybasis;y=0otherwise HILDA JSPC

Theestimatedresultsarepresentedintables4.1through6.2.

Table 4.1 - Estimated results for permanent or ongoing employment, 2005-2009 (1) Dependent variable: employed on a permanent or ongoing basis in wave t

Random-effect Random-effect probit estimation probit estimation Marginal MarginalHILDA Survey effects z-value JPSC effects z-valueAge 0.0041 1.08 Age 0.0012 0.04Children0_4 -0.2265 -16.31*** Children0_4 -0.0822 -6.81***Children5_14 -0.1132 -7.84*** Children5_14 -0.0452 -3.46***Spousalincome 0.0002 0.89 Spousalincome -0.1311 -2.63***Postgraduate 0.2470 5.41*** Postgraduate 0.3690 2.22**Bachelor 0.2586 6.94*** Bachelor 0.2450 4.27***Advanceddiploma, 0.1607 3.23*** Juniorcollege,college 0.0979 2.15**diploma oftechnologyCertificateIII/IV 0.0068 0.15 Specialisedtrainingcollege 0.1318 2.41**Nofobservations 3247 Nofobservations 2359Nofgroups 749 Nofgroups 546Waldchi2 336.730 Waldchi2 84.54 Prob>chi2=0.0000 Prob>chi2=0.0000

Note:Author’sestimationbasedontheHILDASurvey,waves5-9andtheJPSC,waves13-17.***,**and*denotesignificanceatthe1%,5%,and10%levels,respectively.

247TOMOKO KISHI


Table 4.2 - Estimated results for permanent or ongoing employment, 2005-2009 (2)

Dependent variable: employed on a permanent or ongoing basis in wave t

Pooled probit Pooled probit estimation, clustered estimation, clustered with individuals with individuals Marginal MarginalHILDA Survey effects z-value JPSC effects z-valueAge 0.0027 0.56 Age 0.0003 0.06Children0_4 -0.2168 -11.64*** Children0_4 -0.0962 -4.44***Children5_14 -0.1075 -5.90*** Children5_14 -0.0631 -3.31***Spousalincome 0.0004 1.09 Spousalincome -0.2945 -3.73***Postgraduate 0.2280 4.61*** Postgraduate 0.4065 2.51**Bachelor 0.2721 6.96*** Bachelor’sdegree 0.2405 4.02***Advanceddiploma,diploma 0.1425 2.57** Juniorcollege,collegeof 0.0996 2.11** technologyCertificateIII/IV -0.0049 -0.09 Specialisedtrainingcollege 0.1453 2.56**Nofobservations 3247 Nofobservations 2359Nofgroups 749 Nofgroups 546Waldchi2 247.25 Waldchi2 82.50Logpseudo-likelihood -1927.305 Logpseudo-likelihood -1104.57


Table 5.1 - Estimated results for employment on any basis, 2005-2009 (1)

Dependent variable: employed in wave t

Random-effect Random-effect probit estimation probit estimation Marginal MarginalHILDA Survey effects z-value JPSC effects z-valueAge 0.0121 3.57*** Age -0.0023 -0.62Children0_4 -0.2005 -16.44*** Children0_4 -0.2461 -13.97***Children5_14 -0.0916 -7.81*** Children_514 -0.0128 -0.96Spousalincome -0.0001 -0.44 Spousalincome -0.5820 -9.49***Postgraduate 0.1474 4.81*** Postgraduate 0.1075 1.06Bachelor 0.1479 5.45*** Bachelor 0.0775 2.32**Advanceddiploma,diploma 0.1009 2.92*** Juniorcollegeorcollege 0.0005 0.02 oftechnologyCertificateIII/IV 0.0101 0.29 Specialisedtrainingcollege 0.0871 2.76***Nofobservations 3247 Nofobservations 2359Nofgroups 749 Nofgroups 546Waldchi2 318.02 Waldchi2 303.78



Table 5.2 - Estimated results for employment on any basis, 2005-2009 (2)Dependent variable: employed on a permanent or ongoing basis in wave t

