Forschungsinstitut zur Zukunft der ArbeitInstitute for the Study of Labor

DI

SC

US

SI

ON

P

AP

ER

S

ER

IE

S

The Impact of Family Friendly Policies in Spain and their Use throughout the Business Cycle

IZA DP No. 9550

December 2015

Sara De la RicaLucía Gorjón

The Impact of Family Friendly Policies in Spain and their Use throughout the

Business Cycle

Sara De la Rica University of the Basque Country, FEDEA

and IZA

Lucía Gorjón

University of the Basque Country and FEDEA

Discussion Paper No. 9550 December 2015

IZA

P.O. Box 7240 53072 Bonn

Germany

Phone: +49-228-3894-0 Fax: +49-228-3894-180

E-mail: iza@iza.org

Any opinions expressed here are those of the author(s) and not those of IZA. Research published in this series may include views on policy, but the institute itself takes no institutional policy positions. The IZA research network is committed to the IZA Guiding Principles of Research Integrity. The Institute for the Study of Labor (IZA) in Bonn is a local and virtual international research center and a place of communication between science, politics and business. IZA is an independent nonprofit organization supported by Deutsche Post Foundation. The center is associated with the University of Bonn and offers a stimulating research environment through its international network, workshops and conferences, data service, project support, research visits and doctoral program. IZA engages in (i) original and internationally competitive research in all fields of labor economics, (ii) development of policy concepts, and (iii) dissemination of research results and concepts to the interested public. IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character. A revised version may be available directly from the author.

IZA Discussion Paper No. 9550 December 2015

ABSTRACT

The Impact of Family Friendly Policies in Spain and their Use throughout the Business Cycle*

The aim of this paper is to evaluate various aspects of a family friendly law (Act 39/99) approved in Spain in 1999, which grants parents the right to reduce their work-time schedule for childcare issues. Moreover, those who resort to that law enjoy higher protection against dismissal than other workers, which may encourage workers to use the law as a form of job protection, particularly in recession periods. We first test the direct impact of the law, i.e. whether its passing increased the use of work-time reduction for childcare issues. We find that the use of such measures did indeed increase by around 18%. Second, we test whether the passing of the law led to strategic behavior from employers, in the sense of restricting indefinite contracts to potential users of the law to limit the use of work-time reduction. We find that this is indeed the case. Finally, we test whether this law was resorted to more or less in the recent downturn than in the previous economic upturn. We find that its use decreased by around 13% in the recent downturn, which is not consistent with the view that workers would use the law more intensively during recessions for job protection reasons. JEL Classification: C23, J16, J18, J62 Keywords: policy evaluation, family friendly policies, difference-in-difference, business cycle Corresponding author: Sara de la Rica University of the Basque Country Facultad de Ciencias Económicas y Empresariales Avenida Lehendakari Aguirre, 83 48015 Bilbao Spain E-mail: sara.delarica@ehu.es

* The author acknowledges comments and suggestions from participants in the seminar at Universidad Carlos III and from participants in the XX Symposium on Economic Analysis, 2015. The author acknowledges financial support from the Spanish Ministry of Education and Science (ECO2012-35820) and from the Spanish Institute for Women (nº 2011-0004-INV-00081).

2

1.INTRODUCTION

Womenhavemadehugeprogress intheworkplace,especially inthemore industrializedcountries.Goldin(2004)referstothemassincorporationofwomenintotheworkforceduringtheseventiesasthe“quietrevolution”.However,inspiteofthisrevolutionaryprocess,genderdifferencesstillpersist.

Family issuesplayacrucial role inunderstanding thegenderdifferencesobserved in thelabor market. Women combine employment with home responsibilities to a much largerextent than their male partners. This is particularly so in some countries, such as those ofSouthernEurope,becauseofthelackofaccesstoproperchildcareprovisions(DelBoca2002),low levels of participation by men in household tasks (Bettio and Villa 1998; De Laat andSevilla-Sanz2011)and/orlowlevelsofsocialassistance(Adserà2004).Thisgenderasymmetryin reconciling family life and work affects women’s decisions with respect to labor supply,human capital accumulation, and hence their labor-market performance (see Ahn andMira2001;Bertola,Blau,andKahn2007;Adserà2005;DelaRicaandIza2005,amongothers).

The decrease in the working-age population and the high education level achieved bywomeninthepastfewdecadesmakeitessentialtoconsiderwomenasafundamentalpartoftheworkforce.Governmentsand institutionscanplayan importantrole increatingthe legalframeworkforimprovingwomen’schoicesandtheirparticipationintheeconomy,aswellasinhelping societies to break away from themore traditional gender role attitudes that affectwomen’s behavior in many countries. Indeed, in the past few decades many governmentshaveadoptedpoliciesaimedatpromotinggenderequalityandequity in theworkplace.ThispaperseekstoevaluateonesuchpolicyimplementedinSpainin1999.

Thepolicyunderanalysis, calledLaw39/99,wasparticularlyaimedatgivingparentswithchildrenunder6yearsofagetherighttoreducetheirworkschedulewithanequivalentwagereduction1.Thespiritofthislawistomakeitmoreaffordableforparentstostayinthelabormarketandtakecareoftheirchildrenbyreducingtheirworkschedules.2Furthermore,asexplainedbelow,workerswhoavailthemselvesofthispolicyenjoymoreprotectionagainstdismissalsthantherest.Thismayundoubtedlyentailsomeundesirableeffectsthatthepolicymakersdidnotexpect.Thoseeffectsarealsoaddressedinthepaper3.

1Relatedfamily-friendlylawswerealsoimplementedelsewhereinEurope.AustrianLawNr.38/2004hasthesamespiritbutonlyaffectsworkerswithmorethan3yearsoftenureandfirmswithmorethan20employees.Franceimplementeda“supplementaryworkchoicebenefit”lawin2004,providingabenefitthatcanbepaidoutasfromthebirthofthefirstchildforamaximumperiodofsixmonthsatafullorareducedrate,i.e.womencanworkpart-timeandreceivethebenefit.2Althoughthisrightisinprinciplegrantedtobothfathersandmothers,giventhatwomenaretraditionallymainlyresponsibleforchildcarewithincouples,mothershaveinfactbeentheonlyusersofthepolicy.3TheeffectsofthislawwerealsoaddressedinRodriguez-PlanasandFernández-Kranz,2011.Differencesbetweenthetwoapproacheswillbementionedinthecourseofthepaper.

3

Theaimofthepaperistwofold:First,weevaluatetheimmediateimpactofthelaw,inparticularitsdirectandindirecteffects.Bydirecteffectswemeantheextenttowhichthelawhasledtoanincreaseinpart-timeworkingamongparentswithchildrenagedunder6.Withrespect to the indirect effects, we explorewhether employers behave strategically towardspotentialusersofthelawinthefollowingsense:Ifthefamilypolicyiscostlyforfirmsandinaddition its users are more protected against dismissals than other workers some reactionfrom employers might be expected. In particular, they might tend to reduce the indefinitehiringofpotentialusersof the lawand insteadoffer fixed-termcontracts–whosepotentialcosts are much smaller4. Hence, the question to be answered with regards to this indirecteffect iswhetherthelawincreasedtheprobabilityofbeinghiredunderafixed-term(ratherthananindefinite)contractforpotentialfuturelawusers.Toanswerthesetwoquestions,wefocus on salaried employees using the Spanish Current Population Survey (SCPS), themostrepresentative cross sectional sample of the labor force in Spain. We use individualinformationonworkinghoursandcomparetheuseofreducedhoursamongworkersaffectedbythelaw(treatmentgroup)withthosenotaffectedbyit(comparisongroup)justbeforeandafterthepassingofthelaw(diff-in-diffstrategy). The second aim of the study is to characterize the users who have resorted to thefamily friendly policy since its approval (1999) andmeasure the extent to which the GreatRecession has led to a change in the number of policy users and in their personal and jobprofiles.Inprinciple,itmightbeexpectedthatinarecessionframeworkuncertaintyaboutthepossibilityofbeing laidoffmight lead toan increase in theuseofwork time reductionasameasureforprovidinghigherjobprotectionagainstdismissal.Butworktimereductionentailsa proportional reduction in wages, and this negative effect on income is likely to bemoreimportant in a recession context. In addition, fear of reprisals at work during a period ofeconomic instability for the firm might also discourage workers from requesting workreductions for childcare issues. To detect individuals who make use of the law, i.e., whochangetheirtimescheduleforchildcareissues,weneedtofollowworkersovertime.WeusearichlongitudinaldatasetobtainedfromSpanishSocialSecurityrecords(ContinuousSampleofWorkHistories(CSWH))thatcoversworkers’employmenthistoriesandcensusregistrationdata including family characteristics. The dataset contains information on personal and jobcharacteristicsbeforeandafterworkershavechildrenandthusdetectsactualusersofthelawandtheirprofiles.

The first part of the paper is close to that of Rodriguez-Planas and Fernandez-Kranz(2011)althoughthemethodologydifferssomewhat, inparticularwithrespect to thecontrolgroupsused.However,tothebestofourknowledge,therearenostudieswhichcomparetheuseoffamilyfriendlypoliciesindifferentphasesofthebusinesscycle. Ourresultsindicate,inthefirstplace,thatthelawincreasedthelikelihoodofworkingpart-timeforeligiblemothers–i.e.motherswithchildrenunder6–by18%comparedtothe4Workerswithfixed-termcontractsmaynothavetheircontractsrenewedonexpiry,sonojustificationfordismissalisrequiredandthecostissmall–12daysperyearworked.Thisissoalsoforfixed-termworkerswhowouldotherwisemeettherequirementsforusingthefamilyfriendlylaw.

