Discussion PaPer series

IZA DP No. 10867

Sara de La RicaLucía Gorjón

Assessing the Impact of a Minimum Income Scheme in the Basque Country

juNe 2017

Any opinions expressed in this paper are those of the author(s) and not those of IZA. Research published in this series may include views on policy, but IZA takes no institutional policy positions. The IZA research network is committed to the IZA Guiding Principles of Research Integrity.The IZA Institute of Labor Economics is an independent economic research institute that conducts research in labor economics and offers evidence-based policy advice on labor market issues. Supported by the Deutsche Post Foundation, IZA runs the world’s largest network of economists, whose research aims to provide answers to the global labor market challenges of our time. Our key objective is to build bridges between academic research, policymakers and society.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.

Schaumburg-Lippe-Straße 5–953113 Bonn, Germany

Phone: +49-228-3894-0Email: publications@iza.org www.iza.org

IZA – Institute of Labor Economics

Discussion PaPer series

IZA DP No. 10867

Assessing the Impact of a Minimum Income Scheme in the Basque Country

juNe 2017

Sara de La RicaUniversity of the Basque Country, FEDEA and IZA

Lucía GorjónUniversity of the Basque Country and FEDEA

AbstrAct

juNe 2017IZA DP No. 10867

Assessing the Impact of a Minimum Income Scheme in the Basque Country*

In this paper we assess the impact of a Minimum Income Scheme (MIS) which has been

operating in the Basque Country, one of Spain’s 17 regions, for more than twenty years.

In particular, we test whether the policy delays entry into employment for recipients. In

addition, we test the efficacy of policies aimed at enabling recipients of the MIS to re-enter

employment. Our results indicate that on average the Minimum Income Scheme, in

addition to preventing social exclusion by providing financial support, does not delay entry

into employment. However, the impact differs from one demographic group to another.

Furthermore, Active Labour Market Policies designed for this group, in particular training,

have a strong positive impact on finding a new job.

JEL Classification: C14, C21, C52

Keywords: minimum income schemes, active labour market policies, poverty, inverse probability weighting, propensity score matching

Corresponding author:Sara De La RicaUniversity of the Basque CountryFacultad de Ciencias Económicas y EmpresarialesAvenida Lehendakari Aguirre, 8348015 BilbaoSpain

E-mail: sara.delarica@ehu.es

* We appreciate comments from Luis Miller, Javier Gardeazabal and María Alvarez on a previous version of this paper. We are also indebted to participants in the “Workshop on Unemployment and Labor Market Policies” organized in April, 2017, in Barcelona, and particularly to Samuel Bentolila, for their comments. The authors acknowledge financial support from the Spanish Ministry of Education and Science (ECO2015-67105-R).

1.IntroductionMost European Union Member States currently provide some form of MinimumIncome Scheme so as to ensure aminimum standard of living for householdswhenthey lackother sourcesof financial support. Theemergenceof these schemesdatesback to 1992, when a European Council recommendation assessed the need todevelop last resort schemes which recognised the basic right of every individual toensure a decent minimum standard of living. These programmes were part ofcomprehensive,consistentplanstocombatsocialexclusion1.Since then, implementation of Minimum Income Schemes (MIS) across EuropeanCountrieshasvariedincoverageandeffectiveness.Themostwidelyusedaretheso-called "simple and comprehensive schemes", which basically cover everyperson/household in need of support, without confining their effects to particularcategoriesofpeople.Since2008,TheEuropeanCouncilhasendorsedtheobjectiveofcombiningadequate incomesupportwith labormarketactivationmeasures soas tofacilitatere-entryofrecipientsintoemployment.AlthoughtheimplementationoftheseschemesisprogressinginmostEuropeancountries,albeitheterogeneously,thereisnosufficientassessmentoftheirimpactonaspectssuchaspovertyreduction,labourmarketparticipationofrecipientsand/ortheimpactofactivationmeasuresontheirrecipientsintermsofre-entry.ExamplesofsuchstudiesincludeGouveiaandRodrigues(1999)andBrunori,ChiuriandPeragine(2009),whoassesstheimpactofparticularMISsonpovertyreductioninPortugalandinasouthernItalianregion,respectively.Additionally,Clavet,DuclosandLacroix(2013)andCheminandWasmer(2012)assesstheimpactoftwoMISs-onepotentiallyimplementedinQuebecandtheotherinAlsace-Moselle,inFrance,onthelabourmarketparticipationoftheirrecipients.Surprisingly,wearenotawareofanystudythatassessestheimpactoflabourmarketactivationmeasuresforMISrecipientsontheirre-entryintowork.Our paper seeks to fill this gap. Specifically, our study assesses the impact of aMinimum IncomeScheme thatoperates inanorthernRegionof Spain - TheBasqueCountry,calledRentadeGarantíadeIngresos.Thisregionpioneeredtheintroduction

1Formoredetails,seeCouncilRecommendation92/441/EECof24June1992:http://publications.europa.eu/en/publication-detail/-/publication/9953c2cf-a4f8-4d31-aeed-

http://publications.europa.eu/en/publication-detail/-/publication/9953c2cf-a4f8-4d31-aeed-6bf88a5407f3/language-en

ofMISinSpainin1989.TheBasqueCountryiscurrentlytheonlySpanishregionwithaSimple and Comprehensive MIS Scheme2. We assess first whether the Basque MISdelaysentryintolabourmarketforitsrecipients.Thenwetesttheefficacyofpoliciesaimed at enabling its recipients to re-enter employment. We do this by using theInverse Probability Weighting methodology, which enables MIS recipients to becomparedwithasimilar,fictitiousgroupcreatedbyweightingnon-recipients.Bydoingso, the treatment is dissociated from individual characteristics and hence pseudo-randomised.Our results indicate that,onaverage, theBasqueMISdoesnot,per se,delayentryintoworkforitsrecipients.Interestingly,however,theimpactdiffersfromone demographic group to another. Furthermore, Active Labour Market PoliciesdesignedforMISrecipients,inparticulartraining,haveastrongpositiveimpactonre-entryintoemployment.

Therestofthepaperisorganisedasfollows:Section2reviewsinstitutionalaspectsofthe MIS implemented in the Basque Country. Section 3 briefly describes relatedliterature.Section4givesadescriptionof thedataandthemaindescriptivesofMISrecipients.Section5presentsthemethodologicalandanalyticalassessmentmethodsandtheempiricalfindings.Finally,Section6summarisesandconcludes.

2.TheMinimumIncomeSchemeintheBasqueCountry

TheBasqueMISwasintroducedin1989,withtheso-calledIntegratedPlantoCombatPoverty3. In the last fewdecades it hasundergone severalmodifications. In1998 itwasgiventherankoflaw,theconceptsof“poverty”and“exclusion”weredefinedandemployment incentives, penalties and infringementswere established. The amountsprovidedandtherequisitesforrecipientshavealsobeenmodifiedseveraltimes.Thelatestmodificationwasimplementedin2011(Act4/2011).Webasethedetailsofourdescriptiononthatversion.

