The Rate of Return to Higher Education Over the Business Cycle · business cycle and its impact upon the private rate of return to higher education that is the subject of this paper

219AUSTRALIAN JOURNAL OF LABOUR ECONOMICS

Volume 16 • Number 2 • 2013 • pp 219 - 236

The Rate of Return to Higher Education Over the Business Cycle

Michael Corliss, Phil Lewis and Anne Daly,CentreforLabourMarketResearch,UniversityofCanberra

AbstractCurrently there is only a scant literature available on the business cycle effect upon the private rate of return to higher education in Australia. This is despite the expectation that the business cycle would be of a significant consequence to the relative income of degree holders. This paper presents estimates of the rate of return in Australia associated with the completion of Bachelors degrees for a range of fields of study over the business cycle. The results show that for the average person there are strong incentives to complete these degrees, and the private rate of return compares favourably with the real long term real bond rate. However, there are considerable differences in rates of return according to gender and discipline of study. There also appear to be important business cycle effects on the rate of return to a university degree.

JELClassification:J240,J440,J480

1. Introduction InrecentyearsAustraliahasgonethroughaconsiderablechangeineconomicactivity,starting in1991witha recessionandending in2006withaboom.Thesechangesin economic activity coincidewith the Census of Population of Housing (Census)years,conductedbytheAustralianBureauofStatisticsin1991,1996,2001,and2006.Evidenceof thebusinesscycle isgiven infigure1whichshows theunemploymentratefortheperiodusingestimatesfromboththeLabourForceSurveyandfromtheCensus.Theunemploymentrateisagoodindicatorofthelevelofeconomicactivitytakinginthelabourmarket(Lewiset al.,2010).Figure1showstheunemploymentratefromtheLabourForceSurvey,forDecember1991at10.3percent(recession),1996at8.5percent(recovery),2001at6.6percent(expansion),and2006at4.6percent(boom).TheunemploymentratesfromtheCensusesarealsoshownandalthoughtheydiffermarginally the twosetsofestimatesclearly indicate that theCensusperiodscorrespondtophasesofthebusinesscycle.

Addressforcorrespondence:PhilLewis,Director,CentreforLabourMarketResearch,CanberraUniversity,ACT0200.Email:[email protected]©CentreforLabourMarketResearch,2013

220AUSTRALIAN JOURNAL OF LABOUR ECONOMICSVOLUME 16 • NUMBER 2 • 2013

Figure 1 - Unemployment rate, per cent, 1991-1992

Source:AustralianBureauofStatistics,TheLabourForce,CatNo.6202.0andAustralianBureauofStatistics,1991,1996,2001&2006,CensusesofPopulationandHousing.

The business cycle affects the labour market in terms of both wages andunemployment.Atthebottomofthecycle,businesscycleeffectsareconcentratedontheunskilled,increasingunemploymentandthecompetitionforjobs.Lewis(2006)suggeststheeffectsofarecessionaredisproportionallylargerforunskilledworkerscomparedwith theskilledand thus likely to influence therelative incomebetweenunskilledandskilledlabour.Atthetopofthecycle,asthepoolofunskilledlabourhasdriedup, thebusinesscycleeffectsareconcentratedonwagesputtingpressureonwages to rise. If all other effectswereheld constant itwouldbe expected that,on theonehand, theprivaterateofreturn toadegreewouldbegreatestduringaneconomic downturn as the unemployment rate of the unskilled would increase,relative to university educated, and relativewages of unskilledwould fall. On theotherhand,thereturnstoadegreewoulddiminishatthetopendofthebusinesscycleasunemploymentdecreasesandtherelativewagesoftheunskilledincrease.

Thisstoryiscomplicatedhoweverbythelong-termstructuralchangeswhichhastakenplaceintheAustralianlabourmarketoverthelastfewdecadeschangingthe demand for skills. In 1975, services accounted for just over 50 per cent of alljobs,butby2010theservicesectoraccountedforover70percentofalljobs(ABS,2010).Bycontrast,manufacturing’sshareoftotalemploymentalmosthalvedoverthesameperiod, to less than10percent.Thereweresimilar reductions in therelativesharesofjobsinthe‘industrial’servicessuchaselectricity,gasandwater.Structuralchange has increased the number of jobs in the service sectorwhich values skillsthataredevelopedinuniversitiessuchasknowledge,communicationandcognitiveskills(KellyandLewis,2006).Technologicalchange,drivenbytherapiduptakeofInformationandCommunicationTechnologies(ICTs)investment,affectsthedemandforskillsdirectlybyincreasingdemandforICTrelatedskillsandindirectlythrough

1991

12

4

6

8

11

7

5

Labour Force Survey

10

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

11.6

9.2

7.4

5.2

9

Censuses of Population and Housing

221MICHAEL CORLISS, PHIL LEWIS AND ANNE DALY

The Rate of Returns to Higher Education Over the Business Cycle

theenablingeffectsofICTs(Kelly,2007).KellyandLewis(2010)foundthatwhiletheeconomicrecoveryperiodof1991to1996wascharacterisedbyskillchangesresultingfromstructuralchange,fortherecoveryandboomperiodstechnologicalchangewasthemaininfluenceonskillschange.

While there isaconsiderablebodyofresearchon thereturns toeducation,currentlythereisonlyascantliteratureavailableonthebusinesscycleeffectsupontheprivaterateofreturntohighereducation.Thisisdespitetheexpectationthatthebusinesscyclewouldnothaveneutraleffectsonparticipantsinthelabourmarket,butwouldhavedifferentialeffectsonworkersaccordingtotheirdegreeofskill.Peopledecidingaboutgoingtouniversityneedtobeawareofbusinesscycleeffectontheirexpected wages. Policy makers in choosing the level of public subsidy to degreeeducationneedtobeawarethatthereturnstohighereducationvaryovertimeforanumberofreasonsincludingtheeffectsofthebusinesscycle.