Pooled probit Pooled probit estimation, clustered estimation, clustered with individuals with individuals Marginal MarginalHILDA Survey effects z-value JPSC effects z-valueAge 0.0107 2.75*** Age -0.0025 -0.39Children0_4 -0.1985 -13.49*** Children0_4 -0.2468 -9.33***Children5_14 -0.0773 -5.84*** Children_514 -0.0127 -0.56Spousalincome -0.0002 -0.67 Spousalincome -0.5768 -4.09***Postgraduate 0.1579 5.32*** Postgraduate 0.1065 0.64Bachelororhonours 0.1528 5.50*** Bachelor’sdegree 0.0775 1.32Advanceddiploma,diploma 0.0897 2.37*** Juniorcollegeorcollege 0.0004 0.01 oftechnologyCertificateIII/IV -0.0059 -0.14 Specialisedtrainingcollege 0.0872 1.55Nofobservations 3247 Nofobservations 2359Nofgroups 749 Nofgroups 546Waldchi2 275.23 Waldchi2 121.20logpseudo-likelihood -1650.014 logpseudo-likelihood -1451.8503

Note: Author’sestimationbasedontheHILDASurvey,waves5-9andtheJPSC,waves13-17.***,**and*denotesignificanceatthe1%,5%,and10%levels,respectively.

Table 6.1 - Estimated results for permanent or ongoing employment, 2005-2009 (3) Dependent variable: employment on a permanent or ongoing basis in wave t

Random-effect Random-effect dynamic probit dynamic probit estimation with estimation with auto-correlated auto-correlated errors errors Estimated EstimatedHILDA Survey coefficients z-value JPSC coefficients z-valuelag1(permanent) 0.8401 2.74*** lag1(permanent) 2.2230 4.99***Children0_4 -0.5164 -6.19*** Children0_4 -0.2719 -1.82*Children5_14 0.0017 0.03 Children5_14 0.0731 0.77Spousalincomes 0.0001 0.10 Spousalincomes -0.8398 -1.85*Postgraduate 1.0905 4.18*** Postgraduate 1.4015 1.41Bachelororhonours 1.3390 5.66*** Bachelor 0.9580 2.20**Advanceddiploma,diploma 0.7852 3.08*** Juniorcollege,college 0.4658 1.74* oftechnologyCertificateIII/IV 0.1860 0.88 Specialisedtrainingcollege 0.6309 1.70Intercept -0.7392 -3.83*** Intercept -2.1401 -5.40***Nofobservations 4560 Nofobservations 3820Loglikelihood -1264.8608 Loglikelihood -476.574Waldchi2 161.95 Waldchi2 137.76lambda 0.608 8.08*** lambda 0.5436 2.09**AR(1) 0.011 0.07 AR(1) -0.1609 -1.26theta 0.959 5.51*** theta 1.6829 2.01**


249TOMOKO KISHI


Table 6.2 - Estimated results for employment on any employment basis, 2005-2009 (3) Dependent variable: employment in wave t

Random-effect Random-effect dynamic probit dynamic probit estimation with estimation with auto-correlated auto-correlated errors errors Estimated EstimatedHILDA Survey coefficients z-value JPSC coefficients z-valueemployedlag1(employed) 0.7781 2.65*** lag1(employed) 1.8440 6.34***Children0_4 -0.6279 -7.20*** Children0_4 -0.2600 -2.36**Children5_14 -0.0540 -0.88 Children5_14 0.1515 2.25**Spousalincomes -0.0011 -1.17 Spousalincomes -0.3882 -1.82*Postgraduate 0.7725 3.62*** Postgraduate 0.2144 0.32Bachelor 0.7314 4.13*** Bachelor 0.1480 0.79Advanceddiploma,diploma 0.4238 1.97* Juniorcollege,college 0.0174 0.11 oftechnologyCertificateIII/IV 0.0454 0.26 specialisedtrainingcollege 0.1807 0.93Intercept 0.6366 2.18** Intercept -0.7989 -3.15Nofobservations 4560 Nofobservations 3820Loglikelihood -1193.756 Loglikelihood -855.1954Waldchi2 174.56 Waldchi2 148.75lambda 0.4743 5.97*** lambda 0.4155 2.29AR(1) -0.0222 -0.15 AR(1) -0.1713 -1.7theta 0.6674 4.48*** theta 1.2302 2.88