Eligibility Requisites: The first important point to note is that the Basque MIS ishousehold-based. To apply for the aid, applicants must comply with the followingeligibilityrequisites:first,theymustshowthattheirhouseholdincomeisinsufficienttomeetbasicneeds,whichmeans inability toaccess thegoodsandservicesclassedasnecessary for minimum welfare in society according to the Basque Governmentcriterionofpoverty(whichisoutlinedbelow).Thesecondeligibilityconditionconcernsresidency in theBasqueCountry: in principle, the recipient ofMIS in the householdmust be registered on the census and actually have resided in the Basque Country2TheBasqueCountryisasmallregioninthenortheastofSpainwithapopulationof2million(5%oftheSpanishpopulation).Theactivelabourforceisover1millionandtheemploymentrateis50%.TheBasqueCountryisamongtherichestRegionsinSpain,withthesecondhighestGDPpercapitaandthethirdlowestunemploymentrate(12.8%).TheBasqueHumanDevelopmentIndexis0.924,thehighestinthecountry,andatthesamelevelastheNetherlands.3Differentlegislationcanbefoundhere:http://www.lanbide.euskadi.eus/rgi/-/informacion/rgi-legislacion-y-normativa/

withoutinterruptionforthelastthreeyears.IfapplicantscanprovefiveyearsofpaidworkintheBasqueCountrytheresidencerequisitecanberelaxedtooneyearinsteadof three. If none of the above requirements is met, applicants must have beenregistered for a continuous period of five years in the immediately preceding tenyears4.

Furthermore,theMISisconsideredasalastresortscheme,soapplicantsmustalreadyhave applied for all other income aids to which they are entitled. In principle, theschemeiscompatiblewithotherincomeaidsorwagesoffamilymembers,solongasthey do not exceed the defined poverty line. In addition, applicants must own nopropertyotherthantheirhabitualresidence.

Coverage:MISbenefitsaretransferredtofamilyunitsonamonthlybasis.Theamountset by the Basque Government to meet basic necessities varies depending on theminimumwage(MW),thenumberofpeopleinthehousehold,thenumberofretiredpersonsandwhetheritisasingle-parenthouseholdornot.Specifically,itis88%oftheMW for single-member households and can reach 125% of theMW for householdswith threeormoremembers. In thecaseofhouseholdswithat leastonepensionerthosefiguresriseto100%and135%respectively.Single-parenthouseholdsreceiveasupplementarysubsidy5. Ifthereareother incomesinthehousehold,theMIScoversthedifferenceinthatamount.

HouseholdLabourMarketAvailability:Bothholdersandothermemberscohabitinginthesamehouseholdwhoareabletoworkmustcommittobeingavailabletodoso.Inaddition, they must participate in activities that increase their employability. Inparticular, the holder must sign an inclusion-oriented employment improvementagreement. However, although the spirit of the law is that every recipient shouldsearchactively for a jobonly around40%areobserved to receiveany interventionsfrom the public employment service or activating interventions6. We do not knowwhatcriteriathePublicEmploymentServiceusestofollowMISrecipientstomonitortheir activation, i.e. whether individuals are self-selected into different activities orthereissomekindofcompulsoryparticipation.

3.RelatedLiterature

4Forparticulargroups,suchasthosewhoreceiveapublic-sectorpensionorhavebeenvictimsofdomesticabuse,thereisnoneedtoproveworkexperienceandonlyoneyearofresidenceintheBasqueCountryisrequired.5Specifically,theamountin2016variedfrom€625.58forasinglememberhouseholdto€959.70forahouseholdwiththreeormoremembersandatleastonepensioner.Single-parentfamiliesreceiveanadditional€45.6Inparticular,theactivationrateis41.7%forholdersand38.2%fornon-holders.

Verysimilarpolicieshavebeenimplementedinothercountries,thoughfewhavebeenassessed.Furthermore,somepilotprojectsandex-anteorex-postassessmentsofsimilarmeasurestoreducepovertyaroundtheworldcanalsobefound.

The Portuguese Guaranteed Minimum Income scheme (Rendimento MinimoGarantido), set up in 1996, follows a very similar structure to theBasque Renta deGarantíadeIngresos.However,thebenchmarkincomeforthebenefit isverylow,atapproximately50%oftheabsolutepovertylinefigure.GouveiaandRodrigues(1999)providea simulationof its effectonpovertyusing theHouseholdBudget Survey for1994-95. They find that 5% of households and 5.7% of individuals take part in theprogramme.Thecosttothepublicpurseisabout0.18%ofPortugal’sGDPand0.39%oftotalpublicexpenditure,buttheimpactonrecipients’incomeisanaverageincreaseof18.5%intheannualincomeoftheparticipatinghouseholds.Accordingtothisstudy,this policy measure has a modest effect on the reduction of the number of poorhouseholds,butasubstantialeffectonalleviatingtheintensityandseverityofpoverty.Following Beckerman’s model, which analyses the efficiency of income transfers,Rodrigues(2001)estimatestheverticalefficiencyoftheprogramme(astheproportionoftotaltransfersreceivedbyhouseholdsthatwerepoorbeforethetransfers)anditspovertyreductionefficiency(astheproportionoftotaltransfersthatcontributedtoareduction in poverty). The study concludes that the vertical efficiency of theprogramme is 85% and the poverty reduction efficiency is 82%. This means thathouseholds that were initially above the poverty line receive 15% of the total MIStransfers,and18%oftotaltransfersdidnotcontributetoreducingthepovertygap.AsimilaranalysishasnotyetbeencompletedfortheBasqueCountry,butitiscurrentlyinprogress.

ARedditeMinimod’Insermentopilotschemewas implemented inthesmall townofMola di Bari in the south of Italy. Brunori, Chiuri and Peragine (2009) analyse suchissues as (i) eligibility criteria; (ii) targeting choices and results; (iii) distribution andwelfare effect on recipients and on the overall population in the town; and (iv) theincentiveeffectson labourmarketparticipation.Theirmostsignificantfinding is thatthemereuseofanationalmeasureofpovertytendstoobscureindividualsituations,makingitimpossibletodistinguishpoorpeopleinneedofpublicsupport.Bycontrast,thesecondsignificantissueisthatlocaladministrationsseemunabletocorrectlyverifytheincomelevelofhouseholds.Theirstudyalsoshowswhichcategoriesofindividualsare most likely to be activated. The analysis of the MIS shows that a substantialnumberofhouseholds improve theireconomicconditions thanks to theprovisionofpublic funds, even though the coverage rate is insufficient. Finally, they find noevidencetosuggestthat individuals involved intheprogrammetendalsotobecomerecipientsinthefollowingmonths.