Daly and Lewis (2010) showed that during the recession and subsequentrecoveryupto2001therewasanincreaseintheprivaterateofreturnforuniversitydegrees.However,thisdeclinedduringtheboomin2006.Theysuggestthatitmaybeduetostronggrowthintheeconomyoverthatperiodthatraisedtherelativeincomeof year 12 completers. Corliss andLewis (2011) suggested the business cycle hadlittleimpactontherelativeincomeoftradespersonsasawhole.Howeveritdideffecttradeoccupationsdifferentlyandcreatedgreaterdispersionofincomewithinthetradeoccupationsasthemarketprovidedgreaterrewardsforthemorehighlyskilled,bestpaidtradespersonsduringtheeconomicboom.Otherstudies,inthebusinesscyclescontext,havefocusedonskillswagedifferentials.Forinstance,Reder(1955)examinedthecyclicalityofwagedifferentials,hisfindingssuggestingthat theaggregateskillpremiumwascountercyclicalinthe1930sand1940s.Morerecently,KeaneandPrasad(1993),CastroandCoen-Pirani(2008),andLindquist(2004)showthattheaggregateskill premium is acyclical. Therefore it is of interest to investigate whether thesesamephenomenaareobservablefordegreeholders.Itisthisrelationshipbetweenthebusinesscycleanditsimpactupontheprivaterateofreturntohighereducationthatisthesubjectofthispaper.

2. General issues in estimating the returns to education Asaninvestment,thereturnstoeducationcanbeconsideredfrombothanexanteandexpostposition.Forinstance,ayoungpersoncompletingsecondaryschoolmustconsidertheavailableevidenceonthecostsandbenefitsofeducationbeforedecidingwhethertocontinuetothenextlevel.Partofthereturntocompletingaqualificationistheoptionvalueitcreatesforcontinuingontothenextlevelofeducation(Heckman,LochnerandTodd2005,Wei,2007).Oneoftheeasiestmethodsforassessingthefinancialbenefitsofeducationistolookatthecurrentincomelevelsofpeoplewithgivenqualificationsavailableincrosssectiondata.Thereareanumberofassumptionsthatmustbemadeinordertotaketheincomesofpeoplewithagivenqualificationatallsubsequentagesas theexpected future incomeofaperson justbeginning thatqualification.Currentskilldifferentialsmustbeassumedtopersistintothefuture.Eachagecohortmustalsoembodythesameeducationalqualityandthereisalwaystheissueofheterogeneousreturnstoindividualsandtheroleofuncertaintyinactualoutcomes.

Actual outcomes may differ significantly from expected ones. Heckman,LochnerandTodd(2005)presentevidenceofsubstantialdifferencesintheexante


andexpostreturnstoeducationintheUSA.Daly,FlemingandLewis(2006)showthattheexpostreturnstohighereducationforthosestartingtheirdegreesinAustraliain1986werehigherthanwaspredictedexanteasthereturnstoskillgrewinAustraliaoverthe1990s.Themethodsdescribedbelowhavebeenappliedempiricallytocrosssectiondataandtheassumptionsinvolvedinthismethodologyneedtobeborneinmindwhenconsideringtheresults.Thealternativeofcalculatingtheexpostreturntoeducationattheendofworkingliferequiresover40yearsofdataandisoflittleassistancetosomeoneinmakinganinitialinvestmentdecision.

Another major issue arising in estimating the returns to education is theimpactofanyabilitybiasesonresults.Thedifficultyrestswithmeasuringtheeffectsofabilityonincomeindependentofeducationlevels(Wei,2008).Itisoftenassumedthat natural ability and educational attainment are positively correlated but it ispossiblethatthecorrelationgoestheotherwaywiththemoreableleavingeducationinpursuitof incomeopportunitieswhile thosewith lowerabilitycontinuing in thesystem(Leigh,2008).LeighandRyan(2008)inexaminingthereturnstoschoolingestimatedthatbetweenone-tenthandone-fifthofthereturntoschoolinginstudiesisduetoabilitybias,butthisfindingisofonlylimiteduseinsuggestinganestimateforuniversityeducation.However,manyUSstudiessummarisedbyCard(1999and2001)showthatabilitybiasestoestimatesofthereturntoeducationarenotlarge.

The role of family background in the returns to education has also beendiscussedintheliteratureasitislikelytoconfoundthereturnstoeducation.Weale(1993), forexample, reportsBritish resultswhichshow thatonce theoccupationoftheparentsistakenintoaccount,theprivatereturntoanadditionalyearofeducationislowerforpeoplefromahigheroccupationalbackgroundthanforthosefromlessprivilegedbackgrounds.

Thefinancialreturnstoeducationwillalsobeaffectedbytheextenttowhichindividualsparticipate inpaidemployment.Thishasparticular implicationsfor thereturns toeducation identified forwomenwho typicallyhavesome timeoutof thelabourforcewhiletheyraiseafamily.It isalsoimportantforothergroupssuchasIndigenousAustralianswhomayhaveamarginalattachmenttothelabourforce.Leigh(2008)drawsanimportantdistinctionbetweentheproductivityeffectsofeducationasmeasuredbychangesinthehourlywageandtheparticipationeffectwhichisreflectedinhigherannualincomesasthemoreeducatedarelesslikelytospendtimeoutofthelabourforcethanthelesseducated.

Anothernotionrelevanttothestudyofthereturnstoeducationisthegrowthinparticipationineducation.Withtheincreasingnumbersofuniversitygraduates,itispossiblethattheaveragenaturalabilityofstudentsmayhavedecreasedovertime(Wei, 2008). In this context the distinction between the average return to tertiaryeducationandthereturntothemarginalstudentisimportant.