4

similar non-target comparison group. Furthermore, the law applied only to mothers withindefinitecontracts.Inaddition,wefindnegativeindirecteffectsforpotentialusersofthelaw:Whencomparinghiringpracticesasregardspotentialusersofthelaw(targetgroup),i.e.non-mothersofchildbearingage,withasimilarnon-targetcomparisongroup(non-fathersoffertileage)wefindthatthelawincreasedtheprobabilityofthetargetgroupbeinghiredunderfixed-termcontractsbyalmost18%.Third,the2008-2013recessionreducedtheprobabilityofthelawbeingresortedtobyabout13%.

Thepaperisorganizedasfollows.Section2describestheSpanishLaw39/1999.Section3presentsthedatabases(SCPSandCSWH).Section4analyzestheeffectsofthefamily-friendlypolicyonemploymentoutcomesfortheeligiblepopulationandtheunintendedeffectsofthelawafter its implementationon thenon-eligiblepopulationof thepolicy.Section5presentsprofilesofusersandoutlinestheimpactofthe2008inSpainontheuseofthelaw.Section6sumsupandconcludes.

2. FAMILY FRIENDLY POLICY (LAW 39/99) – REDUCED WORK SCHEDULE FORCHILDCARE OnNovember6th1999theSpanishgovernmentpasseda lawwhichgrantedworkingparentswithchildrenunder6yearsoldtherighttoreducetheirworkschedulestoreconcileworkand family life5.Thework timereductiongrantedranges fromone third toonehalfoftheusualfull-timeschedule,withanequivalentwagereduction6.Workersalsohavetherighttochoosethetimeslotduringthedaywhentheywanttowork.Thefirmmusteitheragreeorgotocourt.Underthislawworkerdismissalsforanyreasonrelatedtopregnancy,maternityorpaternity leave, and child-care were declared “unfair”. When employers face “unfair”dismissals for reasons other than pregnancy,maternity or childcare they can solve thembyeither paying the severance payments stipulated for unfair dismissals (45 days per yearworkedatthetimeofthepassingofthelaw)orbyre-hiringtheworker.Inpracticallyallcasesfirms pay the severance payment and do not re-hire the worker7. However, if dismissalsrelated with pregnancy, maternity or childcare issues are declared unfair workers must bereadmitted.Thepossibilityofpaymentforunfairdismissalsisnotenvisagedinthiscase,soinessencethislawprovidesitsuserswithgreaterprotectionagainstdismissals. Thereareseveralissuestobepointedout:First,thefactthatusersofthelawenjoydefacto higher protection against dismissals has raised some debate about the potential

5Moredetailsofthelawcanbefoundin:http://www.boe.es/boe/dias/1999/11/06/pdfs/A38934-38942.pdf.Thelimitonchidren’sageswasrelaxedin2006tocoverthoseagedbetween6and8andagainstin2013forthoseagedfrom8to12.Therightisalsograntedtoworkerswithfamilymemberswhoareclassedasdependentduetophysicalormentaldisability,butinthispaperwefocusonlyonworktimereductionforchildcareissues.6In2013therighttoworktimereductionforchildcarewasextendedtocoverareductionofbetween12%(1/8)and50%(1/2)ofan8-hourworkingday.7InSpainmorethan70%oflayoffstakentocourtaredeclaredillegalandinmostcasesemployersdonotreadmitworkersbutratherpaycompensationforunfairdismissal.

5

inflexibility regarding any future dismissals that firms facewhen hiring potential law users8.Second, thisprotectionagainstdismissal essentially appliesonly toworkersunder indefinitecontracts.Thelawsaysnothingwithrespecttoanyobligationtorenewfixed-termcontracts.In principle, target workers with fixed-term contracts are also granted the right to use thepolicyandhencereducetheirworkinghours.However,giventhatthesituationwithregardtojobprotectiondifferssomuchfromonetypeofcontracttoanother,welookatthepotentialimpactsofthelawforeachtypeofcontractseparatelyhere.Finally,asmentionedabove,thepolicywasaimedathelpingtoreconcileworkandfamilylifeforfamilieswithchildrenunder6.However,wefocusonlyonthepotentialimpactonmothers,giventhatpreliminaryevidenceindicatesthattheproportionoffatherswhoresortedtopart-timeworkbothbeforeandafterthepassingofthelawisconsistentlylowerthan1%.

3.THEDATA

As mentioned in the introduction, we use two main databases to (i) measure theimmediatedirectandindirecteffectsofthepolicy;and(ii)characterizelawusersafterthelawisimplementedandmeasuretheextenttowhichtheirnumbershaveincreasedordecreasedduringtherecessionyears.ForthefirstpurposeweusetheSpanishCurrentPopulationSurvey(SCPS)(EncuestadePoblaciónActiva,EPA),andforthesecondweusetheContinuousSampleofWork Histories (CSWH) dataset from Social Security records (Muestra Continua de VidasLaborales,MCVL),whichhasbeencollectedonanannualbasissince2005.

SPANISHCURRENTPOPULATIONSURVEY(SCPS): SCPS is a cross-sectional database which provides information on demographiccharacteristics(age,gender,yearsofeducation,maritalstatus,regionofworkandresidence,etc.),employmentcharacteristics(currentstatus,typeofcontract,lastwork,tenure,durationofcurrentcontractiffixed-term,numberofhoursworkedinthecurrentjob,currentPTstatus,weekly hours of work, labor status last year, etc.), fertility information (such as number ofchildren, demographic characteristics of children, etc.), household information (number ofadults and children in the household, information about grandparents, etc.). We useinformation for the second quarter of each year from two years before the law wasimplementedtotwoyearsafter.Specifically,wedenoteas“before”theyears1998and1999and as “after” the years 2001 and2002.Wedisregard the year 2000 aswe consider it as areferenceperiodtoguaranteeaclearcutbeforeandafter.Tocheckforrobustnesswerunthefollowingsensitivitychecks:(i)estimatetheimpactofthelawusingaplacebosamplewhichincludes1996and1997;(ii)useonlymotherswithchildrenbornbefore thepassingof the lawas the treatmentgroup in the“after”period, inorder to8This led the formerpresidentofa regionalemployers’association to state recently in themedia that firms faceshieldingwith respect towomenwith children under 12 years of age,which discourages employers fromhiringwomenoffertileageunderindefinitecontracts.

6

avoidanypotentialendogeneityoflawusers;(iii)usetheyears2003and2004insteadof2001and2002asthe“after”group.Thisallowsustocheck,first,whethertheresultsfoundbeforemightbeoverestimatedasaconsequenceofananticipatedeffectofwomenwhowaiteduntiljustafterthelawwaspassedtohaveachildsoastousethefamilyfriendlypolicy.Second,byusing 2003 and 2004 as the “after” period we check the extent to which the impact issustainedovertime.

CONTINUOUSSAMPLEOFWORKHISTORIES9(CSWH): This dataset is compiled annually from 2005 to 2013. It consists of 4% of thepopulation registered with the Social Security (SS) system either as workers, unemployedreceivingbenefitsorpensionersforatleastonedayofthecurrentyearofthesample(overamillionworkhistories).Thecomplete labormarkethistory isreportedforall individuals.Thisdatabase provides highly detailed information about their past and present labor activities,including monthly wages, type of contract, receipt of unemployment insurance benefits,reasons for job termination, and several characteristics of hiring firms such as size, age,ownership, location, and sector of activity. Individual characteristics such as age, gender,residence, nationality and household characteristics such as gender and date of birth ofhousehold members are also provided in the database – they are obtained from censusrecords.Everyindividualinthesampleisfollowediftheymaintainanyrelationship(working,beingunemployedreceivingbenefitsoraspensioners)withSocialSecurityrecords.ThereareseveralcharacteristicsthatmaketheCSWHanappropriatedatabaseforthisstudy.Firstly,itisan administrative dataset that provides highly accurate information on employment for arandom sample of 4%of all Social Security records. The data can be combinedwith censusinformationoneachyear so that it ispossible toobtain informationabout familymembers.Secondly, it is longitudinalsoit ispossibletoobtaininformationontheworker'sentirelabormarket history. Furthermore it assigns an employer identification code that enables firms,sectors,numbersofworkersandlocationstobeidentified.Typeofcontract,entryandleavingdate,andhoursworkedarealsoknown.WepoolalltheinformationregisteredinCSWHfrom2005to2013.Hence,anyindividualwhois includedinthisdatasetforat leastonedayfrom2005to2013appearsinoursample.Weusequarterlydata from2000to2011,keepingthe informationonallvariables fromtheCSWHonthelastdayofthechosenmonths(January,April,JulyandOctober),i.e.wecreateanewpanelofdatawith48observationsperindividual.Wefocusonlyonsalariedworkers(astheyaretheonlyonesaffectedbythe law).AsCSWHprovides informationonthecompleteworking life, if there isno informationonanydateweconsider the individual inquestionasunemployedwithoutreceivingbenefitsorinactive.Forourstudyweonlykeepworkepisodes.Furthermore,wecancapturetheexistence(andage)ofchildren,ifany,aswehavethebirth-rates of all household members. Given all the information available, this dataset is highly

9ForadetaileddescriptionofthissampleseeDuran(2007)andGarcía-Perez(2008).

7

suitableforrecognizingusersofthelawandanalyzingtheimpactofthecrisisontheuseofthepolicy.