An ex-ante assessment of a Proposal in Québec is provided by Clavet, Duclos andLacroix(2013).Everyindividualwouldbeguaranteedanincomeequivalentto80%oftheMarketBasketMeasure.Thestudyfirstestimatesastructurallaboursupplymodeland then simulates the impact of the poverty reduction recommendation by theQuebecCommittee. By predicting labour supply the result shows that the proposedscheme would have strong negative impacts on labour market participation rates,mostly among low-income workers. The so-called Revenue Minimum d’Insertion isassessed by Chemin andWasmer (2012) in Alsace-Moselle in eastern France. Theirestimates, based on double and triple differences, show that the RMI policy isassociated with a 3% fall in employment (among unskilled workers aged 25-55),leadingtoanestimatedlossof328,000jobs;withadeclineinthejob-accessrate;andwitha five-month increase in theaveragedurationofunemployment.Theyalso findconsiderablylargerdisincentiveeffectsforsingleparents.4.TheDatasetandSomeDescriptiveStatistics

4.1.TheDataset

Ourdatasetconsistsofmonthly longitudinal informationonall individualswhowereregisteredwiththeBasquePublicEmploymentServicefromFebruary2015toJanuary2016.Dataarecollectedonthelastdayofeachmonth.Mostofthoseregisteredareunemployed, but some may be employed and searching for another job. Theiremployment status is clearly stated. All MIS recipients and their cohabitants mustregisterwiththeBasquePublicEmploymentServiceasarequisiteforreceivingincomeaid,independentlyoftheiremploymentstatus.

Thedatabaseincludesalltheinformationprovidedbyeachindividualwhenregisteringat the Employment Office, including standard demographic characteristics (gender,age, education level, nationality, postcode and residence, knowledge of otherlanguages), aswell as labourmarket information (previous employment experience,occupational and geographical searches, unemployment duration, etc.). The BasquePublic Employment Service also provides exact information on whether individualsreceiveorhavereceivedunemploymentbenefits(entitledbenefits,assistancebenefitsand/orMIS) and on the duration of entitlement. Finally, the database also recordsinformation on the assistance measures from the public employment services thatunemployed workers have received in the last 12 years to enhance job access.Informationsuchasthetypeofmeasure,numberofhoursandstartandenddatesareprovided.

BasquePublicEmploymentServiceinSpaindividethepoolofunemployedworkersontheirfilesinto"RegisteredUnemployed"and"OtherUnemployedWorkers".Thelattercategory,whichaccountsforaround22%ofallunemployedworkers,includesretireesand pensioners, those not immediately available for work, those registered in the

current month, those who just seek particular kinds of work such as outwork andteleworkingand thosewhoseekwork forunder20hoursaweek.Studentsarealsoincludedinthiscategory.Werestrictouranalysistothe“RegisteredUnemployed”,i.e.thosewithoutajobwhoareseekingworkandimmediatelyavailableforany"regular"job.

TheBasqueCountryhas recordsofaround60.000MIS recipientseachmonth in theperiod under analysis, equivalent to 25% of those registered as unemployed in theBasqueCountry.

In spiteof the richnessof informationof thedataset,an importantdrawback is thatthereisnohouseholdidentifierforMISrecipients.Hence,allwecanassessiswhetheran individual is a MIS holder or not. Hence, although the MIS is provided at thehouseholdlevel,thewholeanalysisisconductedatindividuallevelfordatarestrictionreasons. An additional caveat of the data is the lack of information related tohousehold income, and in particular, to the specific amount of any type of benefitsreceivedbytheunemployed.

4.2.StatisticalDistributionofMISRecipientsvsnon-MISRecipients

To give a precise idea of the differences betweenMIS recipients and otherworkersregistered as unemployed, we present the distribution of each of the two groupsunderatotaloffourcharacteristics:gender,age(<30,30-44and>44),educationlevel(primaryatmost,secondaryandhighereducation)anddurationofunemployment(<3months,3-6months,6-12months,12-24and>24months).Figures1and2showthedistributionofMISandnon-MISrecipients,respectively,acrossthefourcharacteristics.Wedo this for aparticularmonth -October2015– to get abetter idea innotonlyrelativebutalsoabsoluteterms.Anyothermonthfromthesamplewouldgivealmostidenticalpatterns.

[InsertFigures1and2here]

At first sight, the profile for education level and unemployment duration of MISrecipients is quite different from that of the rest of the unemployed. This is notsurprising given thatMIS is seen as a last resort scheme. In particular, 60% ofMISrecipientshavenosecondaryeducationqualificationsandmorethanhalfhavespentmorethantwoyearsunemployed.Theequivalentfiguresarebarelyonethirdandonefourth,respectively,fornon-MISrecipients.Moreprecisely,thebiggestgroupamongrecipients is thatof the very long-termunemployedagedover30withonlyprimaryeducation. This group accounts for a third of all MIS recipients. Among non-MISrecipients the equivalent group accounts for barely 10%. Furthermore, regardless ofeducationlevel,MISrecipientsover30whohavespentmorethantwoyears lookingfora jobaccountfor50%.Focusingontheyoungestgroup, itcanbeseenthatmorethan half have spentmore than two years seeking employment and 70% have only

primary education. However, the pattern is very different among thosewho do notreceiveMIS: those who have been unemployed for a very short time are generallyyoungpeoplewithsecondaryorhighereducation. It is important tonote thatmanyyoungpeoplewithhighereducationcontinuestudyingiftheydonotfindajobandarenot therefore considered as unemployed. This behaviour is not found amongunemployedpeoplewithlowereducationlevels,whoarepreciselythemostcommongroupamongMISrecipients.

4.3.MonthlyExitRatesfromUnemploymenttoEmployment(JobFindingRates)

We now describe the patterns of monthly job-finding rates for recipients and non-recipientsofMIS,makinguseofthelongitudinalnatureofourdataset.Wedefine“exitintoemployment”asatransitionfrom“registeredunemployed”inthecurrentmonthtoalabourstatusof“employed”inthenextmonth7.Therefore,thecharacteristicsoftheunemployedpeoplearefixedinthecurrentmonth.Followingthesamestructureas above, we characterise job-finding rates by comparing recipients with non-recipientsofMISusingthesamefourcharacteristics,i.e.gender,age,educationleveland unemployment duration. Given that we observe unemployed people fromFebruary2015toDecember2015wecompute job-findingrates fromMarch2015toJanuary2016.

Onaverage,themonthlyjob-findingrateforMISrecipients is3%.This issignificantlylower than the rate for non-MIS recipients,which is 9%. Figures 3 and 4 show job-finding rates forMIS andnon-MIS recipients, respectively, fordifferentprofiles. It isimmediatelyapparentthatjob-findingratesincreasewitheducationlevelandstronglydecreasewithunemploymentdurationforbothgroups.Togivesomenumbersonthestrong negative association between unemployment duration and job-finding rates,Figures3and4showthatindividualsunemployedforlessthanthreemonthshaveanaverageexitrateof11%,whiletheverylong-termunemployed(overtwoyears)havearateofonly1%.Interestingly,60%ofMISrecipientsbelongtothegroupofverylong-termunemployed.Anotherpointtonoteisthatalthougheducationlevelisrelevanttounderstandingdifferencesinaccesstojobs,itisfarlesssignificantthanunemploymentduration:theexitrateofMISrecipientswithhighereducationaverages5%,comparedto2%amongthosewithprimaryeducationonly.