Theprinciplesofrateofreturnanalysisarewellknowninlaboureconomicsandtheliteraturedatesbacktothelate1950s(Becker,1960;Schultz1961).Theprivaterate of return is the discount ratewhichmakes the net present value (NPV)of aninvestmentzero.Analternative(equivalent)methodistospecifyadiscountrateandcalculatetheNPVofthefutureincomestreamnetofcosts.ThedecisionruleisthentoinvestwheretheNPVispositive.Inthecaseofthereturnstoeducation,theNPVofan investment is thedifferencebetween thediscountedpresentvalueof lifetimemonetarybenefitsfromobtaininganadditionalqualificationoryearsofeducationandthecostsincurredinthisinvestment.



Thismethod calculates an average rateof return for graduates as awhole.Given the heterogeneity among individuals facing different costs and benefits, theaveragemaynotbeagoodguidetotheexpectedrateofreturntoanygivenindividual.Fromapolicymaker’spointofview,itdoesnotaddressthequestionofthereturntoeducationforthelowestpaidgraduate.Thispaperwillattempttothrowlightonthisquestionbyexaminingtherateofreturnforthosegraduatesreceivingrelativelylowincomecomparedtotheaveragegraduate.

Thereremainotherlimitationstothismethodology.Thereisstilltheproblemthatitisdifficulttocapturethe‘pure’effectofeducationonincomebecauseofthecorrelationbetweeneducationaloutcomesandabilityandfamilybackgroundfactors.Factorssuchasthecurrentstateofthebusinesscycle,selfselectionofindividualsintothelevelofeducationthatismostlikelytoproducethehighestreturnsforthemandthelimitationsoutlinedearlierrelatingtotheuseofcrosssectiondataremainimportantconsiderations.

Itispossibletorefinethecalculationstotestdifferentassumptionsthathavebeenmade.DalyandLewis(2010)forexample,presentprivateratesofreturnusingdifferentassumptionsabout the lengthofadegree, student incomewhile studying,the timing ofHECSpayments and the absence from theworkforce bywomen forseveralyearssometimeduringtheircareers.Itwouldalsobepossibletomakesomeadjustmentstotheseestimatesinthelightoftheavailableevidenceontheeffectofabilityonincome.

TherehavebeenanumberofAustralianstudiesoftheprivaterateofreturntoinvestmentinanundergraduatedegreeoverthepastthirtyyears.Differencesintheunderlyingassumptionsandmethodologiesofthesestudiesmakedirectcomparisonsdifficult,butthereisgeneralagreementthatinvestmentinanundergraduateeducationishighlyprofitablefromanindividualviewpoint(Miller,1982;Maglen,1994;Dalyand Jin, 1997; Chapman and Salvage, 1997; Borland et al., 2000; Borland, 2001;Larkins,2001;Dalyet al.,2004;Lewiset al.,2004;LeighandRyan,2008;Leigh2008;DalyandLewis,2010;Wei,2010).

Inthispaperwearespecificallyinterestedintheextenttowhichthereturnstohighlevelsofskillareinfluencedbythebusinesscycle.Withascarcityofresearchonthisitistheaimofthispapertopartlyrectifythis.

3. Assumptions of this study Thereareanumberofkeyassumptionswhichmustbemadeinordertocalculatetheprivaterateofreturn.Theseincludethefollowing:

• The length of the degree and therefore the period for which income isforegone,

• Whetherstudentsearnanyincomewhilestudying,• Thesizeofdirectcostsincludingthepurchaseofbooksandequipmentandfees,

• Thesizeofanyadjustmentforabilityorotherunmeasuredcharacteristicswhicharelikelytoaffecttherateofreturn.

The following discussion outlines the assumptions made in the base casereported here. The results of further calculations to test the sensitivity to theseassumptionswillbereportedlater.Thedatausedfortheincomecalculationsarefrom


the1991,1996,2001and2006CensusesofPopulationandHousing.Thefollowing12fieldsofstudywereidentified:Humanities,Science,AlliedHealth,MathematicsandStatistics, Information Technology, Engineering, Architecture, Medicine, Nursing,Dentistry,EducationandVisualandPerformingArts.

InthebasecaseforaBachelordegree,thelengthofthedegreeisassumedtobe threeyears forHumanities,Science,Mathematics andStatistics, InformationTechnology,NursingandVisualandPerformingArts; fouryears forAlliedHealthand Engineering; and five years for Architecture, Medicine, and Dentistry. Therehavebeensomerecentchangesintheconfigurationofthequalificationsrequiredtopracticeinsomeoftheseprofessions.Medicineisnowapost-graduatequalificationat some universities andArchitecture has changed from a five year undergraduatedegreetoathreeyearundergraduatedegreefollowedbyatwoyearMasters.Asmostofthecurrentholdersofthesequalificationscompletedtheirdegreesundertheearliersystemasundergraduatedegrees,thisassumptionhasbeenappliedtoallgraduatesinthesefieldsofstudy.WhiletheaggregatedgroupsofHumanitiesandSciencearehomogenous in theminimum lengthof time required tocompleteadegree,AlliedHealthcontainssomevariationwithin.Outofthesevensub-fieldsofstudyidentifiedinthiscategory,fourfieldsrequirefouryearstocomplete(Radiography,RehabilitationTherapies, Complementary Therapies, and Pharmaceutical), two fields take threeyearstocomplete(PublicHealthandOtherHealth)andonefieldofstudytakesfiveyearstocomplete(OpticalScience).

Tocarryouttheanalysis,specialtabulationsofmedianincomebyqualificationforeachagebetweenandincluding18to64yearswerepreparedfromthe1991,1996,2001and2006Censuses.Theseincomesarethemediansforallpeoplewithagivenqualification inanagecategoryandall labourmarket statesare included; full andpart-timeemployed,unemployedandnot in the labourforce.They thereforereflectthedifferentemploymentoutcomesforgraduatescomparedwithYear12completersandmalescomparedwithfemalesaswellasanyincomedifferencesthatexistwithinaparticularlabourforcestatus.Asimpleregressionrelatingmedianincometoageandagesquaredwasestimatedtoprovideasmoothseriesformedianincomesateachage.Netincomeswerethencalculatedusingthe1990/91,1995/96,2000/01and2005/06incometaxationrates.