Tables4.1and4.2areobtainedfromthestaticestimationbasedonmodel1,type1.Theyindicatethatmostofthevariablespertainingtoeducationorqualificationshave significant positive effects on the probability that a respondent works on apermanentorongoingbasis.FortheHILDASurvey,threecategoriesofqualifications– postgraduate degrees, bachelors or honours, and advanced diplomas or diplomas–have significantpositive effects.For the JPSC, four categoriesofpost-secondaryeducationhavesignificantpositiveeffectsonpermanentemployment.ThecomparisonbetweentheHILDAandJPSCresultsrevealsthatthemarginaleffectsofthebachelor’sdegreesarealmostthesameforbothsamples,whilethosefortheadvanceddiplomasordiplomasintheHILDASurveyarehigherthanthoseforjuniorcollegeorcollegesoftechnologyintheJPSC.

Tables5.1and5.2displaytheestimatedresultsformodel1,type2.Table5-1,which isbasedonrandom-effectsestimation, indicates that for theHILDASurvey,the three dummy variables for qualifications – postgraduate degrees, bachelor orhonours,andadvanceddiplomas/diplomas–havesignificanteffectsontheprobabilitythattherespondentworks.Ontheotherhand,fortheJPSC,onlybachelor’sdegreeshavesignificantpositiveeffects.However,themarginaleffectofthebachelor’sdegreefor theJPSC,0.078, is lower thanthatfor itsHILDAcounterpart,0.148.Table5.2,basedonapooledprobitestimationwithclusters,showsthatfortheJPSC,noneofthevariablespertainingtoeducationalattainmenthavesignificanteffectstoemployment.


A comparison between table 4.1 and 5.1 reveals that for both surveys,marginal effects of all dummy variables for higher education or qualifications arehigher for employment on a permanent/ongoing basis than for other employmentstatuses.However,thedifferencesinthemarginaleffectsofeducation/qualificationsdummiesbetweenthetwotablesaregreaterfortheJPSCthanfortheHILDAsurvey.Forexample,thedifferencesinthemarginaleffectsofbachelor’sdegreesarelargerfor theJPSC(0.2450 in table4.1and0.0775 in table5.1) than that for theHILDASurvey(0.2586intable4.1and0.1479intable5.1).Similarly,thedifferencesinthemarginaleffectsofjuniorcollegesorcollegesoftechnologyfortheJPSC(0.0979intable4.1and0intable5.1)arealsolargerthanthedifferencesinthemarginaleffectsofadvanceddiploma/diplomafortheHILDASurvey(0.1607intable4.1and0.1009intable5.1).Similarobservationsareobtainedfromacomparisonbetweentable4.2and5.2.ThissuggeststhatfortheJPSCs,thereturntohighereducationorqualificationismoredependentonwhethertheemploymentisonapermanentorongoingbasisthanitisfortheHILDASurvey.

Tables6.1and6.2depictthedynamicestimationresultsbasedonmodel2,types1and2,respectively.Table6.1demonstratesthat,fortheHILDASurvey,postgraduatedegrees,bachelor’sdegrees,advanceddiplomasordiplomashavesignificantpositiveeffectsonpermanentorongoingemployment.Ontheotherhand,fortheJPSC,onlybachelor’sdegreessignificantlyraisetheprobabilityofworkingonapermanentbasis.Table6.2showsthat,fortheHILDASurvey,postgraduatedegrees,bachelor’sdegrees,andadvanceddiploma/diplomahavesignificantpositiveeffectsonemploymentonanybasis.For the JPSC,noneof the educationdummieshas significant effectsonemploymentinthedynamicestimationresults.

Comparing the HILDA and JPSC results leads to several observations.Firstly,theeffectsofeducationonemployment(foralltypesofemploymentstatus)are significant andpositive in the estimations usingHILDAdata but not for thoseusing JPSC data. The effects of higher education on permanent employment areobservedfortheJPSCbuttheresultsarenotrobust.Secondly,fortheJPSCresults,theeffectsofpostgraduatedegrees,bachelor’sdegrees,andjuniorcollege/collegesoftechnologyonpermanentemploymentarelesssignificantstatisticallyforthedynamicmodel(table6.1)thanforthestaticmodel(tables4.1and5.1).Thissuggeststhattheeffectsofeducationarecancelledbytheeffectsofthepreviousperiod’semploymentstatus.Thisinturnsuggeststhat,inJapan,thepreviousperiod’semploymentstatusismoreimportantforemploymentthaneducationalattainmentorqualificationsinthepresentperiod.Thirdly,fortheHILDASurvey,theeffectsofdegreesordiplomasaresignificantinthedynamicestimationresultsaswellasinthestaticresults.Fourthly,themarginaleffectsofbachelor’sdegreesonemploymentareweaker for theJPSCthan for theHILDA Survey. One possible explanation is that university educationinJapanisnotnecessarilycareer-oriented,aspointedoutbyEdwardsandPasquale(2003). Another explanation is that the labour demand forwomenwith bachelor’sdegreesissoweakinsomeindustries(Hori,2009;andSano,2009)thatlaboursupplyofwomenwithbachelor’sdegreestendstobeexcessive.However,theinterpretationofthisresultneedsfurtherresearch.