Figure4focusesonthecomparisonbetweenMISrecipientsandnon-MISrecipientsonjob-findingrates.Asmentionedabove,thereisadifferenceof6percentagepointsonaverage between the job-finding rates of the two groups. However, that differencevariesmarkedlydependingon individualprofiles.Forexample,amongtheveryshort

7Inparticular,BasquePublicEmploymentServicedoesnotremovenon-MISrecipientsfromtheregisteroncetheybecameemployedunlesstheindividualspecificallyasksforit.Inthatway,weensurewearecapturingmostre-entriesintoemploymentfornon-MISrecipients.NotethatforallMISrecipientsitsregisteriscompulsory.

termunemployedthereisadifferenceof7.5points,whileamongtheverylongtermunemployedthedifferenceisbarelyonepercentagepoint.

[InsertFigures3and4here]

4.4.Determinantsoftheprobabilityoffindingajob:MISrecipientsvsnon-recipientsFinally,weestimatetheprobabilityoffindingajobbythelastdayofeachmonthforallthoseregisteredunemployedonthelastdayofthepreviousmonth.Asabove,wecalculate the probability of finding a job from March 2015 to January 2016. Thedependentvariable,therefore,takesavalueof1iftheunemployedpersongetsajobinthenextmonth,and0otherwise.

Toperformthisexercise,wetakeintoaccountallobservablevariablesthatmayaffectthe employability of people registered with the Public Employment Service. Inparticular, we include demographic characteristics such as sex, age, nationality,disability, education and language skills; job characteristics such as requestedoccupations,experience,activityinthepreviousfieldofwork,unemploymentduration,geographicalscopeofthenewjobsearch,month(s)inwhichtheindividualisobservedas unemployed, whether individuals are MIS holders or not 8 , and province ofregistration.

We add a dummy indicating whether individuals have ever been referred to socialservices.Thereceiptofbenefitsinthecurrentorinpreviousmonthsisalsoincluded.We include in our estimation an indicator for whether individuals have receivedactivationservicesat leastonce in the last sixmonths.40.7%ofMIS recipientshavereceivedsomekindofmeasureinthelastsixmonths,ascomparedto13.75%ofnon-recipients9 . We divide activation service into the following categories: guidance,monitoring,informationonself-employmentandtraining.

Table 1 presents the results of the estimation (marginal effects are shown) using apooled probit model with month and province fixed effects10 . The first columnestimates the probability of finding a job for MIS recipients and the second doeslikewisefornon-recipients.Notethatthisestimationdoesnotaccountforunobservedheterogeneity. It should be taken as a preliminary view of the importance of thecharacteristicsofunemployedpeopleinthejobsearchprocess.

[InsertTable1here]

8NotethattheMISisprovidedathouseholdlevel,whereastheanalysisisconductedatindividuallevel.Thedatadonotenableustoidentifyeachofthehousholdmembersandtheirlabourmarketsituation,onlywhethertheyareMISholdersornot.9ThesefiguresareforOctober.Anyothermonthfromthesamplewouldgivepercentages.10Populationaveragewitharobustestimatorofthevarianceisused.

Themostnoteworthyresulthasbeenalreadyanticipated:unemploymentdurationisthevariablethatmostaffectstheprobabilityofexitingunemployment.Thechancesofentering employmentdecreasedramatically as the time forwhich aperson remainsunemployedincreases.Thelargestdecreaseintheprobabilityofgettinga joboccursafterthebarrierof3months(referencegroup)withareductionof5percentagepointswhenindividualsareunemployedforbetween3and6months.Beingunemployedforbetween6monthsand1yearreducesthelikelihoodbyonepoint(6.5pointslesslikelythan for those unemployed for less than 3 months) and for those unemployed forbetween 1 and 2 years the probability falls by 1.6 points (8 points less likely). Thenegativeimpactincreasesto9pointsifthedurationofunemploymentgoesbeyond4years.Acomparisonoftheseresultswiththeimpactofthesamevariableonthetotalnumber of unemployed people who do not receiveMIS (column 2) shows that theduration of unemployment also has the greatest negative impact in this group. Inparticular,beingunemployedformorethan3monthsreducestheexitprobabilitybyalmost 8 percentage points. As occurs with the MIS group, the likelihood of exitcontinuestodecreaseasthedurationofunemploymentincreases,withexitbeing15.7pointslesslikelyamongthoseunemployedformorethan4years.Ascanbeseen,noothervariablehasasimilarimpact.

Consideringlevelsofstudies, ingeneralthelikelihoodoffindingajobcanbeseentobe correlated with the education level of each unemployed individual: havingsecondaryeducationqualifications(comparedwithprimaryornoeducation)increasesthe probabilitymidpoint; completing high school increases it by 0.8 points;mediumlevelvocationaltrainingincreasesitby1.2pointsandhigherlevelvocationaltrainingand higher university degrees raise it by 1.9. Notice that the impact of beingunemployed formore than 3months is double that of having university studies (ascomparedtoprimaryornoeducation)forMISrecipients.

A separate section below is dedicated exclusively to a counterfactual assessment oftheimpactofActivationServicesontheprobabilityoffindingajob,soherewepresentonlyapreliminaryassessmentofactivationinterventions.Itisimportanttonotethatinformationonself-employmenthasaclearlydifferentiatednature,sincepeoplewhouse it are practically on their way towards self-employment. Thus, measuring itseffectivenessviaitsimpactontheprobabilityofleavingforajobdoesnotmakemuchsense.Fromnowon,weassesstheeffectivenessofonlytheotherthreeinterventionsinexitsintoemployment.

5.AssessingtheImpactoftheBasqueMinimumIncomeSchemeontheLabourMarket:ACounterfactualAssessment

Any Minimum Income Scheme is by nature a passive policy, as its main aim is toguarantee all individuals the resources required to meet their minimum needs.

However,asmentionedabove,theBasqueMIS,followingthedictatesoftheEuropeanCouncilsince2008,requiresrecipientstoparticipate(inprinciple)inactivepoliciestomaketheirre-entryintoemploymentasfastandsuccessfulaspossible.Inviewofthistwo-foldscopeoftheMIS,withbothpassiveandactiveaspects,ourassessmentofthepolicyisalsotwo-fold.

Firstly,althoughthegoalofanypassivepolicyisnottoacceleratetheemployabilityoftheunemployedbuttosupplementtheir incomesoastoalleviatepoverty,empiricalevidence generally finds that most income transfers to the unemployed result in adelay in job-finding. Reservationwages increase for anyonewho receives additionalincome,andthistypicallydelaysjobentry,henceloweringjob-findingrates.However,therearetwoaspectsoftheMISwhichmightaccelerateratherthandelayjobaccess:oneisthattheMIScanalsobereceivedbyemployedworkerswithinsufficientincometomeetminimumneeds,soMISrecipientsmightbewillingtoacceptjobswith"low"wagescompatiblewithretainingthetransfer.TheotheristhatrecipientscanlosetheirMIS if it is proved that they have rejected job offers. For these reasons, the typical"delay" effect of a passive transfer such as theMISmaybepartially offset by somekind of “acceleration effect” for reasons other than the activation measuresimplemented.