Theresultspresentedherearebasedoncross-sectionaldatafromtheCensusesand,therefore,maketheassumptionthatthefutureincomestreamofan18yearoldchoosingbetweenundertakingadegreeorleavingtheeducationsystemattheendofYear12isbestrepresentedbywhatpeopleofdifferingageswithdifferentlevelsofqualificationsarecurrentlyearning.DuetothechangeintheCensusquestionrelatingtohighestlevelofschoolingattained,the1991and1996‘ageleftschoolequals18’istakentobeequivalentof‘highestlevelofschoolingcompletedequal’sYear12orequivalent’in2001and2006.Theforegoneincomeofundergraduateswereassumedtobethemedianincomeofahighschoolgraduate.

Thebasecaseisthesimplestcasewhichcanbemodifiedtotestthesensitivityofresultstochangesinassumptions.Itassumesthatuniversitystudentsdonotearnincomewhilestudyingandthatthereareexplicitcostsforstudents;undergraduatesincur$1,720peryear(in2006dollars)indirectstudycost.Thesedirectcostestimateswere taken from theAustralianUniversitiesStudentFinancesReport2006 (AVCC2007). In addition students are assumed to pay upfront the student fees applicable



foreachfieldofstudyandyearattherelevantrate(i.e.1991,1996,2001,or2006)oftheHigherEducationContributionScheme(HECS)chargeandcollecta20percentdiscountforpayingfeesupfront.Thebasecaseresultsdonotmakeanyadjustmentforabilityorotherunmeasuredcharacteristicswhichmayaffecttheprivaterateofreturn.

4. The labour market for graduates Table I shows the number of graduates per 1000 population over 15 years of ageincreasingdramaticallyovertheperiod1991to2006withnearlyan80percentincreaseintheproportionofgraduatesfrom87graduatesper1000in1991to156graduatesper1000in2006.Thisincreaseinthenumberofgraduatesper1000populationhappenedatthesametimetherewasconsiderablepopulationgrowthintheorderof22percentduringtheperiod1991to2006.BothofthesefactorshavesignificantlyincreasedthenumberofgraduatesinAustralia.

Holding everything else constant, such a large increase in the number ofgraduates in theAustralianpopulationover 15years of age,wouldbe expected toreducetherelativemedianincomesofgraduatescomparedtothoseholdingayear12certificate.However,thisisnotthecaseasshowninthetable.Infacttheearningsratioofgraduatestoyear12certificateholdersincreasedfrom1991to2001formalesandwhilefemalesdidseeafallfrom1991to1996theearningsdifferentialbecamelargerfrom1996to2001.

Asdiscussedearlierthebusinesscyclecouldpartlyexplainthetrendintheearningsratioofgraduatestoyear12certificateholders.Atthebottomofthecycle,businesscycleeffectsareconcentratedontheunskilled,increasingunemploymentandthecompetitionforjobsreducingrelativewages.Atthetopofthecycle,asthepoolofunskilledlabourhasdriedup,thebusinesscycleeffectsareconcentratedonputtingpressureonrelativewagesoftheunskilledtorise.Table1showsaconsiderablefallin the earnings ratio forbothmale and femalegraduates from the2001expansionto2006topof thecycle.Whileforfemalesthehighestearningsratioisduringthe1991recessionaspredicted,formalestheearningsratioofthe1991recessionisnotthehighestpointand in fact theearnings ratiogrowsupuntil its2001peakat theexpansion phase of the business cycle.No doubt there have been cyclical changesalongsidelongertermstructuralandtechnologicalchanges.

Table 1- Summary statistics on the graduate labour market in Australia

1991 1996 2001 2006Population15+ 13,085,092 14,040,303 14,856,774 15,918,082Graduates/1000population 87 103 146 156MalesMedianincome-graduates $47,273 $53,159 $67,729 $62,921Medianincome-year12 $25,978 $28,598 $34,148 $42,356Ratiograduate/year12 1.82 1.86 1.98 1.49FemalesMedianincome-graduates $34,117 $37,632 $46,108 $42,259Medianincome-Year12 $18,410 $21,894 $25,765 $27,164Ratiograduate/year12 1.85 1.72 1.79 1.56

Source:Dalyet al.,(2006);AustralianBureauofStatistics,1991,1996,2001,2006CensusesofPopulationandHousing,unpublished.


However,asstatedbefore,thegraduatelabourmarketisfurthercomplicatedby structural and technological change taking place during the same period. Asdiscussedearlier,bothstructuralandtechnologicalchangehaveincreasedthedemandforgraduatesinAustralia.

5. Results Webeginbycomparingtheratesofreturnformalesandfemalesinthemostrecentyearforwhichwehaddata,2006,toprovideabenchmarkforlookingatchangesovertime.Table1showsquiteclearlythereissubstantialincentiveforprivateindividualstoattainaBachelordegreeinmostdisciplinesandthatthereisconsiderablevariationbetweendifferentfieldsof study. Italsoshows that thereare importantdifferencesin ratesof returnaccording togender.The rateof return in2006 forallgraduateswas12percentforfemalesand15percentformales.Clearly,failuretoaccountforthesedifferencescanbeverymisleading.Thehighest ratesof returnformenwereinDentistry,Nursing,andInformationTechnology.Forwomenthehighest ratesofreturnwereinDentistry,Medicine,andInformationTechnology.ThelowestratesofreturnforbothmalesandfemaleswereintheVisualandPerformingArts(negativerateofreturn),HumanitiesandArchitecture.Inmostcasesbutnotall,theprivateratesofreturnwereslightlyhigherformalesthanforfemales,mostnotablyinDentistryandNursing.OntheotherhandfemaleswithaHumanitiesqualificationreceivedaprivaterateofreturnofninepercentcomparedtothreepercentfortheirmalecounterparts.