4.IMPACTOFTHELAW–DIRECTANDINDIRECTEFFECTS

4.1.DIRECTEFFECT

Themain question to be answered here is whether the implementation of the lawresultedinanincreaseintheuseofreducedworkinghours(part-timework)byparentswithchildren under 6, as its spirit intended. As mentioned above, we investigate this using theSpanishCurrentPopulationSurvey for the twoyearsprior to thepassingof the lawand thetwoyearsfollowingit10.Wefirstdescribethespecificsampleusedtoaddressthisquestion:

First,wefocusonsalariedemployeeswithindefinitecontracts.Inprincipletherighttoask forwork time reduction also applies to fixed-term contractworkerswho are parents ofchildrenunder6yearsofage.However,evidencefromtheCSWHrevealsthattherearealmostno femaleworkersunder fixed-termcontractswhohavechildren fulfilling theconditions forthemtobepotentialusersofthelaw.Hencewerestrictourstudytoworkerswithindefinitecontracts.Wealsorestrictittoindividualsagedbetween25and45,i.e.offertileage.Over45swithverysmallchildrenorwithnochildrenmaybeoutliers intermsoftheirbehavior inthelabormarket(theyaccountfor11%ofthewholesample).Wedropworkersyoungerthan25becausesomeofthemmightbestudentsorlivewiththeirparents,whichwouldchangetheirprofiles:wewant to avoid parentswho are students in the analysis (2.5%of the sample ofparents). In addition we focus on married people, as single mothers11 might face differentpersonalcircumstancesfromtherestinresortingtothelaw12.

Preliminaryevidenceindicatesthattheproportionoffatherswhousepart-timeworkisconsistentlylowerthan1%,bothbeforeandaftertheintroductionofthelaw,wefocusontheimpactofthelawonmothers.Hence,ourreference(treatment)groupismothersworkingunderindefinitecontractswhohavechildrenunder6yearsold.

Asacomparison(control)groupweselectwomenwithoutchildren,i.e.peopleunderverysimilarconditions(withindefinitecontracts,marriedandoffertileage)butnotaffectedbythelaw.Thediscussionofwhatgroupisthemostappropriateforcomparisonpurposesisnot trivial. Inprinciple,anotherpotentialcontrolgroupmightbemotherswithchildrenwhoalsoneedchildcare(forexamplebetween6and8yearsold)butarenotaffectedbythelaw13.

10WedonotusetheCSWHtomeasuretheimmediateuseofthelawbecausethedatasetisfirstcompiledin2005,sothesampleofparentsthatcanbeobservedreducingworkinghoursduetochildcareissuesintheyearsprevioustothepassingofthelawisverysmallandbiased,asitisrestrictedtopersonsincontactwithSocialSecurityin2005.11Theyconstitute6%ofallmotherswithchildrenyoungerthan6years.12Wealsoexcludeindividualswholiveinthesamehouseholdasgrandparentsorhavedomesticservants.Thelawalsoallowsworkerstoreduceworkinghourswhentheyneedtotakecareofoldpeople.Giventhescopeofourpaper,wewanttoensurethatpotentialusersofthelawuseittotakecareofchildrenandnotothermembersofthefamily.13Previoussimilarpapers(Rodriguez-Planas,N.andFernández-Kranz,D.2011)developtheiranalysisusingmotherswithchildrenbetween7and12astheircontrolgroup,butasshownthisisnotapplicableinouranalysis.

8

To justify our choice of a comparison group,we compare the use of part-timework by thethreepotential groups – the target group and the twopotential control groups, before andafterthepassingofthelaw.Figure1depictsthattrend:

[Figure1]

Thered linedepicts theproportionofpart-time(PT)workers forour targetgroupofmotherswithchildrenbetween0and5yearsold. Asteady increase inPTcanbeobserved,whichbecomesclearlymorepronouncedfrom2000onwards.ThegreenlinerepresentsPTformotherswithchildrenbetween6and8yearsoldandthebluelineshowsPTforwomenwithnochildren14.ItisimmediatelyapparentthattheproportionofPTworkerswhoaremotherswithchildrenbetween6and8showsasignificantincreasearoundtheyearswhenthelawwaspassed which is not mirrored in the other two groups. This means that the parallel trendsassumption is not satisfied. This different behavior does not disappearwhenwe control forother observable covariates such as education, age, and type of job. Indeed the target ortreatment group seems to follow a similar trend to that of women without children: theproportionsofPTworkersrunparallelbeforetheimplementationofthelaw,withaconsistentgap of approximately 4 percentage points. This gap remains the same when observableindividualandjobcharacteristicsarecontrolledfor,butitisconsistentintheyearspriortothepassing of the law. This is themain reasonwhywe choosewomenwith no children as thecontrolgroup.

Ourfinalsamplecovers9520femaleworkersagedbetween25and45,allofwhomaremarried and hold indefinite contracts. 6329 of them have children under 6 years old(treatmentgroup)and3191havenochildren(controlgroup).

METHODOLOGY

To conduct this analysis we use the Spanish Current Population Survey. The maindisadvantageofcross-sectionaldata is the lackof longitudinal informationon individuals.Toaddress this drawback we use a difference-in-differences (DiD) method. The DiD design isusuallybasedoncomparingtwodefactodifferentgroupsbeforeandaftertheoccurrenceofthe treatment, i.e. a total of four groups. Three of these groups are not affected by thetreatment. Time is an important variable in distinguishing between the groups. Besides thegroupwhichhasalreadyreceivedthetreatment(mothersafterthepassingofthe lawinouranalysis) these groups are the following: (i) those treated prior to the current treatment(mothers before the implementation of the policy); (ii) those not treated, i.e. the controlgroup, justbeforethetreatment isappliedtothosetreated(non-mothersbefore2000);and(iii)thosenottreatedafterthetreatment(non-mothersafter2000).Theideaofthisempiricalstrategyisthatifthetwotreatedgroupsandthetwocontrolgroupsaresubjecttothesametime trends, and if the treatment has no effect before the passing of the policy, then anestimateofthe'effect'ofthetreatmentinaperiodinwhichitisknowntohavenonecanbe

14Needlesstosay,allthreegroupscontainmotherswhoaremarried,aged25-45andhaveindefinitecontracts.

9

used to remove the effect of interference factors towhich a comparison of post-treatmentoutcomesoftreatedandnon-treatedsubjectsmaybeprone.

Thisempiricalstrategyconsistsofestimatingthefollowingprobabilityequationforthelikelihoodofworkingpart-time.

wheretindexestheyearanditheindividual;andwhere =1iftheindividualworkspart-

timeandzeroifsheworksfull-time; =1ifindividualireceivesthetreatment

(motherswithchildrenunder6)andzerootherwise(womenwithoutchildren); =1if

observationisafterthetreatment(2001,2002)andzeroiftheobservationisbeforethetreatment(1998and1999)15. isavectorofcovariatesthatincludedemographic,

employmentandfamilyinformationsuchasage,levelofeducation,adummyindicatingwhethertheindividualisthefamily-head,birth,unemployment,proportionoffixed-term&part-timeworkersbyregionofworkandyearandsectorofwork.Finally, isazeromean

disturbance.

Thecoefficientβrepresentsthepre-treatmentdifferencesbetweenthetreatmentandcontrolgroups.Thisisthegapthatweobservebeforepolicyimplementation,i.e.theunobservablevariablesthataffectthetreatmentandcontrolgroupsdifferentlythecovariatesarecontrolledfor(asshowninFigure1).γcapturesthepost-treatmenteffectonthecontrolgroup.Thatis,howthelawaffectsnon-mothers.Finally,δisthetreatmenteffect.Thisisthediff-in-diffestimator.Itshowstheincreaseinthegapthatcomesspecificallyfromthepolicyimplementationforthetargetgroupandnotfromexternalfactors.

RESULTS

Before presenting the results of the direct impact of the law, we present somedescriptives which help characterize the main demographic and job characteristics of thetreatment and control groups before and after the lawwas passed. These are presented inTable1.

[Table1]

The first twocolumnspresentchanges for the treatmentgroupbeforeandafter thepassingof the law.Thedependentvariable is theproportionofpart-timeworkers,andtheirnumber increasesby 29.5%, from12.7% to 16.5%.With respect to the covariates, it canbeseenthattheproportionofheadsofhouseholdalsoincreases(rememberthatthislawaimstohelppeopletocombinechild-careandwork,somothersinparticularseemtoincreaseslightlytheir roles as heads of households). It is noteworthy that education levels increasesignificantly,withtheproportionofhighlyeducatedmothersupfrom41.86%to53.73%(arise

15Rodriguez-Planas,N.andFernández-Kranz,D.2011usedifferentperiods:astheirbeforegrouptheytakesixyearsfrom1994to1999andastheiraftergroupthreeyearsfrom2001to2003.

10

of28.35percentagepoints).Theconstructionandindustrysectorsalsobecomemorecommonformothers16.

Comparingmothers with non-mothers, note that before the passing of the law theproportionofmothersworkingPTwas3.82percentagepoints(p.p.)higherthanthatofnon-mothers(controlgroup),asshowninFigure1.However,after1999thisdifferencerisesto7.6p.p., i.e. 3.78 p.p. higher than before, which means a 100% increase. This increase can berelatedtotheeffectofthelaw.Withrespecttodifferencesindemographiccharacteristics,thetreatmentgroup issomewhatolder thanthecontrolgroup,which isexpectedaswe imposethe condition that the latter must be non-mothers. On the other hand, before 1999 thetreatmentgrouphadahigherlevelofeducationbutafteritthedifferencewasreduced.Non-mothers were employedmore in the construction and industry sectors before the law butafterthelawtreatedmothersincreasedtheirparticipationinconstruction.