OurfirstassessmentwithrespecttotheimpactoftheMISintheBasqueCountrylooksatwhethertheMIScausesadelayoranaccelerationeffect,andifsoonwhatscale.Thisisthefirstobjectiveaddressedinthissection.

Secondly, andperhapsmore interestingly,we seek to assesswhether activepoliciesofferedtoMISrecipientsmakeforbettertransitionstowardsemployment.Thisisthesecondobjectiveofthesection.

5.1.EmpiricalAssessmentStrategy

In both analyses the aim is to assess the impact either of the MIS itself or of theactivation measures aimed at MIS recipients on the probability of exitingunemployment.As inpreviousestimations,thedependentvariable(Y)takesavalueof1 if theunemployed individualgetsa job inthenextmonthand0otherwise.Thetreatment(D),whichisadummyvariable,takesavalueof1firstlywhentheindividualisanMIS recipientandsecondly if the individual receivesactivationmeasures11.Thecovariatesincludedinouranalysesarethesameasinpreviousestimations(X).

Themainproblemthatwefaceinboththeanalysescarriedoutinthispaperissampleselection.Inthefirstoneunemployedpeopleneedtocomplywithstrictrequirementsto receive MIS. In the second analysis, the profile of the unemployed people who11Weareunabletodrawupadurationanalysisbecauseofdatalimitations.Ourdatasetincludesonly2015informationandforMISbeneficiaries(morethan70%ofwhomhavebeenunemployedformorethanoneyear)wewouldneedlongerlongitudinalinformation.

receive activation measures differs broadly from that of non-activation measuresrecipients (as shown below). Consequently, given that individuals are not randomlychosen,ameandifferencebetweentheoutcomesoftreatedandcontrolgroupcannotbeusedtolearnthecausalityinthecorrespondingtreatment.Onlywhenparticipationinthetreatmentdependsonobservablecharacteristics(X)cantheAverageTreatmentEffectontheTreated(ATT)beestimatedbyconditioningonthesevariables,renderingthecounterfactualoutcomeindependentofthetreatment(conditional independenceassumption, CIA). However, the probability of finding a job for recipients and non-recipients ofMISmight be affected by confounding factors. Therefore, it is hard tojustify the validity of CIA in this analysis. In the second analysis, our lack ofunderstandingof the selectionprocess for receivingactivationmeasuresmeans thatweareunabletoargueastowhetherCIAissatisfiedornot.

Propensity Score methods are useful for estimating treatment effects usingobservationaldatasincetheyenableobservationalstudiestobedesignedalonglinessimilartorandomisedexperiments(Rubin,2001)12.RosenbaumandRubin(1983)showthat instead of conditioning on the covariates, conditioning on the probability ofpotential treatment conditional on observable covariates, the propensity score(𝑝 𝑥 = 𝑃 (𝐷 = 1/𝑋)), suffices to achieve a balance between the treatment andcontrol groups as long as other requirements are met. Firstly, the covariatesinfluencing assignment and outcome should not predict the treatment participationdeterministically(weakoverlap,𝑃 𝐷 = 1/𝑋 < 1forallX).Secondly,theparticipationinthetreatmentofoneindividualmustnothaveanimpactontheoutcomeofothertreated or control individuals. Our two samples confirm the weak overlap.Furthermore, it seems reasonable to think that being an MIS recipient or servicerecipientdoesnotaffectotherpeople’sprobabilitiesoffindingajob.Forthesereasons,webelievetheuseofPropensityScoretechniquestobeappropriate.

Different propensity score approaches have been suggested for estimating anadequatecounterfactualoutcome.Themostwidelyusedmethodsarematchingandreweighting (Imbens, 2004). These methods seek to remove observed systematicdifferences between treated and control subjects. In our first analysis, InverseProbabilityWeighting(IPW)makesthedistributionofobservablecovariatessimilarin

12Anotheralternativethathasbeensuggestedtousistheregressiondiscontinuityapproach,usingasacontrolgroupthosehouseholdsthatareclosetofulfillingthetotalincomerequirementsofthehouseholdtoreceivetheMISbutdonotcomply.Unfortunately,thismethodologycannotbeusedhereforseveralreasons.Firstly,aspreviouslymentioned,wedonothaveinformationonthetotalhouseholdincome,therefore,wecannotmeasurehowfarindividualsorhouseholdsaretocomplywiththeincomerequirementtoreceiveMIS.Secondly,wearenotabletoidentifythoseindividualsbelongingtothesamehousehold,thus,matchingMIShouseholdswithnon-MISonesisnotpossible.Finally,thelackofothercrucialhouseholdvariables,suchasthenumberofchildrenorotherdependentsinthehouseholdisnotavailableeither.

the treated and control groups.13Furthermore, as explained below, IPW is the onlyvalidmethodologyinourfirstanalysisduetothecharacteristicsofthetreatment.Forthesecondpartofourresearch,ourlackofknowledgeoftheselectionmechanismandthe characteristicsof the sampleassessed leadsus to calculate the treatmenteffectusingtwodifferentmethods:InverseProbabilityWeighting(IPW)andPropensityScoreMatching(PSM).

The idea behind Inverse ProbabilityWeighting is the following: random assignmentguarantees that thedistributionof the covariates amongunitsofobservation in thetreatment and control groups is probabilistically equivalent, i.e. all units are equallylikelytobeinthetreatmentorcontrolgroups.However,whentheassignmentisnotrandomsomeindividualsaremorelikelytobetreatedthanothers,dependingontheirparticular characteristics. To account for these differences in the regressionformulation observations must be weighted according to the inverse probability ofreceivingtreatment.Thisgivesapseudo-randomsamplebyweightingobservationsbytheinverseoftheprobabilityofbeingtreated.Therefore,thedistributionofcovariatesbetweenthegroupswouldbeprobabilisticallyequivalent(GardeazabalandVega-Bayo,2015).Inshort,weightingindividualsbytheinverseprobabilityoftreatmentcreatesasynthetic sample where treatment assignment is independent of the observedcovariates. Inverse Probability Weighting enables unbiased estimates of averagetreatmenteffectstobeobtained.However,theseestimatesareonlyvalidiftherearenoresidualsystematicdifferencesinobservedvariablesbetweentheweightedtreatedandcontrolgroups(AustinandStuart,2015).Weprovethistobethecasehere.It isthus assumed that when the observable differences are reduced, so are theunobservable factors. It stands to reason that a more efficient estimator can beobtainediftheregressionofthereweightedsampleincludesallmeasuredcovariatesas additional regressors. This other estimator is known as Augmented InverseProbabilityWeighting(AIPW).