Figure 2 - Private rates of return by gender, 2006

Source:AustralianBureauofStatistics,2006CensusofPopulationandHousing,unpublished.

Total

Nursing

Mathematics and Statistics

Humanities

Allied Health

Architecture

0%

Sciences

Medicine

Engineering

Female Male

Information Technology

Visual and Performing Arts

Dentistry

Education

5% 10% 15% 20%



6. Business cycle effects Turningnowto theeffectsof thebusinesscycle, theCensusesprovideuniquedatacoveringdifferentstagesofthebusinesscyclefrom1991,themiddleofarecession,1996,therecovery,2001,theboomand2006,labourshortages.Figures3and4showtheprivateratesofreturnformalesandfemalesoverthebusinesscycle.

Figure 3 - Private rates of return for males, 1991–2006

Source:AustralianBureauofStatistics,1991,1996,2001,2006CensusesofPopulationandHousing,unpublished.

Theprivaterateofreturnformales,showninfigure3,experiencedadramatic

increasebetweenthe1996recoveryto2001expansionphaseofthebusinesscycleformanyofthefieldsofstudyunderinvestigation.Duringthistimetherateofreturnforthewholesampleincreasedfrom11percentin1996to15percentin2001.Thiswasmost pronounced inNursing, Education andMathematics.Other degrees recordedmoremodestincreasesintheprivaterateofreturn,withtheexceptionofMedicineobservingaslightdecrease.WiththeexceptionofInformationTechnology,wheretherateofreturnfellsharplybetween2001to2006,allthefieldsofstudyrecordedanupwardmovementorremainedrelativelyconstantduringthe2001expansionto2006boomphaseofthebusinesscycle.Whilethereisconsiderablevariationintheprivaterateofreturnbetweenthedegreesshownhere, thebusinesscycleappears toaffecttheminafairlyconsistentmanner.

Figure4 suggests a slightlydifferent story forwomen than for theirmalecounterparts over thebusiness cycle.Consistentwithmales, femaledegreeholdersincreased their private rate of return from 1996 to 2001, although it was not aspronouncedasformales.Thesampletotalrateofreturnincreasedfrom12percent

Total


Architecture


Engineering

Dentistry

0%


Education

Nursing

1991

Medicine

Humanities

Sciences

Allied Health

5% 10% 15% 20% 25%

1996 2001 2006


in1996to14percentin2001.However,forfemales,therateofreturnfellbetween2001and2006back to the1996 levelunlikemaleswhichonlysawaslight fall inrateofreturnoverthisperiod.ThefieldsofstudyhavingthegreatestincreasesintheprivaterateofreturnforfemaleswereVisualandPerformingArtsandEngineering.A further differencewas that the downward adjustment in the female private rateofreturnoccurringbetween2001and2006wasmoredramaticthanfortheirmalecounterparts and affected more fields of study than just Information Technology,includingMathematics andEducationwithminor reductions noted in some of theotherfieldsofstudy.

Figure 4 - Private rates of return for females, 1991-2006


Perhapsthemostinterestingresultsarethedramaticvariationintheprivaterate

ofreturnforholdersofInformationTechnologydegrees.Itispossiblethattheraisedexpectationsofpotentialstudents in2000due to theveryhighprivaterateof returnof Information Technology degrees compared with other fields of study created anoversupplyofInformationTechnologygraduates.Thisoccurredatatimewhendemandwascomingoffahighin2006puttingdownwardpressureonwagesthusdramaticallyreducingtheprivaterateofreturnforInformationTechnologydegreeholders.

7. Low income degree holders Sofarthefocushasbeenonthemedianincomeofindividualsattainingabachelordegreetounderstandthereturnstoinvestmentineducationandtrainingforthetypicaldegreeholder.Wenowexaminethereturnsforthosegraduateswhoarelesswellpaid,specificallythelowest20percentofgraduates(20thpercentile).Hereasinthebase

Total


Architecture


Engineering

Dentistry

0%


Education

Nursing

1991

Medicine

Humanities

Sciences

Allied Health

5% 10% 15% 20% 25%

1996 2001 2006



casetheopportunitycostofstudyistakentobethemedianincomeoftheYear12certificateholders.Inthisparticularcasetheresultsoftheprivaterateofreturnarenearlyallnegativeand,therefore,havebeenomittedfromthepaper.Thatis,thelowestpaiduniversitygraduateswouldbebetteroffnotgoingtouniversity,inallstagesofthebusinesscycle.

However, the net present value can be calculated and the estimates arepresentedinfigures5and6.Asexplainedearlierinthepaperthenetpresentvalueiscalculatedwitharealrateofinterestoftwopercentwhichisapproximatelythelongrunrealbondrate.Theconclusionsarenotchangedmuchbythechoiceofinterestratewithinreasonablebands.

Clearly formost of theworst paid bachelor degree holders the net presentvalueoftheirlifetimeincomewouldbegreaterleavingschoolaftergraduatingfromYear12andobtainingfull-timeemployment.Asshowninfigure5thesampletotalsuggestsin2006theworstpaidgraduatesare$132,210worseoffcomparedtothosewithonlyayear12certificate.However,thislossinlifetimeincomehasbeenreducedconsiderably from its peak in1996of$253,457. In2006 formen inMedicine andDentistry, however, the rewards for the lowest paid obtaining a degree were stillworthwhileandforlowerpaidwomenuniversitygraduatesthereisstillincentivetostudyMedicine,DentistryandEducation.