Table 2 presents the main coefficients of interest from the estimation of equation[1]17. The first columndisplays the results from theunconditionalestimation -nocovariatesare included.Thesecondpresents theconditional resultsof theanalysis,wherecontrolsareincluded.Thesearetheresultsfromestimatingequation[1].

[Table2]

Thefirst issuetobepointedout isthattheunconditional (column1)andconditional(column2)impactsareverysimilar.Thissuggeststhatthecovariatesareuncorrelatedwiththetreatmentandjustgeneratemorepreciseestimatesofthecausaleffectofinterest18.

Morespecifically,andconsistentwithFigure1,conditionalonobservables,beforetheenactment of the law mothers of young children were more likely to work PT than non-mothers,as thevariabletreated (β) is statisticallysignificantat the1%significance levelandstands at 5.2percentagepoints (marginal effects shown in theTable),which represents thegap in the figure. The coefficient of interest reported in the third row (treated*after), δ,estimatestheeffectofthepolicyonPTworkforeligiblemotherscomparedtonon-mothers.Itcanbeclearlyobserved that the lawaffects the targetgroup.Allelsebeingequal,after theimplementationofthelawawomanwithsmallchildrenandanindefinitecontractisaround3percentagepointsmorelikelytoworkPT19).Thisisequivalenttoanincreaseof18%inthelikelihoodofworkingpart-timeforthetargetgroup,consideringthatthepredictedprobability(controlling for covariates) of working PT before the implementation of the law for thetreatment group is 15.74% (10.58+5.16). The table also shows that the lawdidnot affectnon-targetwomenunderindefinitecontractsbecausethevariableafter(γ)isnotstatistically

16Theservicesectorisbrokendowninto5subgroups:wholesale,education,health,housingandotherservices.17Marginaleffectsarereportedinthetables.18Noticethatthestandarderroroftheestimatedtreatmenteffectincolumn2issmallerthanthecorrespondingstandarderrorincolumn1.Furthermore,theR-squarevalueincreaseswhenthesecontrolvariables ,areincluded19Ifthe“after”periodisextendedtothreeyearstheimpactis28.57%stronger(increaseof3.6percentagepointsversus2.8).Tochecktherobustnessofthisanalysiswealsoruntheestimationconsidering1997-1999asthe“before”periodand2001-2004and2001-2003asthe“after”period.Theresultsarebroadlythesame.

11

significant. Thismeans that is can be strongly argued that the unobservables in the controlgroupremainconstant,whichmakesourcomparisonmorerobust.

To obtain the probability ofworking part-time for treatedmothers after the law alleffects(β+γ+δ)mustbeaddedtogether.Afterthe lawtreatedwomenwere8percentagepoints(5.16+0+2.86)morelikelytoworkPTthansimilarwomenwithoutchildren,i.e.theyshowapredictedprobabilityof18.6%.

ROBUSTNESSCHECKS

Asarobustnesscheckweimplementseveraltests,allofwhichareshowninTable3.

[Table3]

Column1present theresultsofaplacebotest,conductedtoensurethat the impactestimatediscausedbythefamilyfriendlypolicyandnotbyother,externalfactorscorrelatedwithit.Thisplaceboanalysisconsistsof“pretending”thatthetreatmenthappenedearlierandthenmeasuring the outcome after the pretend treatment but before the actual treatmenttakesplace. If thisartificial treatment is found tohaveaneffect then thateffectbecomesaspecification test for the common trend assumption, because any estimated nonzero effectwould have to be interpreted as selection bias and thus would cast serious doubts on thevalidityoftheidentifyingassumptions.

Hence,weestimatethesamedifference-in-differencesmodelsforaperiodinwhichnochange in family-friendly laws tookplace.Weuse thesamepre-reformperiodof1998-1999(excludingpost-1999data)fortheseestimatesbutwesimulateafalsepost-reformperiod.Fortheplacebotestweconsidertheperiods“after”1996and1997.Thisperiodischoseninorderto find themost similar group to 1998-1999 (nearest in time) for comparison, and to avoidchanges in the likelihood of part-time work for external reasons. For the placebo test thesampleincludes8,784females:6,078mothersand2,706non-mothers.

Theresultsoftheplaceboregressionareshownincolumn2ofTable2.Itcanbeseenthat the coefficientof interest - treated*after, is not significantlydifferent from zero,whichmeansthatifthe1996-1998periodhadbeenthe“after-the-law”years,nodifferentialuseofpart-timeworkwouldhavebeenfoundacrossmotherswithchildrenyoungerthan6yearsofage and non-mothers. This implies that if the law had not actually been implemented nochangeswouldhavebeenobserved in the incidenceofPTversus full-timework formotherswithchildrenyoungerthan6yearsoldcomparedtonon-mothers.Thisconfirmsthattheresultpresented above is caused by the implementation of the law and not by any spurious orunobservablefactor.

Thesecondrobustnesschecklooksforanybiasasaresultofpotentialendogeneityofthetreatmentgroup:Thismightariseifsomenon-mothers,knowingthattheimplementationof the family friendly policy is imminent and attracted by its advantages, decide to becomemothers and hence switch from the control to the treatment group. To checkwhether this

12

potential effect plays any role,we eliminate from our treatment group thosemotherswithchildrenbornjustafterthepassingofthelaw,i.e.thosewithchildrenunder1yearoldin2001andunder2 in2002. The result is shown in the second column. There is hardly any changewith respect to the original one. Therefore, it seems thatwomendo not decide to becomemothersattractedbytheadvantagesofferedbythepolicy.

The third and last robustness check is intended to evaluate whether the impactreported above is just a short-run effect or is sustained over time. To that endwe use the2003-2004asthe“after”group.ResultsareshowninColumn3.Thetreatmenteffect(δ)inthemedium-term is statistically significant at the 1% significance level and amounts to 5.92percentagepoints.Thisisequivalenttoanincreaseof36.4%inthelikelihoodofworkingpart-time for our target group in the medium-term, considering that the predicted likelihood(controlling for the covariates) of working part-time for the treatment group before theimplementationofthe law is16.28%(10.98+5.3). Itcanbeconcludedthattheeffect isnottransient. Indeed, the likelihood of resorting to the law seems to increase over time. Onepossibleexplanation for that increase is that it takes time forknowledgeof theexistenceofthepolicytospread,anditsusehasincreasedasithasbecomeprogressivelybetterknown.

Therefore,themainconclusionregardingthisfirst,directimpactofthelawisthatthepassingofthepolicy increasedtheuseofpart-timeworkbytargetmothersbyabout18%.Tosomeextent,thus,thelawsucceededinitsmainaim.

4.2.IMPACTOFTHELAW:INDIRECTEFFECT Nextweanalyzetheextenttowhichthepassingofthelawhashadpervasiveeffectson itspotential futureusers (whatwedenoteby its“indirecteffect”).Thepotentialusersofthe law are non-mothers of fertile age, given that it is mothers with small children whoaccountfortheincreaseintheuseofpart-timeworkforchildcarereasonsafterthepassingofthelaw,asshownintheprevioussection. Thepervasiveeffectanalyzedinthissectioniswhetherthislawhasledemployerstobehave strategically in the senseof anticipating the lawwhenhiringworkers, dependingonwhethertheyarepotentialusersornot.Thismightbethecaseifemployersfeelthattheuseof part-time reduced and the increased protection from dismissal that these workers mayenjoyimposeadditionalrestrictionsonthem.Ifso,andifthefamilyfriendlypolicyiscostlyforfirms, employers might be more likely to hire potential users of the law under fixed-termrather than indefinite contracts. Under fixed-term contracts employers are not forced torenewthecontractswhenworkersbecomeeligibletousethelaw,sothoseworkerswouldnotenjoythegreaterprotectionagainstdismissal.Thereforethetreatmentgroupinthisanalysiscomprisesnon-mothers(whoarepotentialfutureusersofthelaw)andtheoutcomevariableis the probability of having a fixed-term contract, measured before and after the law. Asbefore,wefocusonsalariedworkersoffertileage(between25and45yearsold).

13

Asthecontrolgroup,weusenon-fathers inthesameagebracket.Anotherpotentialcontrolgroupmightbeallmalesalariedworkers,notonlynon-fathers20.Todecidewhichofthetwocontrolgroupsseemmoreappropriate,wecomparethetrendsintheproportionsoffixed-termcontractsforthesetwogroups(non-fathersandallmalesalariedworkers)andthetreatmentgroup.Figure2depicts thoseproportions. Itcanbeseenthat the trends fornon-fathersandnon-mothers lookverysimilarbefore thepassingof the law (30.5%)butdivergeafter it: They increase for non-mothers but not for non-fathers. This is not the case for allsalariedworkers,whoexhibitremarkabledifferencesintheuseoffixed-termcontractsintheyearsprevioustothepassingofthelaw.Hence,wefinditmoreappropriatetousethegroupofnon-fathersagedbetween25and45asthecontrolgroup.Ourfinalsamplecovers20,118individuals:11,332non-fathersand8,786non-mothers.

[Figure2]

METHODOLOGY

To conduct this second analysis we again use SCPS and a difference-in-differencesmethod.Theregressionisthesameasinthepreviousanalysisbutnowthedependentvariableistheprobabilityofbeinghiredunderafixed-termcontract.Andasmentioned,wenowfocusonasampleofnon-mothers(treatmentgroup)andnon-fathers(controlgroup)agedbetween25and45.

wheret indexestheyearand itheindividual;andwhere =1ifindividual ireceivesthe treatment (potential mothers) and zero otherwise (potential fathers), = 1 ifobservationisafterthetreatment(2001or2002)andzeroifitisbeforethetreatment(1998and1999). isavectorofcovariates(asinthepreviousanalysis).Finally, isazeromeandisturbance.