251TOMOKO KISHI


7. Concluding remarks Weestimatedfemalelaboursupplyfunctions,payingspecialattentiontotheeffectsofeducationandqualificationsonlabourmarketandemploymentoutcomesinAustraliaand Japan.Weadopted adynamicmodel aswell as staticmodels toobtain robustresultsforthetwocountries.

Asthequalificationsystemsforthetwocountriesdiffer,itwasnotpossibletocomparetheeffectsofeducationorqualificationsonemploymentinasymmetricmanner.However,wewere able tofind somedifferencesbetween the econometricresultsforthetwocountries.

One of the major dissimilarities is that postsecondary education does notleadtoahigherprobabilityofemployment(foranyemploymentstatus)intheJPSCresults,unlikefortheHILDAresults.Anotherimportantdissimilarityisthat,fortheJPSC, the effects of education areweaker for the dynamic estimationwith laggedemploymentvariablesthanforthestaticestimation.Incontrast,fortheHILDASurvey,theeffectsofeducationarenotsignificantlydifferentbetweenthedynamicandstaticestimationresults.Thisresultcouldbeinterpretedinthefollowingway.FortheJPSC,theeffectsofpreviousemploymentstatusweresostrongthattheycancelledtheeffectsofeducation.However,thisisahypothesis;futureeconometricstudywillberequiredtoisolatetheeffectsofpreviousemploymentfromthoseofeducation.

Themarginaleffectsofbachelor’sdegreesonpermanentemploymentwereweakerfortheJPSCthanfortheHILDASurvey.Theinterpretationsforthisfindingalsoneedtobeinvestigatedfurther.

For Japan, specialised training colleges appear to have no effects onemploymentinthedynamicestimationresults.Incontrast,forAustralia,qualificationssuchasadvanceddiplomasordiplomasdidhaveaneffect.ThissuggeststhatskillsacquiredfromvocationaleducationhavenotyetbeenfullyutilizedinJapan.

Theneedforcertaintypesofeducation,skillformation,andskillevaluationvariesonthebasisofthelabourmarkettype.Inthepast,theJapaneselabourmarketwascharacterisedbylowlabourmobilityandwell-developedinternallabourmarkets.Forthisreason,experiencesacquiredwithinfirmshavebeenweighedhigherthantheskillsacquiredoutsidefirms.However,thelabourmarketinJapanhasbeenchangingsincethebeginningofthe1990sintermsoflabourmobilityanddiversificationsinthetypesofemployment,thusraisingtheimportanceofskillsacquiredoutsidefirms.Currently,thesystemsofeducation,qualifications,andlabourmarketinstitutionsareundergoingreformsothathumancapitaldevelopedbothoutsideandinsidethefirmscanbefullyutilizedtobearreturns.TheAustralianexperiencescouldbeusefulasanexampleofadvancedcasesforthisreform.


Appendices Appendix A1 - Characteristics of the Household Income and Labour Dynamics in Australia (HILDA) Survey and the Japanese Panel Survey on Consumers (JPSC)

HILDA Survey JPSCStartingyear 2001 1993Genderofrespondents Bothmenandwomen OnlywomenAgegroupforthefirstyear 14-92years 24-34yearsNumberofrespondentsinfirstwave 13,969 1,500Interviews Conductedannually NotconductedContinuity Panelmembersare Panelmembersare followedovertime followedovertimeToppingofthedata 5,477newmemberswere 500newmemberswere addedinwave11 addedinwave5 836newmemberswere addedinwave11Information Economicandsubjective Familydynamics, well-being,labourmarket labourmarketdynamics, dynamics,andfamily householdincomes, dynamics savings,andexpenditure