Figure 5 - Net present value of the 20th percentile of males, 1991-2006


Total


Architecture


Engineering

Dentistry

-$600


Education

Nursing

1991

Medicine

Humanities

Sciences

Allied Health

-$400 -$200 $0 $200 $400

1996 2001 2006

Thousands


Figure 6 - Net present value of the 20th percentile for females, 1991-2006


Over time the incentive for the worst paid degree holders remains fairly

consistent.Menhadasmallbutpositivenetpresentvaluein2001andfemalesobservedasmallbutpositivenetpresentvaluein1991and2001,however,thenetpresentvaluedoesnotremainpositivefortheentireperiodunderquestion.AdditionallytheonlypositivenetpresentvalueforwomenstudyingEducationtookplacein2006.Theseresults suggest that for the lowest paidmen and women graduates onlyMedicineconsistentlyoffersanyincentivetoundertakethetrainingovertheentireperiod.If,asapotentialuniversitystudent,youknewyouweregoingtobea lowearner thenyouwouldbebestnottostudyauniversitydegree.Thereisabigriskinvolvedwithuniversity education including thepossibilityyoumightnot complete and thatyoumightnotearnmuchasaconsequenceofcompletion.

9. Degree holders employed full-time Inthebasecaseeveryoneisincludedintheprivaterateofreturnestimatesnomatterhowmuch theywork or even if they do notwork at all. There are two effects ofhigher education for individuals - they are higher wages and lower incidence ofunemployment.Bylookingatonlythoseworkingfull-timetheunemploymenteffectcanbeheldconstantandtheeffectuponwagesisolated.Forpeoplefromsomefieldsofstudyunemploymentoragreaterincidenceofcasualworkmaysignificantlyaffectthereturnstoadegree.Also,many(perhapsmost)womenleavethelabourforceforperiodsoftimetolookafterchildren.Thereforeitmaybeusefultocontrolforthisby

Total


Architecture


Engineering

Dentistry

-$400


Education

Nursing

1991

Medicine

Humanities

Sciences

Allied Health

-$200 -$100 $0 $100 $300

1996 2001 2006

Thousands-$300 -$200



selectingonlythoseworkingfulltime.Boththeuniversitygraduatesgroupandthecomparatorgroup(Year12graduates)onlyincludethoseworkingfull-time.Figures7and8showtheprivaterateofreturnestimatesforindividualswithabachelordegreeandwhoareemployedfulltime.Thebusinesscycleeffectswouldbeexpectedtobegreaterwhenexaminingallgraduatesratherthanthoseemployedfulltime.

Overall the recession through to theboomphaseof thebusiness cyclehasincreasedtherateofreturntomaledegreeholdersthatareworkingfulltimeasshowninfigure7.Thelargestchangeoccurredduringfrom1996(recovery)to2001(expansion)phaseofthebusinesscyclewithanincreaseintherateofreturnforthesampletotaloffivepercentagepoints.Forthevariousfieldsofstudy,themostremarkableoftheseincreaseswasfoundintheInformationTechnology,AlliedHealth,ArchitectureandMathematicsupfivepercentfrom1996to2001.Thereweresomenotableexceptionstothisthough-thosewithanInformationTechnologydegreerecordedasignificantreductionintherateofreturntotheirdegreedroppingoffourpercentagepointsfrom2001to2006despitetheirimpressivegaininthepreviousfiveyearperiod.Theresultsimplythatthebusinesscyclehasagreateffectonwagesnotjustemployment.

Figure 7 - Private rates of return for full-time workers, males, 1991-2006


Forfulltimefemaledegreeholdersshowninfigure8therecessionthroughto

boomphaseofthebusinesscyclehadasimilareffectontheratesofreturnforfemaledegreeholdersasexperiencedbymaledegreeholders.Thelargestchangeoccurredfrom1996(recovery)to2001(expansion)phaseofthebusinesscyclewithanincreasein therateofreturnfor thesample totalof threepercent.For thevariousfieldsofstudy, themost remarkable increases in the rateof returnwere for thoseholdinga

Total


Architecture


Engineering

Dentistry

0%


Education

Nursing

1991

Medicine

Humanities

Sciences

Allied Health

5% 10% 15% 20%

1996 2001 2006


degreeinInformationTechnologyandEngineering.Although,therewasafallintheprivaterateofreturnforInformationTechnologydegreeholdersfrom2001to2006,formostoftheotherfieldsofstudy,theyobservedafurtherincreaseintheirprivaterateofreturnfrom2001to2006.Thisisincontrasttotheeffecttheboomhadforallwomenwhereareductionintheprivaterateofreturnwascommonplaceamongthevariousfieldsofstudy.

This can be explained by distinguishing between salary effects andemploymenteffects.Ontheonehand,whenexaminingwomenworkingallhoursaboomintheeconomyfavouredpart-timeemploymentforwomencausingtherelativewages of unskilled women to increase. On the other hand, when examining onlywomenworkingfull-time,employmenteffectsareeliminatedandonlyapuresalaryeffectremains.Thus,duringtheboomtherewasafallintheprivaterateofreturnforwomendegreeholderswhenexaminingallwomen,andanincreaseintheprivaterateofreturnforwomendegreeholderswhenconsideringonlywomenworkingfull-time.

Interestingly, controlling for hours worked had a much greater impact inreducing the private rate of return for females than it did onmales. This suggestsunemploymentandchildraisingimpactsuponfemaleincomeconsiderablymorethanitdoesonmales.