βisthenon-treatmenteffect.Anon-significantlydifferentfromzerocoefficientwould

reveal that, all else being equal, the treatment and control groups exhibited similarproportions of fixed-term contracts before the passing of the law. γ captures a potentialindirectimpactofthelawonnon-fathers–i.e.achangeintheprobabilityoffixed-termhiring.δ is the treatment effect, i.e. the change in the likelihood of being hired under fixed-termcontracts for non-mothers after the law versus before the law compared to non-fathers. Asignificant positive coefficientwould indicate that the law increased the probability of non-mothers being hired under fixed-term contracts as compared to the corresponding non-fathers.

RESULTS

20Rodriguez-Planas,N.andFernández-Kranz,D.2011computeitinthatway.Theanalysiswasalsoconductedinthiswayandthetreated*aftervariablewasfoundtobehigherandstatisticallysignificant.

14

Beforepresentingtheresultsoftheestimation,wepresentsomedescriptivestatistics.

Table 4 shows the proportions of fixed-term contracts and other average values of theindependent variables for the treatment and control groupsbefore andafter thepassingofthelaw.

[Table4]

A comparison of non-mothers (treatment group) before and after the law shows aclear increase in the proportion of fixed-term contacts - from 30.48 to 34.91% (15.5%).Wediscussbelowwhetherthisincreaseisduetothe“indirecteffect”ofthelawthatweseektotest.Inregardtothecovariates,itmustbepointedoutthateducationlevelishigherafterthepassingof the lawthanbefore fornon-mothersagedbetween25and45.Theproportionofnon-motherswithuniversitydegreeincreasesfrom37.6%in1998-1999to52.4%threeyearslater.

A comparison between non-mothers and non-fathers (control group) reveals thatbefore the passing of the law the proportion of fixed-term contracts is the same for both(30.5%),butfornon-fathers itremainsunchangedafterwards.Thisreinforcesourconfidenceinthesuitabilityofthisgroupasacontrolgroup.Sectoralseparationisnoticeablewhennon-mothersarecomparedwithnon-fathers,bothbeforeandafterthepassingofthelaw.Thisisalso the case with the whole sample of female and male workers. Women are highlyconcentratedinservices,andtheirincidenceinindustryandconstructionisreallylow.

Table5presentsthemaincoefficientsofinterestfromtheestimationofequation[2].Column1displaystheunconditionalimpact,i.e.withnoadditionalcovariates.Column2showstheconditionalimpactontheobservablecovariates.

[Table5]

Thecoefficientsofthetwoestimationsareveryclose,whichindicatesthatthesetofcovariates are uncorrelated with the treatment. On the other hand, the estimation fitincreasesnotably,asindicatedbytheincreaseintheR-square.

FromColumn2 it can be concluded that firms seem to behave strategically: all elsebeing equal, after the passing of the law non-mothers between 25 and 45 years are 5.33percentage points more likely to be hired under a fixed-term contract than thecorresponding non-fathers,whichmeans that an unintended and unexpected effect of thislawhasbeentoincreasethehiringofpotentialmothersunderfixed-termcontractstopreventthemfromhavingtherighttoreducetheirworkinghoursandenjoygreaterprotectionagainstdismissal.Thatincreasemakesthelikelihoodofwomenbeinghiredunderfixed-termcontractsaround18%higher,giventhatbeforethecrisisthatlikelihoodwas30%.Inaddition,giventhatthe impact of the variable treated is not significant, it can be concluded that similar non-mothers and non-fathers were equally likely to be hired under indefinite contracts before1999.Finally,astheimpactoftheindicator“after”isnotsignificanteither,itcanbeconcluded

15

that fornon-fathers the lawhadno impacton the likelihoodofbeinghiredunder indefinitecontracts.

ROBUSTNESSCHECKSThe tests shown in Table 6 were conducted as robustness checks on the estimated

indirecteffect.

[Table6]

Asbefore,wefirstrunaplacebotest,presentedinthefirstcolumn.Inparticular,weuse1996-1998asafictitious“after”interval,aswedidwiththedirecteffect.Fortheplacebotestthesampleincludes16,858individuals:9,565menand7,293women.Column1indicatesthat the treatment effect – the δ coefficient - is not statistically significant. So if a fictitious“after” period is used no significant increase is found in the use of fixed-term contracts fornon-mothers with respect to non-fathers. This supports the assumption that our previousresultsontheeffectsofthefamily-friendlylawwerenotspurious,andaddsrobustnesstothepreviousresultthatthepassingofthelawmadeitmorelikelyfornon-mothersthanfornon-fatherstobehiredunderfixed-termcontracts.

Thesecondrobustnesscheckevaluateswhethertheimpactisfeltonlyintheshortrunorissustainedovertime.Again,weusetheyears2003and2004asthe“after”group.Resultsare shown in Column 2. The treatment effect (δ) is statistically significant at 10%. Thismayariseiftheanticipatorybehaviorbyemployersvanishesovertime.

Therefore,themainconclusionreachedconcerningtheindirectimpactofthepolicyisthat its passing made it more likely for non-mothers than non-fathers in similarcircumstancestobehiredunderfixed-termcontracts,butitseemsthattheeffectdisappearsovertime.

5.FAMILYFRIENDLYPOLICY:PROFILESOFUSERSOFTHELAWANDTHEIMPACTOFTHE

GREATRECESSION

ThesecondaimofthepaperistocharacterizethepersonalandjobprofilesofusersofFamily Friendly Law39/99 (referred tohereafter as “users of the Law”) and toquantify theextenttowhichtheGreatRecessionhaschangedthenumberandcharacteristicsoftheusersof the Law. Asmentioned in the introduction, on one side itmight be expected that in aneconomicdownturnincreasedfearsofbeingdismissedwouldleadtoanincreaseintheuseofwork time reductionas awayofobtaininggreaterprotectionagainstdismissal.By contrast,income effects might have a negative impact, as work time reduction entails proportionalwagereduction,andthisnegativeincomeeffectislikelytobemoreimportantinarecession.

16

DATAANDDESCRIPTIVES

TheContinuousSampleofWorkHistoriesenablesindividualswhomakeuseofthelawtobeidentified.AsmentionedintheDataSection,wemergealltheyearsintheCSWHfrom2007to2013andkeepallindividualswhoregisteranyworkcontractfrom2000onwards.Werestrictoursampletoworkerswhohavehadchildrenatsometimebetween2000and2011.We compute a quarterly panel and divide the whole sample into two periods: contractsbetween2004and2007 (denotedas theexpansionaryperiod),andcontractsbetween2008and2011(denotedastherecessionperiod).Accordingly,weonlykeepepisodesfrom2004inorder to avoid possible bias caused when we require individuals to remain in the SocialSecurityrecordsin200521.

Under the legal context in our reference period, employers cannot change theirworkers’workinghourswithouttheworkers’consent22.The2012LaborReform changesthelegalcontextandallowsemployers toreducetheworkinghoursof theiremployees insomespecificsituations,soweonlymeasuretheimpactofthecrisisuptotheendof2011.Hence,fortheperiodunderconsiderationweassumethatallchangesfromfull-timetopart-timeinthe same firmarevoluntary.Therefore, inprinciplewe identifyanyemployeewhouses therighttochangeher/hisfull-timecontracttoapart-timecontractwithinthesamefirm23whenhavingachildofthepermittedageasauseroftheLaw.

However,thenumberofmenwhodeclarethattheyworkPTduetochild-careissuesisless than1%24,and thispattern seems to remainconstantover time.Therefore,weexcludemenfromthisanalysis,astheproportionofusersofthelawamongthemisnegligible.

Second,onlyworkersunderindefinitecontractsarereallyprotectedagainstdismissalsince under fixed-term contracts employers are not forced to renewworkers´ contracts. Infact, we find an insignificant number of fixed-term contract25 workers who reduced theirworkinghoursonhavingachild.ForthatreasonwefocusonworkerswithindefinitecontractsinanalyzingusersoftheLawandtheimpactoftherecession.

Toidentifynon-usersoftheLawwefocusonallthosepotentialuserswhodecidednottochangetheirworkinghourseventhoughtheywerelegalentitledto.Wedefinea“non-user”as anymotherwith children of the permitted agewith a full-time (indefinite) contractwhomaintainsthesametypeofcontractinthenextperiodifsheremainsatthesamefirm.

The sample contains 2,578 different users26. It covers 835,713 observations (womanperquarter),20,259ofwhicharefromusersoftheLaw27.

21Werunsomecheckstoensurethatwedonotendupwithabiasedsamplefrom2004.22Article12.4.eEstatutodelosTrabajadores[“Workers’Statute”]Exactdetailscanbefoundhere:http://www.estatutodelostrabajadores.com/a12-contrato-a-tiempo-parcial-y-contrato-de-relevo/23Changestodifferentfirmsarenewcontractsandnotreductions,sotheydonotprovideprotectionagainstdismissal.24FigureobtainedfromtheSpanishCurrentPopulationSurvey.25380observationsin8years.26IftheCSWHrepresents4%oftheSpanishpopulation,thenitshows64,450realusersofthelaw.Onaveragealmost300,000womendeclaredthattheyworkedpart-timeduetochild-careissuesinSpainbetween2004and

17

Figure3depictstheproportionofmothersinthepoolofeligiblepotentialuserswhomakeuseoftheLawquarterbyquarter.Onaverage,intheperiod2004-2007theproportionof users is 2.46%, while in 2008-2011 it decreases to 2.40%. As Figure 3 shows, there is asudden jump of 17% preciselywhen the law extends the permitted age of children by twoyears (23rd March 2007). This jump is due to the fact that the increase in the number ofmotherseligibleisgreaterthanthatinthenumberofactualusers.Gradually,theproportionofusersdropsbackto former levelsbecauseusersof theLawextendtheir reducedworkinghoursuntiltheirchildturns8yearsold.