Appendix A2 - Sequence patterns for the HILDA Survey, waves 5-9

Sequence pattern Number of observations Percentage Cumulative percentage

22222 158 17.32 17.3200000 108 11.84 29.1711111 91 9.98 39.1422221 25 2.74 41.8910000 19 2.08 43.9701111 18 1.97 45.9412222 16 1.75 47.7021111 16 1.54 49.4500111 14 1.54 50.0010111 13 1.43 52.52Others 433 47.48 100.00Total 912 100.00

Notes:Author’scalculationsbasedontheHILDASurvey,waves5-9.2,1,and0denoteemploymentonapermanentbasis,othertypesofemployment,andnotworking,respectively.

253TOMOKO KISHI


Appendix A3 - Sequence patterns for the JPSC, waves 13-17

Sequence pattern Number of observations Percentage Cumulative percentage

22222 185 24.21 24.2100000 162 21.20 45.4211111 98 12.83 58.2501111 17 2.23 60.4700001 16 2.09 62.5710000 15 1.96 64.5322220 15 1.96 66.4900011 10 1.31 67.8011100 10 1.31 69.1111222 10 1.31 70.42Others 226 29.59 100.00Total 764 100.00

Note: Author’scalculationsbasedontheJPSC,waves13-17.2,1,and0denoteemploymentonapermanentbasis,othertypesofemployment,andnotworking,respectively.

References AustralianBureauofStatistics,LabourForce,Australia,September2014,Catalogue

6202.0.http://www.abs.gov.au/AUSSTATS/[email protected]/mf/6202.0(accessedon16October2014).

AustralianBureauofStatistics,GenderIndicators,2014,DataCatalogue4125.0.http://www.abs.gov.au/ausstats/[email protected]/mf/4125.0(accessedon16October2014).

AustralianFairWorkOmbudsmanwebsite.http://www.fairwork.gov.au/employment/casual-full-time-and-part-time-work/pages/default.aspx (accessed on 16October2014).

AustralianQualificationFrameworkwebsite.http://www.aqf.edu.au/(accessedon16October2014).

Birch,E.(2005),‘StudiesofTheLabourSupplyofAustralianWomen:WhatHaveWeLearned?’Economic Record,81,65-84.

Breunig,R.,Gong,X.andKing,A.(2012),‘PartneredWomen’sLabourSupplyandChildCareCostsinAustralia:MeasurementErrorandtheChildCarePrice’,Economic Record,88(1),51-69.

Buddelmeyer,H.,Wooden,M. andGhantous, S. (2006), ‘Transitions FromCasualEmployment inAustralia’, Report Prepared for theAustralianDepartmentofEmploymentandWorkplaceRelationsUnder theSocialPolicyResearchServicesAgreement,Melbourne Institute ofAppliedEconomic andSocialResearch.

Cai, L. (2010), ‘Work Choices of Married Women: Drivers of Change’, VisitingResearcherPaper,ProductivityCommission,Canberra.

Daly,A.,Meng,X.,Kawaguchi,A.andMunford,K.(2006),‘TheGenderWageGapinFourCountries’,Economic Record,82,165-176.

Edwards,L.andPasquale,M.(2003),‘Women’sHigherEducationinJapan:FamilyBackground, Economic Factors and the Equal Employment OpportunityLaw’,Journal of the Japanese and International Economies,17,1-32.


Goodman,R.,Hatakenaka,S.andKim,T.(2009),‘TheChangingStatusofVocationalHigher Education in Contemporary Japan and the Republic of Korea; ADiscussion Paper’,UNESCO-UNEVOC International Centre for TechnicalandVocationalTraining,Bonn,Germany.

Heckman,J.(1981),‘HeterogeneityandStateDependence’,inS.Rosen,(ed.),StudiesinLaborMarkets,UniversityofChicagoPress,Chicago,91-140.

Hori, H. (2009), ‘Labor Market Segmentation and the GenderWage Gap’, Japan Labor Review,6,5-20.

InstituteforResearchonHouseholdEconomicswebsite,http://www.kakeiken.or.jp/jp/index.html(accessedon16October2014).