The IPW estimator uses a two-step approach to estimate treatment effects. ThespecificationfortheAverageTreatmentEffectontheTreated(ATT)isasfollows:

1) Estimate the probability of being treated based on the covariates by a probit14regression.Denote𝑝!(𝑥),i.e.thepropensityscore.Usetheinverseprobabilityweightstocomputethenewpseudo-randomsample.Buildregressionweights(𝑤!)as:

𝑤! = 1 𝑖𝑓 𝐷! = 1

𝑤! =𝑝!(𝑥)

1− 𝑝!(𝑥) 𝑖𝑓 𝐷! = 0

13Table2belowshowsthedistributionofthecharacteristicsinthereweightedsample.14 Alogitmodelcanbealsoused.

Theideabehindthisreweightingprocedureisquitestraightforward.Theobjectiveistoapproximate the distribution of the covariates of the control group to those of thetreated group. For that reason all treated individuals have weights of 1. Controlindividualswitha0.5probabilityofbeingMIS recipientsareassignedaweightof1;those with a probability higher than 0.5 have weights of more than 1 with anincreasing pattern and thosewith a probability lower than 0.5 haveweights of lessthan 1 with a decreasing pattern. By doing this, the outcome of those controlindividualswiththehighestprobabilitiesofbeingMISrecipientswouldgraduallyweighmoreandtheoutcomeofthosecontrolindividualswiththelowestprobabilityofbeingMISrecipientswouldweighexponentiallyless.

2)CalculatetheATTofthenewsample,i.e.runaprobitregressionoftheoutcomeona constant and the treatment using the weights calculated. The coefficient of thebinarytreatmentinthepreviousregressionisaconsistentestimationofATT,providedthat the propensity-score is correctly specified. Adding all confoundersmeasured asadditionalcovariatestheAugmentedInverseProbabilityWeighting(AIPW)estimatorisobtained.

In the second assessment, an additional Propensity Score approach is applied:Propensity Score Matching (PSM) here helps us also to estimate the impact ofactivationmeasures.Thismethodologyentailsmatchedsetsoftreatedanduntreatedsubjects who share similar propensity scores (Rosenbaum and Rubin, 1985), and itenablestheATTtobeestimated(Imbens,2004).Themostcommonimplementationisone-to-onepairmatching,inwhichpairsoftreatedandcontrolindividualsareformedinsuchawaythattheyhavesimilarpropensityscores. Onceamatchedsamplehasbeenformed,thetreatmenteffectcanbeestimatedbydirectlycomparingoutcomesbetweenmatched treated and control individuals. Schafer and Kang (2008) suggestthattreatedandcontrolsubjectsshouldberegardedasindependentwithinmatchedsamples.Bycontrast,Austin(2011)arguesthatthepropensityscorematchedsampledoes not consist of independent observations.Hemaintains that in the presence ofconfoundingfactorscovariatesarerelatedtooutcomes,somatchedsubjectsaremorelikelytohavesimilaroutcomesthanrandomlyselectedsubjects.

BasedonAustin’sargument,werejecttheuseofthePropensityScoreMatchinginthefirst analysis. Non-observed factors such as family income differ systematicallybetweenthetreatedandcontrolindividualsastheyarecrucialdeterminantsforbeingselectedforthetreatment.However,thesecondassessmentusesPSM,aswefind itreasonable toargue that theunobservable factorsof treatedandcontrol individualsresemble each othermore (given the selected control group used) than in the firstanalysis.

5.2.ImpactofMISonjob-findingrates-DoesMISreducetheprobabilityoffindingajob?

Asshown inprevioussections,MISrecipientshaveamonthly job-findingrateof3%,comparedto9%forthenon-MISunemployedgroup.However,asalreadystated,thecompositionofthegroupofMISrecipientsdiffersnotablyfromthatoftherestoftheunemployed,andthosedifferences(mainlylongerunemploymentdurationandlowereducationlevel)maybecausingatleastpartofthedifferencesobservedinjob-findingrates. To isolate compositional differences from the income scheme, we use theInverse Probability Weighting Methodology as detailed above. This enables us toassesstheextenttowhichthedifferenceobserved in job-findingratesareexplainedby(i)compositionaldifferencesbetweenthetwogroups;and(ii)bytheMIS.

Tothatend,weincludeinthetreatmentgroupallthoseindividualswhoarerecipientsoftheMISinthecurrentmonth.Giventhattheobservationunitisoneindividualpermonth, an individualmay belong to the treatment group in somemonths (inwhichhe/shereceivestheMIS)butnotinothers(inwhichhe/shedoesnotreceiveit).Hence,an individualmay belong to the treated group in a givenmonth and to the controlgroup inanother.Tosetupanadequatecounterfactual,wemustdefine thecontrolgroupsothatitprovidesthebestpossiblesimulationofjob-findingratesforthegroupofMISrecipientsiftheyhadnotreceivedthebenefit.Accordingtothedata,for93%ofMIS recipientsMIS is theONLY incomeaid received;a further6%also receiveotherwelfarebenefitsand the remaining1% receivecontributorybenefits. In the last twosituations,theyreceivebothtypesofincomeaidbecausetheotherbenefitsreceivedarestilllowerthanwhatitisconsiderednecessarytomeetbasichouseholdnecessities.Wethinkthatitmakessensetoassumethatiftheincomeschemedidnotexistthe93%currently receivingonlyMISwouldnotbegettinganyadditional incomeaidandtheremaining7%wouldreceivean insufficientamount.Forthisreason,wehavechosento include unemployed individuals who do not receive ANY benefit in the currentmonthinthecontrolgroup15.Forthisgroup,theobservedmonthlyjob-findingrateis6.5%. Consequently, the outcome of the assessment must be interpreted as thedifferential impact of MIS on the job-finding rate compared to not receiving anybenefit.

However, the treatment (receivingMIS) isbynomeans random.As specifiedabove,there are specific requirements. Some of them are observable in our dataset butothersarenon-observedconfoundervariables, suchas totalhousehold income, thatmustbecontrolledfor.To"correct"forthesedifferencesbetweenthetreatmentandcontrol groupsweuse the InverseProbabilityWeightingmethod.Table2 shows thedistribution of the reweighted control group, which validates the use of the IPWmethodology. This table shows that the differences in the main characteristics areeliminatedbyusingthesaidweightingprocedure.

15InOctoberthisgroupconsistsofatotalof69,961unemployed,comparedwith38.345unemployedMISrecipients.

[InsertTable2here]

TheresultsoftheInverseProbabilityWeightingEstimationandofanextendedversionofit(theAugmentedInverseProbabilityWeightingEstimator)arepresentedinTable3.Applying such methodology, we find that the impact of MIS is not significantlydifferent fromzeroatanysignificance level.Theresult is thesameforboththe IPWand the AIPW estimators, which makes it more reliable16. This indicates that themonthly job-findingprobability forMIS recipientswouldhavebeen the same if theyhadnotreceivedanybenefit.WecanthusconcludethattheMISitselfdoesnotreducetheprobabilityoffindingajob.Inotherwords,thedifferencesobservedinjob-findingratesbetweenthetreatmentandthecontrolgroupareduesolelytothedifferenceinthecompositionsofthetwogroupsandnottotheeffectofthepolicy.