[Figure3]

Table5characterizesusersoftheLawfordifferentperiods. Intheupturnperiodthetypical profile of a user of the Law is a women in her thirties, of Spanish nationality, withbetween2and7yearsoftenure,workingasclericalofficerorassistantinasmallfirmintheservicesector.However,withtheonsetoftherecessiontheprofilechangestosomeextent:First,theproportionofover40sincreasesrelativelyintherecession.Second,theproportionofforeign workers among users increases by 60% in the recession period. Third, the averagetenureofusersoftheLawalsoincreases28.Intermsofoccupationalclassificationgroups,userscanbedividedintowhite-collar(thefirstsevengroups)andblue-collar(thelastfourgroups).Exceptfortechnicalengineersandexpertsandqualifiedassistants,agreaterdecrease inthenumberofusersof theLaw isobservedamongwhite-collarworkers thanamongblue-collarones in the recession, comparedwith thepreviousupturn. In addition, usersof the lawareover-representedinsmallfirms(withfewerthan10employees)beforeandafter2008.

Ascanbeseen,thecharacteristicsofusersoftheLawseemtochangeconsiderablyinthe recessionperiod. This change is perfectlyunderstandable if it is taken into account thatalthoughthe large-scaledismissalsbroughtonbytheGreatrecessionhitworkerswithfixed-termcontractsharder,theyalsoaffectedthosewithindefiniteones.

[Table7]

METHODOLOGYANDRESULTS

Our first aim is to estimate the determinants of “being a user of theLaw” and howtheychangeoverthebusinesscycle.Wecomparemotherswhoreducetheirworkinghoursintheir firms (users)with thosewhodecide to stay full-time (non-users)even though theyareentitled to take reductions. Table 5 shows a compositional change in users, whichmust becontrolled forwhen estimating the determinants of being a user of the Law.Wedo this byestimating two non-linear probit models separately. Among the covariates used in theestimationswe includeage(in intervals),adummyindicatingwhetherthemother isSpanishor a foreign national, tenure (in intervals), occupational classification group (white and blue

2011,i.e.accordingtoourresultsalmostoneoutoffivepart-timersmaybeusersoftheLawandhenceprotectedagainstdismissals.27Noticethatthesamewomencanbeusersandnon-usersindifferentperiods.Wefind63,938differentwomenwhoareeligiblebutneverusethelawduringtheperiodobserved.28Wecomputethisclassificationbasedonitsdistribution.

18

collar),sizeofthefirm(inintervals),9indicatorsofsectorofactivity,andregionalfixedeffects.Ourreferenceprofileisawomanbetween30and34,ofSpanishnationality,with2-7yearsoftenure,inawhite-collarjob,andworkingatasmallfirm.Table6presentstheresultsfortheexpansionaryperiod(2004-2007)initsfirstcolumn,andfortherecessionperiod(2008-2011)initssecond.

[Table8]29

Table6presents thedifferent impactsof the covariateson the likelihoodofbeingauser of the policy throughout the business cycle. After 2007women under 30 use the Lawmorethanourreferencegroup,contrarytothesituationbeforethecrisis.Mothersover40are56%more likely tousetheLaw intherecessionperiodthanthereferencegroupofwomen.Havingmore than7yearsof tenure,beingablue-collarworker,andworkingata small firmseem to be stronger determinants for using the Law in the recession than in the precedingperiod.Summarizing,Tables5and6revealthatthereisnotonlyacompositionalchangeinthesampleofnon-users,butalsoachangeintheimpactofthedeterminantsofbeingauseroftheLaw.Thismustbetakenintoaccountwhenestimatingtheimpactoftherecessiononusers.

ThesecondaimofthissectionistoquantifytheextenttowhichtheGreatRecessionled toachange in thenumberofusersof theLaw. Inotherwords,weseek toestimate theeffect of the recession on the use of the Law. As mentioned above, on the one hand theRecessionmightbeexpectedtoleadmoreworkerstousetheLawtoprotectthemselvesfromdismissal, but on the other hand income effects and probably also a fear of reprisalsmighthavetheoppositeeffect.

[Table9]

Toanswer thisquestionweestimate theprobabilityofbeingauserof the Law,butincludeindicatorsfortherecessionperiod.Table7displaystheresults.Eachcolumnshowstheeffectofthefamilyfriendlypolicyallowingfordifferentialimpactsbeforeandafter2008.Thevariabledenotedbycrisistakesavalueof0iftheobservationbelongstotheexpansionperiod(2004-2007) and 1 if it belongs to any quarter in the recession (2008 onwards). Column (1)representstheeffectofthecrisisvariableitselfwithnocontrolsforobservables(rawimpact).Inthatcontext,thevariablecrisisdoesnotcaptureanychangesinuse-noticethatR-squareis0,whichmakesthemodelunreliable.Inthesecondcolumnwecontrolforthesamevariablesas in previous estimations (in Table 6) and maintain the profile of the reference group ofwoman. The impact of the variable crisis is statistically significant at 1% and the impactsamountto-0.0027;i.e.intherecessionperiodmothersare0.27percentagepointslesslikelytousetheLawthanintheexpansionperiod.Tomeasurethescaleoftheimpactweneedtocompare it with the likelihood of using the Law in the pre-crisis period. For example, theaverage predicted likelihood of a woman in the reference group using the Law is 1.98%(predictedlikelihood).Hence,theimpactoftherecessiontakestheformofadecreaseinuseof13.6%.

29Marginaleffectsarereportedintheestimationtables.

19

Thisresultassumesthattheimpactofeachcontrolvariableisinvarianttotheperiodunder consideration, i.e. pre-recession or recession period. However, thismight not be thecase.Asshownpreviously,thereisnotonlyasignificantcompositionalchangeinthegroupofusersoftheLawbetweenthepre-crisisandrecessionyearsbutalsoachangeintheimpactsofthedifferentdeterminants.Hence,theimpactoftherecessionfoundheremightbebiasedasitmightcapturenotonlytheimpactoftherecessionitselfbutalsocompositionalchangesthathavenotbeentakenintoaccount.

To control for these changes in composition,we conduct awithin-cell estimation asfollows:

We identify thecells forwhich relevant changes in composition (and in impacts)areobserved in users of the Law before and during the recession, and then we estimate thelikelihoodofbeingauserwithincellssoastocomparewomenwithverysimilarcharacteristicswhen measuring the impact of the recession on the use of the law. This means that thevariable “crisis” captures the difference in the likelihood of using the Law among motherswithincells,hencepreventingthecoefficientestimatedfromcapturingtheeffectscausedbythechangeincompositionorthenon-normaldistributionoftheunobservables.

Given that compositional changes mainly affect the proportion of foreign/Spanishnationals,tenureandjobqualifications,wecreate12cellswithallpossiblecombinationsof(i)foreign/Spanish nationality; (ii) tenure (three groups); and (iii) job qualifications (whitecollar/blue collar)30. In addition to controlling by cells31, we also include indicators of agegroup,firmsize,sectorofactivity,andregionalfixedeffects.Thereferenceprofileisthesameasbefore.Theresultsofthewithin-cellestimationandtherestofthecovariatesareshowninthethirdcolumn.Thereisbarelyanychangewithrespecttocolumn(2):thereisareductionof0.262percentagepointsintheuseoftheLawduringtherecession.Inotherwords,awomeninthereferencegroupbecomes13.2%lesslikelytobeauseroftheLaw.

Finally, the last column presents a similar analysis using a more restricted cellgrouping.Inparticular,weaddagegroups(4)andfirmsize(4).Giventhatthegroupofforeignworkersisnotbig,wedonotincludeforeign/Spanishnationalityasanadditionalcharacteristicforthecellsoastopreventemptycells.Hence,weendupwith72differentcells32.Resultsofthewithin-cellestimationwithmorerestrictivecellcharacterizationarepresented incolumn(4).Theresultsdonotchange:0.265percentagepointsintheuseoftheLawafter2007.Thismeansthatduringtherecessionlikelihooddropsby13.7%forthereferencegroupofwomen.

In summary,with regard to the impactof the recessiononusersof theLawwe findthattherecessionhasledtoadecreaseintheirnumberofaround13%.Thisresultsuggests

30Noneofthese12cellsisemptyandtheyallcontainusersandnon-usersbeforeandduringtherecession.31Inspiteofthisclassification,thecellmadeupofforeignnationals,bluecollarandmorethan7yearstenurehasnousersinthepre-crisisperiod.32Underthisclassificationallcellsarefilledwithusersandnon-usersbeforeandafter2008.

20

thatnegativeincomeeffectsandperhapsfearsofreprisalhaveoutweighedpotentialgreaterprotectionagainstdismissal.