Figure 8 - Female private rates of return for full-time workers


Additionallyaftercontrollingforhoursaconsiderablereductionintherates

ofreturntodegreeholderswasobserved.Itiswellknownthattherearesignificantlydifferentunemploymentratesforpeoplewithdifferentqualifications(seeforinstance,Lewis, 2006). Those holding a degree have on average a considerably reduced

Total


Architecture


Engineering

Dentistry

0%


Education

Nursing

1991

Medicine

Humanities

Sciences

Allied Health

5% 10% 15% 20%

1996 2001 2006



incidenceofunemploymentwhencomparedwiththosewhosehighestqualificationisayear12certificate.Figures7and8areperhapsevidenceofthis.Itwouldbeexpectedthatthosewithayear12certificatehavehigherratesofunemploymentcomparedtodegreeholders.Thisunemploymenteffectwouldcause themedian incomeofyear12certificateholders tobe loweredbymore thandegreeholders.By removing theunemployedfromthesamplereducestheimpactunemploymenthasonincome.Asexpectedtheresultsshowareducedrateofreturntodegreeholderswhensamplingonlythoseworkingfulltime.Thereisevidencethatwomenshiftbetweennotinthelabourforceandpart-timeemploymentwhenthelabourmarketisgood(Norriset al.,2004).Alsoyouwouldexpectareductionintherateofreturnwhenmoreemploymentopportunitiesareavailable.

10. Conclusion Thispaperpresentsestimatesof therateofreturntoBachelordegreesinAustraliaovereachphaseofthebusinesscycle.Theresultsshowthat,generally,adegreeisaprofitableinvestmentforindividuals.Variousassumptionshavebeenadoptedinordertoprovideasensitivityanalysisoftherobustnessoftheseestimatesandimportantlywhetherornottheseresultsremainconsistentovertime.

While returns to university education are generally high, there are largedifferences between disciplines. For instance, the rates of return to a Bachelordegree are particularly large in some fields of study notably Dentistry, Medicineand InformationTechnology.The results show that thecompletionofaVisualandPerformingArtsdegreeisnotagoodfinancialinvestmentforeithermenorwomen.Also,notallstudentsreceiveconsiderablebenefitsfromobtainingauniversitydegree.For instance,manyof the lowestpaiduniversitygraduateswouldhavedonebetterfinishingtheireducationatYear12andenteringfull-timeemployment.Theresultsshownherealsosuggestthatthisconclusionisrobustovertime.

Itisinterestingthattheprivaterateofreturnwasatitslowestduringthebottomof the business cycle. This is against expectations as typically the contractionaryphaseofthebusinesscyclewouldbeexpectedtodisproportionatelyimpactuponthelowerskilled(Year12certificateholders)intermsofbothemploymentstatusandthemedianwagethusdecreasingtheopportunitycostofstudyandincreasingtheprivaterateofreturn.ItisalsopossiblethatthelargeincreaseinsupplyofgraduatelabourinthepostDawkinseracoincidingwiththedownturntosomeextentoutweighedthebusinesscycleeffect.

Additionally, itwouldbeexpected that at the topof thebusiness cycle theprivate rate of return to a degree would diminish as the pool of unskilled labourdried up alongwith a decreasingunemployment rate puttingpressure onwages torisedisproportionatelyfasterfortheleastskilled.Anotherreinforcingfactorexpectedtoreduceratesofreturntouniversityeducationwouldhavebeenthehugegrowthinsupplyofuniversitygraduatespost1990.ThisisonlyreallyevidentforInformationTechnology degreeswhere a dramatic increase in the private rate of return in theprevious periodmay have caused an oversupply of undergraduates by 2006whichresultedinalargefallintheprivaterateofreturnduringtheboomperiod.Formostdegreestheimpactofthebusinesscycleonratesofreturnhavebeenpro-cyclical.


References Arrow,K.J.(1973),‘HigherEducationasaFilter’,Journal of Political Economy II,

July,193-216.Ashenfelter, O. and Krueger, A. (1994), ‘Estimates of the Economic Return to

SchoolingfromaNewSampleofTwins’,The American Economic Review,84(5),1157-1173.

AustralianViceChancellorsCommitteeAVCC(2007),Australian University Student Finances 2006,AVCC,Canberra.

Becker,G.(1964),Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education,ColumbiaUniversityPress,NewYork.

Borland,J.,Dawkins,P.,Johnson,D.andWilliams,R.(2000),Returns to Investment in Higher Education,UniversityofMelbourne,availableonthewebatwww.melbourneinstitute.com

Borland, J. (2001), New Estimates of the Private Rate of Return to University Education in Australia, University ofMelbourne, available on the web atwww.melbourneinstitute.com.

Card,D.(1999),‘TheCausalEffectsofEducationonEarnings’,inO.AshenfelterandD.Card(eds.)Handbook of Labor Economics,NorthHolland,Amsterdam.

Card,D.(2001),‘EstimatingtheReturnstoSchooling:ProgressonSomePersistentEconometricProblems’,Econometrica,69(5),1127-1160.

Castro,R.andCoen-Pirani,D.(2008),‘WhyHaveAggregateSkilledHoursBecomeSoCyclical Since theMid-1980s?’, International Economic Review, 49(1),135-185.

Chapman, B. and Salvage, T. (1997), The Consequences of Recent Changes in Financing for Australian Higher Education,mimeo,Centre forEconomicPolicyResearch,AustralianNationalUniversity.

Daly, A., Fleming, D. and Lewis, P.E.T. (2003), ‘The Labour Market for LegalProfessionals’,Law Institute Journal.

Daly,A,FlemingDandLewisP.E.T.(2004),‘Isalegaleducationabetterinvestmentthananeconomicsdegree?’,Australasian Journal of Economics Education,1(2),183-197.

Daly,A.,FlemingD.andLewisP.E.T.(2006),‘ACohortAnalysisofthePrivateRateofReturn toHigherEducation inAustralia’,Australian Journal of Labour Economics,9(3),257-268.

Daly,A.andJin,L.(1997),‘EstimatingthePrivateRateofReturntoEducationforIndigenousAustralians’,Australian Economic Papers.

Daly,A.andLewis,P.(2010),‘ThePrivateRateofReturntoanEconomicsDegreeinAustralia:anupdate’,Economic Papers,29(3),353–364.