JapanInstituteforLabourPolicyandTraining(JILPT)(2008),‘ShokugyoNoryokuHyokaToRodoShijo’(TranslatedfromJapaneseas‘EvaluationofVocationalCompetenceandLabourMarket’,Forewordbytheeditorialboard),Japanese Journal of Labour Studies,577,2-3.JILPT,Tokyo.

Kambayashi,R. (2013), ‘Differences InDefinitionsofNon-RegularEmployment inGovernmentStatistics’,Japan Labor Review,10,55-66.

Kishi,T.(2014),‘FemaleLabourSupplyinAustraliaandJapan;AnAnalysisBasedOnLongitudinalData’,Journal of Australian Studies, 27, 29-44,AustraliaStudiesAssociationofJapan.

Kohara,M.(2010),‘TheResponseofJapaneseWives’LaborSupplyToHusbands’JobLoss’,Journal of Population Economics,23,1133-1149.

Koike, K. (1991), ‘Daisotsu White-Collar no Jinzai Kaihatsu’ (Translated fromJapanese as ‘The Career Formation ofWhite-Collar EmployeesWhoAreUniversityGraduates’),ToyoKeizaiShimposhaPublishing,Tokyo.

Koike,K.andInoki,T.(2003),‘CollegeGraduatesinJapaneseIndustries’,TheJapanInstituteofLabour,Tokyo.

Mavromaras,K.,McGuinness,S.andFok,Y.K.(2009),‘OverskillingDynamicsandEducation Pathways’, IZADiscussion paper,No.4321,The Institute for theStudyofLabor(IZA),Bonn.

Matsuura,K.andShigeno,Y.(2001),‘JoseinoSentakutoKakeiChochiku’(Translatedfrom Japanese as ‘Women’s Choices and Household Incomes’), NihonHyoronshaPublishing,Tokyo.

MelbourneInstituteofAppliedEconomicsandSocialResearchwebsite,http://www.melbourneinstitute.com/hilda/(accessedon16October2014).

MinistryofEducation,Culture,Sports,Science,andTechnologyinJapanwebsite;Onthespecialisedtrainingcolleges,http://www.mext.go.jp/component/a_menu/education/detail/__icsFiles/afieldfile/ 2013/10/11/1322361_2_1.pdf (accessedon16October2014).

Ministry of Education, Culture, Sports, Science, and Technology in Japan, SchoolBasicSurvey,2001-2013.

Mitchell,W.andWelters,R.(2008),‘DoesCasualEmploymentProvideA“SteppingStone” toBetterWorkProspects?’,CentreofFullEmployment andEquityWorkingPaperNo.08-11,UniversityofNewcastle.

Moschion,J.(2013),‘TheImpactofFertilityonMothers’LabourSupplyinAustralia:EvidencefromExogenousVariationinFamilySize’,Economic Record,89,319-338.

255TOMOKO KISHI


Nawata, K. and Ii, M. (2004), ‘Estimation of the Labor Participation and WageEquationModelofJapaneseMarriedWomenbytheSimultaneousMaximumLikelihoodMethod’,Journal of the Japanese and International Economies,18,301-315.

Okamura,K.andIslam,N.(2011),‘Inter-TemporalLabourForceParticipationAmongMarriedWomeninJapan’,The Japanese Economic Review,62,562-580.

Sano,S.(2009),‘TestingtheTaste-BasedDiscriminationHypothesis:EvidencefromDataonJapaneseListedFirms’,Japan Labor Review,6,36-50.

StatisticsBureauofJapan,MinistryofInternalAffairandCommunications(2014),Labour Force Survey, Historical Data, http://www.stat.go.jp/data/roudou/longtime/03roudou.htm(accessedon17October2014).

Stewart,M.(2006),‘MaximumSimulatedLikelihoodEstimationofRandom-EffectsDynamicProbitModelsWithAutocorrelatedErrors’,The Stata Journal,6,256-272.

Waldfogel,J.,Higuchi,Y.andAbe,M.(1999),‘FamilyLeavePoliciesandWomen’sRetention After Childbirth: Evidence from the United States, Britain, andJapan’,Journal of Population Economics,12,523-545.

Female Labour Supply in Australia and Japan: The Effects ...€¦ · AQF qualificationsaligned as...

Documents

Transcript of Female Labour Supply in Australia and Japan: The Effects ...€¦ · AQF qualificationsaligned as...