6.SUMMARYANDCONCLUSIONS

Family issuesplayacrucial role inunderstanding thegenderdifferencesobserved in thelabor market. Women combine employment with home responsibilities to a much largerextentthantheirmalepartners.Governmentsandinstitutionsmayplayanimportantrole increating the legal framework for improving women’s choices and their participation in theeconomy, and in helping societies to break away from the more traditional gender roleattitudes that affectwomen’s behavior inmany countries. Indeed, in the past few decadespoliciesaimedatpromotinggenderequalityandequityintheworkplacehavebeenadopted.TheevaluationofonesuchpolicyimplementedinSpainin1999istheaimofthispaper.

The policy under analysis, called Law 39/99, was particularly aimed at granting parentswith children younger than 6 years of age the right to reduce theirworking hours,with anequivalentwagereduction.Thespiritofthislawistoenableparentsmoreeasilytoaffordtostay in the labor market and take care of their children by reducing their working hours.Moreover, users of the Law enjoy greater protection against dismissal than other workers,which may encourage workers to use the law as a job protection particularly in recessionperiods.

Inthispaperweevaluatetheimpactofthelaw,inparticularitsdirectandindirecteffects.Ourresultsindicatefirstthatthelawincreasedthelikelihoodofworkingpart-timeforeligiblemothers– i.e.motherswith childrenunder6 -byaround18% (almost3percentagepoints)comparedtosimilarnon-targetgroups.Second,wetestwhetherthepassingofthelawledtostrategic behavior from employers in the sense of offering fewer indefinite contracts topotentialusersof the lawsoas to limit theuseof reducedworkinghours.A comparisonofhiringpractices involvingpotentialusersofthe law(targetgroup), i.e.non-mothersoffertileage,withasimilarnon-targetgroup(non-fathersoffertileage)revealsthatthelawincreasedthe likelihoodofthetargetgroupbeinghiredunderfixed-termcontractsby5.33percentagepoints(18%).

ThesecondaimofthestudyistocharacterizetheworkerswhohavemadeuseoftheLawsince its approval (1999)andmeasure theextent towhich theGreatRecessionhas led toachangeinthenumberofusersandintheirpersonalandjobprofiles.WefindthattheprofilesofusersoftheLawinthedownturnhavechangedfromthoseinthepreviousupturn.Beforethecrisistheyaremainlywomenintheirthirtiesinwhite-collarjobs,butduringthedownturnthey are low-qualifiedworkers older than 40whowork in small firms. Themost importantfindingisthattheGreatRecessionhasreducedthelikelihoodofresortingtotheLawbymorethan 13%. This is not consistentwith the view that eligibleworkers use the Lawduring therecentrecessionmainlytoprotectthemselvesagainstdismissal.

21

References

Adserà, Alicia (2004). “Changing Fertility Rates inDevelopedCountries. The Impact of LaborMarketInstitutions.”JournalofPopulationEconomics17(1):17–43.

Adserà, Alicia (2005). “Where are the Babies? Labor Market Conditions and Fertility inEurope.”EuropeanJournalofPopulation27(1):1-32

Ahn,Namkee,andPedroMira(2001).“JobBust,BabyBust?:EvidencefromSpain.”JournalofPopulationEconomics14(3):505–21.

Angrist, Joshua D. and Pischke Jörn-Steffen (2009). “Mostly Harmless Econometrics. AnEmpiricist’sCompanion.”PrincetonUniversityPress.

Ariza, A., de la Rica, S. andUgidos, A. (2003). “The effect of flexibility inWorking hours onFertility: A Comparative analysis of selected European Countries.”DFAE-IIWP Series, ISSN-e1988-088X,Nº.8

Barry T.Hirsch (2005). “WhyDo Part-TimeWorkers Earn Less? TheRole ofWorker and JobSkills”Industrial&LaborRelationsReview,58(4):525-551

Bustillo,R.M.,Macías,E.F.,andPérez,J.I.A.(2008).“EltrabajoatiempoparcialenEspañaenelcontextodelaUniónEuropea”.MinisteriodeTrabajoyAsuntosSociales.

Bertola,Giuseppe,FrancineD.Blau,andLawrenceM.Kahn.(2007).“LaborMarketInstitutionsandDemographicEmploymentPatterns.”JournalofPopulationEconomics20(4):833–67.

Bettio,Francesca,andPaolaVilla (1998)“AMediterraneanPerspectiveontheBreakdownofthe Relationship Between Participation and Fertility.” Cambridge Journal of Economics. 22(2):137–71.

DelaRica,S.,andFerrero,M.D.(2003).“Theeffectoffertilityon labourforceparticipation:theSpanishevidence”.SpanishEconomicReview,5(2),153-172.

De laRica,S.andAmaia I. (2005). “CareerPlanning inSpain:DoFixed-TermContractsDelayMarriageandParenthood?”ReviewofEconomicsoftheHousehold3(1):49–73.

De Laat, Joost, and Almudena Sevilla-Sanz (2011). “The Fertility and Women’s Labor ForceParticipation Puzzle in OECD Countries: The Role of Men’s Home Production.” FeministEconomics17(2):87–119.

DelBoca,Daniela(2002).“TheEffectofChildcareandPart-TimeonParticipationandFertilityofItalianWomen.”JournalofPopulationEconomics15(3):549–73.

Escobedo,A.,Navarro,L.andFlaqueriVilardebò,L.(2008).“Elimpactodelamaternidadylapaternidadenelempleo:Itinerarioslaboralesdemadresypadresenhogaresconmenoresdetresaños.”

22

Fernández-Kranz, D., and Rodríguez-Planas, N. (2011). “The part-time pay penalty in asegmentedlabormarket.”LabourEconomics,18(5),591-606.

GarcíaPérez,J.I.,(2008).LaMuestraContinuadeVidaslaborales(MCVL):unaguíadeusoparaelanálisisdetransicioneslaborales.RevistadeEconomíaAplicadaXVI(E-I),5–28.

Goldin, Claudia. (2005) “From the Valley to the Summit: The Quiet Revolution thatTransformedWomen’sWork”RegionalReview.14:3,pp.5-12.

Jaumotte,F.(2003).“Femalelabourforceparticipation:PasttrendsandmaindeterminantsinOECDcountries.”

Lapuerta,I.,Baizán,P.AndGonzález,J.(2008).“WhobenefitsfromparentalleaveinSpain?ALifeCourseAnalysis.”

Michael L. (2011). “The Estimation of Causal Effects by Differences-in-DifferencesMethods”DiscussionPaperUniversityofSt.Gallen.

Nollenberger, N. and Rodríguez-Planas, N. (2011) “Child Care, Maternal Employment andPersistence:ANaturalExperimentfromSpain.”IZAdiscussionpaper5888

Rodriguez-Planas,N.andFernández-Kranz,D.(2011).“UnintendedEffectsofaFamily-FriendlyLawinaSegmentedLaborMarket.”IZAdiscussionpaper5709

Rodriguez-Planas, N., D., Lacuesta, A. and Fernández-Kranz, D. (2013). “The motherhoodearningsdip:Evidencefromadministrativerecord.”JournalofHumanResources48(1),169-197

Treviño,R.,Vidal,E., andDevolver,D. (2009). Factorese indicadoresdevulnerabilidaden laconciliacióndeempleoyfamilia.EstudiosFIRPOS.

Visser,J.2002.“Thefirstpart-timeeconomyintheworld:amodeltobefollowed?”JournalofEuropeanSocialPolicyFebruary2002vol.12no.123-42

23

FIGURESANDTABLES

Figure1.Part-timerate.SCPS(1992-2004)Marriedwomenbetween25and45yearsoldunderindefinitecontract

0.00%$

5.00%$

10.00%$

15.00%$

20.00%$

25.00%$

1992$ 1993$ 1994$ 1995$ 1996$ 1997$ 1998$ 1999$ 2000$ 2001$ 2002$ 2003$ 2004$

Without$children$ With$children$(between$0$and$5)$ With$children$(between$6$and$8)$

Figure2.Fix-termrate.SCPS(1992-2004)Individuals between 25 and 45 years: non-fathers and non-mothers without children and all men

Law

Law

24

Figure3.ProportionofLawUsers-CSWH(2007-2013)

Mothersunderindefinitecontractthathavechangedtopart-timeinthesamefirmvsthosethatstayedworkingfull-time.

2.46%&2.48%&2.55%&

2.65%&2.77%&2.75%&2.71%&2.65%&2.70%&2.60%&2.58%&

2.66%&2.61%&

2.15%&2.20%&2.27%&

2.39%&2.33%&2.27%&2.27%&

2.45%&2.41%&2.45%&2.50%&2.55%&2.51%&2.52%&2.54%&

2.61%&2.54%&2.53%&2.54%&

0.00%&

0.50%&

1.00%&

1.50%&

2.00%&

2.50%&

3.00%&

31jan2004&

30apr2004&

313jul304&

313oct304&

31jan2005&

30apr2005&

313jul305&

313oct305&

31jan2006&

30apr2006&

313jul306&

313oct306&

31jan2007&

30apr2007&

313jul307&

313oct307&

31jan2008&

30apr2008&

313jul308&

313oct308&

31jan2009&

30apr2009&

313jul309&

313oct309&

31jan2010&

30apr2010&

313jul310&

313oct310&

31jan2011&

30apr2011&

313jul311&

313oct311&

25

Table1.DescriptiveStatistics.