Haveman,R.H.andWolfe,B.L.(1984), ‘SchoolingandEconomicWell-Being:TheRoleofNonmarketEffects’,Journal of Human Resources,19(3),377-407.

Heckman,J.,LochnerL.andToddP.(2005),‘EarningsFunctions,RatesofReturn,andTreatmentEffects: theMincerEquation andBeyond’,NBERWorkingPaper11544availableathttp://www.nber.org/papers/w11544

Holzer, H., Raphael, S. and Stoll, M. (2006), ‘Employers in the Boom: How DidHiringofLess-skilledWorkersChangeDuring the1990s?’,The Review of Economics and Statistics,88(1),283-299.



Katz,L.andAutor,D.(1999),‘ChangesintheWageStructureandEarningsInequality’,Handbook of Labor Economics,3,Chapter26,1463-1555,NewYork:North-Holland.

Kelly,R.andLewis,P.(2010),‘TheChangeinLabourSkillsOvertheBusinessCycle’,Australian Bulletin of Labour,36(3),260-277.

Keane,M.andPrasad,E.(1993),The Relation Between Skill Levels and the Cyclical Variability of Employment, Hours and Wages, Working paper 93/44,InternationalMonetaryFund.

Larkins,F.(2001),‘TheEconomicBenefitsofAustralianUniversityDegrees:BachelorandResearchHigherDegrees’,Australian Economic Review,34(4),403-14.

Leigh,A.(2008),‘ReturnstoEducationinAustralia’,Economic Papers,27(3),233-249.

Leigh, A. and Ryan, C. (2008), ‘Estimating Returns to EducationUsingDifferentNatural Experiment Techniques’,Economics of Education Review, 127(2),149-160.

Lemieux,T.(2006),The“MincerEquation”ThirtyYearsAfterSchooling,ExperienceandEarnings,inJacobMinceraPioneerofModernLabourEconomics2006,PartIV,127-145.

Lewis,P.,Daly,A.andFleming,D.(2004),‘Whystudyeconomics?Theprivaterateofreturntoaneconomicsdegree’,Economic Papers,23(3),234-243.

Lewis,P.(2006),MinimumWagesandEmployment,AustralianFairPayCommissionResearchReport1/06.

Lewis,P.,Garnett,A.,Hawtrey,K.andTreadgold,M.(2010),The Australian Economy: Your Guide,PearsonEducation,Sydney.

Lindquist,M.(2004),‘Capital-SkillComplementarilyandInequalityovertheBusinessCycle’,Review of Economic Dynamics,7,519.540.

Maglen, L. (1994), ‘Education Expansion and the Private Returns to a UniversityDegree’,Economic Papers,13(4),57-71.

Miller,P.(1982),‘TheRateofReturntoEducation:Evidencefromthe1976Census’,Australian Economic Review,15,3rdquarter,23-32.

Miller,P.Mulvey,C.andMartin,N. (1995), ‘WhatDoTwinStudiesRevealAbouttheEconomicReturns toEducation?AComparison ofAustralian andUSFindings’,American Economic Review,85(3),586-599.

Mincer,J.(1974),Schooling, Experience and Earnings,NationalBureauofEconomicResearch,NewYork.

Murphy,K.M. andWelch, F. (1990), ‘EmpiricalAge-earnings Profiles’, Journal of Labor Economics,8(2),202-229.

NATSEMandAMP(2008),WhatPricetheCleverCountry?TheCostsofTertiaryEducation in Australia, AMP.NATSEM Income and Wealth Report 21availableatwww.amp.com.au/ampnatsemreports

Norris,K.,Giles,M.andKelly,R.(2004),Economics of Australian Labour Markets,6th(ed.),PearsonEducation,French’sForest,NSW.

Polachek,S.(2007),EarningsovertheLifecycle:TheMincerEarningsFunctionandItsApplications,discussionpaperno.3181.

Psacharopoulos, G. (1973), Returns to Education: an International Comparison,Elsevier.


Psacharopoulos, G. (1981), ‘Returns to Education: an Updated InternationalComparison’,Comparative Education,17(3),321-341.

Psacharopoulos,G.(2009),ReturnstoInvestmentinHigherEducation:AEuropeanSurvey,aContributiontotheHigherEducationFundingReformProjectfortheEuropeanCommission,ledbyCHEPS.

Psacharopoulos,G.andPatrinos,A.(2004),‘ReturnstoInvestmentinEducation:AFurtherUpdate’,Education Economics,12(2),111-135.

Reder,M. (1955), ‘TheTheoryofOccupationalWageDifferentials’,The American Economic Review,45(5),833-852.

Schultz,T.(1961),‘InvestmentinHumanCapital’,The American Economic Review,51(1),1-17.

Weale M. (1993), ‘A Critical Evaluation of Rate of Return Analysis’, Economic Journal,103(May),729-737.

Wei.H.(2007),Measuring Option Values and the Economic Benefits of Completing Secondary Education, Methodology Advisory Committee Papers cat. no.1352.0.55.082,AustralianBureauofStatistics,Canberra.

Wei,H.(2008),‘MeasuringHumanCapitalFlowsforAustralia:ALifetimeLabourIncome Approach’, Methodology Advisory Committee Papers cat. no.1351.0.55.023,AustralianBureauofStatistics,Canberra.

Wei,H.(2010),‘MeasuringEconomicReturnstoPost-schoolEducationinAustralia’,Research Paper cat. no. 1351.0.55.032, Australian Bureau of Statistics,Canberra.

Weisbrod,B.A.(1964),External Benefits of Education,PrincetonUniversity,IndustrialRelationsSection.

Documents

The Rate of Return to Higher Education Over the Business Cycle · business cycle and its impact upon the private rate of return to higher education that is the subject of this paper