[InsertTable3here]

As a second step, we analyse whether the MIS has different impacts on differentdemographic groups. Specifically, we assess the impact ofMIS onmen andwomenseparately,onthreeagegroups(<30,30-44and>45)andonthreeeducationgroups(primary,secondaryandhigher)17.Theresults,presentedinTable4,confirmthattheimpact of MIS is not homogeneous across demographic groups. In particular, forwomenMISdelaysexittoemploymentslightly(0.2p.p)whereasithasnoimpactonmen.Second,theMISaccelerates job-findingforolderworkers (0.2p.p)whereasforyoungworkers(<30)itdelaysexittoemployment(1p.p).Finally,wefindadelayasanimpactofMISforlesseducatedworkers(0.2p.p),whereasitacceleratesjobentryforthose with more than primary education (0.2 p.p for workers with secondaryeducationand0.5p.pforthosewithhighereducation).

[InsertTable4here]

Ourresultscoincidepartiallywiththeex-anteassessmentinClavet,DuclosandLacroix(2013) and with the findings (double and triple difference estimation strategy) inChemin and Wasmer (2012). Both find a negative impact on labour marketparticipation,particularlyamongspecificgroupssuchaslow-skilledworkers.However,theirresultsarenotdirectlycomparabletooursasthemethodologyandthedesignofthepoliciesintheregionsthattheyexaminearedifferent.Toourknowledgethereisnocomparableassessmentofasimilarpolicy.The main conclusion of this exercise is as follows: by definition, the MIS reducespoverty and promotes social cohesion. Our analysis leads us to conclude that on

16TheassessmentisalsoconductedusingthePropensityScoreMatchingmethodology.However,theresultsaredivergent,corroboratingtheargumentofAustin(2011).17Thesameanalysisisnotcarriedoutfordurationofunemploymentbecauseoftheendogeneityofthevariable.

average the MIS per se does not delay exit to employment. However, we do finddifferences in its impactondifferentdemographic groups. Inparticular, it causesanundesireddelayeffect(alsocommonlyfoundinotherpassivepolicies)forwomen,thelesseducatedandyoungpeople,butacceleratesentry intoemploymentformediumandhigh-educatedworkersandforthoseagedover45.

5.3.TheImpactofActivePoliciesonjobfindingprobabilityforMISrecipients

InthissectionweassesstheeffectivenessoftheactivationinterventionsreceivedbyMISrecipients.Suchanassessmentishighlyrecommendedgiventhatingeneralactivepoliciesarequitecostly. Itenablesus tocheckand ifnecessarymodifyand improvetheefficiencyoftheBasquePublicEmploymentServiceinprovidingrecipientswiththetoolsthattheyneedtore-enteremployment.Thisinformationcancertainlyhighlightwhatactionsshouldbestrengthened,modifiedoreveneliminated.

RecallthatwefocusonthreetypesofActivePolicy:guidance,monitoringandtraining.Firstofall,individualsareclassedasusersofactivationservicesiftheyareobservedtohavereceivedsuchmeasuresatleastonceinthelastsixmonths(includingthecurrentmonth).

Secondly,wepresentsomedescriptivestatistics toshowtheextentofactivation forthe MIS group. As in the descriptive section, we focus (in order to present thecharacteristicsoftheunemployed)onaparticularmonth(October2015)soastoavoidoverrepresentationof the long-termunemployed.Of the38.345unemployedpeopleregisteredasMIS recipients in thatmonth,15.630had received somekindofactivepolicy intheformofguidance,monitoringortrainingatsometimeintheprevious6months.Thisamountsto40.8%ofthetotal.Asregardsthetypesofservicesreceived,15,106 people (39,4% of all unemployedMIS recipients) received guidance services,265 (0.7%)monitoringservicesand881(2.3%)trainingcourses.Thismeansthat728individualsreceivedmorethanonetypeofservice.Giventhelowfigureformonitoring,from here on we focus our results on activation through guidance or traininginterventions.

Abriefprofileisgivenbelowofhowindividualsinvolvedineachofthesetwopoliciescompare to individuals who receive no activation measures. Table 5 shows thedistributionof the fourmain characteristics (sex, age, educationandunemploymentduration)dependingonthetypeofactivepolicyreceived.

[InsertTable5here]

In general men receive more activation than women: around 65% of those whoreceived trainingweremen. Theage range variesdependingon the typeof service.Guidance and training predominate in the 30-45-age range (their relative incidenceamongMISreceiversis46%).Ingeneral,youngpeopletendtoreceivefeweractivation

interventions.Therearealsosubstantialdifferencesbetweeneducationlevels:60%ofMIS recipients have at most primary education, 27% secondary and 13% highereducation,whichmeans that on average fewer activationmeasures are received byhighly educated MIS recipients. In addition, activation measures decrease asunemploymentdurationincreases.

Furthermore, we find distributional differences per type of activation measure.Guidance measures are distributed similarly across education levels, but we findsignificant differences in training measures, as recipients with secondary or highereducation levels receive more training measures than those with at most primaryeducation.

To assess the impact of each of these activation interventions, we place thoseMISrecipientswhohave receivedeachparticularactivationpolicybeingassessed (eitherindividualguidanceortraining)inthelastsixmonthsinthetreatmentgroup.Asbefore,wemeasure the impactof receiving the activationmeasuresonmonthly job-findingrates. As a control group we useMIS recipients who have not participated in ANYactivationmeasuresfromthePublicEmploymentServiceinthelastsixmonthssoastogetacleanerimpactofeachspecificactivationmeasure.Theresultsmustthereforebeinterpreted as the impact of the intervention on the probability of finding a jobcomparedtonotreceivinganyactivationserviceinthelastsixmonths.

As shown in Table 5, the treatment and control groups differ in importantcharacteristicssuchasthedurationofunemploymentandeducationlevel.WeassesseachinterventionfollowingtheIPWmethodologydescribedabove.Theinterventionsarethus"pseudo-randomised",sothedistributionofthecovariatesbetweenthetwogroups is balanced and the treatment is probabilistically equivalent. Therefore, theimpactofeachtypeofinterventioncanbeproperlyassessedwithouttheresultsbeingbiasedbydifferencesincomposition.

InadditiontotheIPW(andAIPW)method,wealsouseaPropensityScoreMatchingtechniquetoenhancerobustness.Giventhat thecontrolgroupnowconsistsofMIS-recipients(althoughtheydonotreceiveactivationmeasures),wefinditreasonabletoassumethatunobservedconfoundingfactorsoftreatedandcontrolindividualsdonotdiffer substantially from one group to the other. This assumption is essential tovalidatetheuseofthePropensityScoreMatchingtechnique.