SCPS(1998-2002)

Pre$Law Post$Law Pre$Law Post$Law

Part+Time+rate 12.76 16.53 8.94 8.93

Age 34.27 34.87 32.39 32.61(3.99) (4.05) (5.09) (5.20)

Head 10.22 12.16 14.8 16.64(0.10) (0.12) (0.15) (0.17)

EducLow 24.68 21.56 27.14 23.4Medium 33.46 24.17 39.4 25.69High 41.86 53.73 33.46 50.9

SectorPrimary 0.53 0.32 0.46 0.24Industry 11.6 12.42 13.18 14.78Construction 0.97 2.04 1.44 1.81Services 86.89 85.22 84.93 83.17

Treated Control

Thesamplecontainsmarriedfemalesunderindefinitecontractbetween25and45.Treatedgroup:Motherswithchildrenbetween0and5yearsold.Controlgroup:Non-mothers.

Table2.Resultsofdirecteffect.SCPS(1998-2002)

Equation)[1] Equation)[2]VARIABLES Part7time Part7time

treated 0.0409*** 0.0516***(0.0101) (0.00914)

after 70.000131 70.0162(0.0128) (0.0227)

treatedafter 0.0351** 0.0286**(0.0161) (0.0150)

Covariates No Yes

obs.)P 0.1272059 0.1272059pred.)P 0.1244879 0.1058416R7square 0.0115 0.084Observations 9,520 9,520Robust)standard)errors)in)parentheses***)p<0.01,)**)p<0.05,)*)p<0.1

Note: Control variables include age, a dummy indicatingwhether the individual is the household head, sector,level of education, tenure, unemployment, partiality,temporary and birth rate by region and year. Marginaleffectsarereported.

26

Table3.Robustnesschecksofdirecteffect.SCPS(1996-2004)

[1]$$$$ [2]$ [3]VARIABLES Part3time Part3time Part3time

treated 0.0462*** 0.0519*** 0.0530***(0.00803) (0.00919) (0.00936)

after 0.000437 30.00549 0.0734(0.0124) (0.0232) (0.0552)

treatedafter 0.00510 0.0327** 0.0592***(0.0142) (0.0154) (0.0159)

obs.$P 0.1125911 0.1260045 0.1300417pred.$P 0.0916341 0.1042417 0.1097903R3square 0.0865 0.0863 0.0816Observations 8,784 8,960 10,066Robust$standard$errors$in$parentheses***$p<0.01,$**$p<0.05,$*$p<0.1

Note:Additionalcontrolsincludeage,adummyindicatingwhetherthe individual is the household head, sector, level of education,tenure, unemployment, partiality, temporary and birth rate byregionandyear.Marginaleffectsarereported.Columns:[1]Placebo:“after”periodincludesyears1996and1997.[2]Treatedgroupincludesonlymotherswithchildrenbornbeforethepassofthelaw.[3] Medium-term effect: “after” period includes years 2003 and2004.

Table4.DescriptiveStatistics.SCPS(1998-2002)

Pre$Law Post$Law Pre$Law Post$Law

Fixed$Term1rate 30.48 34.91 30.55 30.64

Age 32.82 32.91 33.27 33.46

Head 36.82 39.01 80.28 82.17

EducLow 27.74 24.74 42.36 38.88Medium 34.63 22.85 35.49 24.38High 37.62 52.41 22.15 36.74

SectorPrimary 1.05 1 3.87 3.97Industry 11.8 11.16 24.4 23.85Construction 1.2 1.72 11.49 14.89Services 85.95 86.12 60.24 57.28

Treated Control

Individualswithoutchildrenbetween25and45yearsoldTreated:Potentialmothers.Control:Potentialfathers.

27

Table5.Resultsofindirecteffect.SCPS(1998-2002)

[1] [2]VARIABLES Fixed2term Fixed2term

treated 20.000753 0.0197*(0.00981) (0.0110)

after 0.000831 0.000818(0.00880) (0.0101)

treatedafter 0.0435*** 0.0533***(0.0136) (0.0140)

Covariates No Yes

obs.KP 0.3159331 0.3159331pred.KP 0.3157668 0.3007602R2square 0.0013 0.0919Observations 20,115 20,115RobustKstandardKerrorsKinKparentheses***Kp<0.01,K**Kp<0.05,K*Kp<0.1

Note:Controlvariablesincludeage,adummyindicatingwhetherthe individual is thehouseholdhead, sector, level of education,tenure, unemployment, partiality, temporary and birth rate byregionandyear.Marginaleffectsarereported.

Table6.Robustnesschecksofindirecteffect.SCPS(1996-2004)

(1) (2)VARIABLES Fixed2term Fixed2term

treated 0.0319*** 0.262(0.0113) (0.376)

after 0.0113 20.00484(0.0103) (0.0115)

treatedafter 0.0131 0.0457*(0.0149) (0.0252)

obs.EP 0.3155334 0.3201716pred.EP 0.2967836 0.3068873R2square 0.1024 0.0834Observations 16,854 21,907RobustEstandardEerrorsEinEparentheses***Ep<0.01,E**Ep<0.05,E*Ep<0.1

Note:Additionalcontrolsincludeage,adummyindicatingwhetherthe individual is the household head, sector, level of education,tenure, unemployment, partiality, temporary and birth rate byregionandyear.Marginaleffectsarereported.Columns:[1]Placebo:afterperiodincludesyears1996and1997.[2] Medium-term effect: after period includes years 2003 and2004

28

Table7.DescriptiveStatistics.PanelfromCSWH(2004-2011)

Expansion))))))2004-2007

Recession))))))))))))2008-2011

N.)of)observations 8,032 12,227Proportion)of)users 2.46% 2.40%

Age34.253 35.436(5.036) (5.925)

<)30 15.39 15.0230-34 36.89 29.4935-)39 34.85 31.94≥)40 12.87 23.55

Foreign 6.47 10.39

Tenure)(years) 5.143 5.778(3.785) (3.966)

<)2)year 20.06 15.572)-)7)years 57.26 51.44≥)7)years 22.68 32.98

Contribution)groupGraduates,)engineers)and)senior)management 6.47 5.47Technical)engineers,)experts)and)qualified)assistants 6.01 7.17Administrative)and)workshop)managers 2.64 1.87Unqualified)assistants 2.94 2.41Administrative)officers 22.34 18.25

Subaltern 5.39 4.24Administrative)Assistants 25.76 23.42First)and)second)officers 7.11 8.64Third)officers)and)specialists 9.51 11.3Laborers 11.79 16.61Workers)under)18 0.04 0.61

Size)firm<)10 34.77 36.6810-49 24.91 23.850-499 24.02 25.71≥500 16.3 13.81

Sector

Primary 0.36 0.2Industry 11.21 9.52Construction 8.63 11.7Wholesale 0 8.85Housing 0.41 0Administrative 2.29 5Education 5.49 5.11Health 15.29 15Communication)and)transports 26.82 16.88Finances 1.52 1.75Other)services 27.99 24.88

The sample contains females under indefinite contract that reducedtime-scheduleinthesamefirm,thatis,usersofthelaw.

29

Table8.Probabilityofbeinga“39/99LawUser”PanelfromCSWH(2004-2011)

2004$2007 2008$2011VARIABLES user user

<530 $0.00165** 0.00295***(0.000715) (0.000643)

35$539 $0.00168*** $0.00281***(0.000572) (0.000444)

≥540 $0.00913*** $0.00397***(0.000586) (0.000459)

Foreign 0.00559*** 0.00403***(0.00122) (0.000711)

<525year $0.0213*** $0.0226***(0.000523) (0.000397)

≥575years 0.00102 0.0111***(0.000658) (0.000567)

Blue5collar 0.00392*** 0.00753***(0.000609) (0.000446)

10$49 0.000546 $0.00635***(0.000659) (0.000414)

50$499 $0.00377*** $0.00842***(0.000608) (0.000411)

≥500 $0.00284*** $0.0105***(0.000695) (0.000408)

obs.5P 0.0245996 0.0240249pred.5P 0.0209251 0.0185842R$square 0.037 0.0563Observations 326,509 508,931Standard5errors5in5parentheses***5p<0.01,5**5p<0.05,5*5p<0.1

The sample containsmothersunder indefinite contract that stay in the same firm.The two columns estimate the probability of being a lawuser (have reduced timeschedule)along2004-2007and2008-2011,respectively.Referencegroups:belongto30-34agegroup,native,tenurebetween2and7years,whitecollar,smallfirms,healthsectorandworkinginMadrid.Sectorofactivityandfixregiondummiesarealsoincludedintheestimations.Marginaleffectsarereported.

30

Table9.Probabilityofbeinga“39/99LawUser”overthebusinesscycle.PanelfromCSWH(2004-2011)

(1) (2) (3) (4)

VARIABLES user user user user

crisis 50.000554 50.00270*** 50.00262*** 50.00265***(0.000346) (0.000318) (0.000317) (0.000313)

obs.?P 0.0242416 0.0242416 0.0242416 0.0242416pred.?P 0.0242403 0.0197943 0.0197137 0.0193161R5square 0 0.0451 0.0458 0.0498Observations 835,713 835,713 835,713 835,713Standard?errors?in?parentheses***?p<0.01,?**?p<0.05,?*?p<0.1 Thesamplecontainsmothersunder indefinitecontractthatstays inthesamefirm.Groupofage,foreigndummy,tenure,size,contributiongroup,sectorofactivityandfix region dummies are also included in the estimations. Marginal effects arereported.Columns:(1) Withoutcovariates.(2) SamecovariatesthaninTable8.(3) GroupA:cellsofforeign/native,3groupsoftenureandblue/whitecollar.Restofcontrolvariablesincludedintheestimationinavectorofcovariates.(4) Group B: cells of 4 groups of age, 4 groups of size, 3 groups of tenure andblue/whitecollar.Restofcontrolvariables includedintheestimationinavectorofcovariates.