The resultsof theassessmentofeachactivepolicy forMIS recipients (guidanceandtraining) are shown in Table 6. Inverse Probability Weighting (IPW), AugmentedInverse Probability Weighting (AIPW) and the Propensity Score Matching (PSM)estimators are presented. The first three columns correspond to the threespecifications for the impactof guidance service. It canbe seen thatguidancehasapositive impactonexit intoemployment.This impact is statistically significant forall

threeapproaches,although itsmagnitudediffersslightlyfromonetotheother.Asageneralresult,weconcludethatguidanceincreasestheprobabilityofgettingajobbyabouthalfapercentagepointovernotreceivinganyactivationinterventioninthelastsixmonths18.

The last three columns in Table 6 show the impact of training programmes on job-findingrates.Unfortunately,wehavenoinformationonthetypeoftrainingprovidedor on whether there is any selection process prior to participating in a trainingprogramme. Given this information limitation, all that we can assert is whetherparticipatinginanykindoftrainingprogrammehelpsindividualsfindajob.WhatwefindisthattrainingisundoubtedlythefactorwithgreatestimpactontheprobabilityoffindingajobfortheMISgroup.Individualswhousetheseprogrammesincreasetheirlikelihoodoffindingajobbyaround3percentagepoints.Giventhattheaveragejob-finding rate for MIS recipients is 3%, the probability of finding a job increases byaround 100%when an unemployedMIS recipient attends a training course. Due totheir potential for job-finding, it would be most helpful to have more detailedinformationregardingtrainingprogrammessoastoassessinthefuturemorepreciselywhichtypesoftrainingprogrammeseemtoworkbest.

[InsertTable6here]

In line with the literature on Active Labour Market Policies, we also find that anadequatedesignofactivationpoliciesacceleratesre-entryintoemployment.Inshort,activepoliciessignificantlyacceleratetheprobabilityoffindingajobforMISrecipients.However,onlyaround40%ofthemusesuchmeasures,eventhoughparticipation inthemissupposedlycompulsory.Specifically,trainingisthemosteffectivepolicy:thosewho undergo it are twice as likely to find a job. This conclusion emphasises theimportanceoflinkingpassivepolicieswithactivepolicies,becausethoseMISrecipientswho use active policies enhance their chances of finding a job compared to similarunemployedpeoplewhodonotreceiveanyaid.

6.SummaryandConclusions

In theBasqueCountry (a region innorth-easternSpain)aMinimum IncomeSchemehas been in place continuously since 1989. Its main objective is to guarantee allindividuals the resources required to cover their basic necessities, and at the sametime to provide for their progressive integration into society and employment.Furthermore,inlinewithEuropeanCouncilrecommendations,theBasqueMIShasaninteresting feature: recipients are in principle required to participate in activemeasurestomaketheirre-entryintoemploymentasfastandsuccessfulaspossible.

18Theimpactofguidanceisalsoaddressedforthepopulationsubgroups.Theresultsarenotshownhereasallprofileshavesimilarresults,sotheyaredeemedtobeoflittleinterest.Trainingprogrammesarenotassessedforthedifferentpopulationgroupsforreasonsofsamplesize.

In2015therewereabout62,000MISrecipients,60%ofwhombelongedtothegroupdenotedas“registered-unemployed”at thePublicEmploymentService.The restareworkers,retiredrecipientsandnon-workingpersonswhofordifferentreasonsdonotfitintothecategoryofthoseregisteredasunemployed.MISrecipientsaccountfor25%ofalltheregisteredunemployedintheBasqueCountry.

GiventhattheBasqueMISisalastresortscheme,individualswithloweducationlevelsandthe(very)long-termunemployedareprevalentamongrecipients.Specifically,60%ofMISrecipientshaveatmostprimaryeducationand52%havebeenlookingforajobfor more than two years. Unsurprisingly, low education levels and particularly longunemployment durations are themain determinants that delay job-finding. Indeed,MISrecipientshaveanaveragemonthlyjob-findingrateof3%,whileforunemployedpeoplewhodonotreceivetheMIS,theratestandsat9%.

Thefirstempiricalstrategyinthispaperistomeasurewhetherthisdifferenceissolelyduetothedifferentcompositionoftheunemployedor,whethertheMISdelaysentryintoemploymentasempiricalevidencehasproventhatpassivepoliciesdoingeneral.

ThesecondaimofthepaperistomeasuretheeffectivenessofactivepoliciesonMISrecipientsintermsoftheirimpactontheprobabilityoffindingajob.EventhoughallMIS recipients are supposed to engage in activationmeasures, the fact is that onlyaround40%of them (16,000outof38,000unemployed recipients)havedone soatanytimeinthelastsixmonths.Guidanceisthemostcommonservice:itisreceivedby39% of all unemployed MIS recipients. It is followed at some distance by training(received by only 2.3%). The profiles of the participants differ from one kind ofactivationmeasuretoanotherandalsowithrespecttothosewhodonotparticipateinsuchservices.

PropensityScoremethodsareappliedinbothassessments.Inbothanalyseswefollowan Inverse Probability Weighting methodology. In the second exercise we alsosupplement our assessment with a Propensity ScoreMatching. Bothmethodologieshelp us deal with confounding effects and differences in composition between thetreatedandcontrolgroupsinthemostsuitablewayaccordingtothecharacteristicsofthecorrespondingsample.

OurresultsconfirmthatonaveragetheMISdoesnotdelayentryintoemployment,sothe difference in the job-finding rates observed are due solely to the differentcompositionsofthetreatedandcontrolgroups.Iftheanalysisisconductedforspecificpopulation groups, we find that its impact differs. The undesired delay effectcommonlyfoundinpassivepoliciesisobservedamonglesseducatedandyoungerMISrecipients, but the MIS accelerates entry into employment for medium and high-educatedpeopleandfortheover45s.Tothebestofourknowledgetherearenootherassessmentsofsimilarpolicyimplementationsthatwecouldcompareourresultswith.

The second finding is that all types of public employment activation services havepositive impacts on job-finding rates, but the extent of that impact varies fromonemeasure to another: the most effective services are training programmes (whichdouble the probability of finding a new job), followed by guidance services (whichincreasetheprobabilitybyaround20%).Hence,asapolicydevice,thisstudysupportsthe conclusion that training services forMIS recipients should be enforced, as theyhelp recipients to re-enter employment, which is the ultimate aim of activationmeasures. Moreover, it is essential to emphasise the importance of linking passivepolicieswithactivationmeasuresforrecipients.

Finally, for future researchwehave two relatedprojects towork on, both ofwhichrequiremoreinformation.Thefirstistoextendourstudytoadurationtypeanalysis,wherethequestiontobeansweredisnotbasedoninstantaneousjobfindingratesbutratherontimetoexitfromunemployment.Fornowwearelimitedbythefactthatweonlyhave informationonallunemployedworkersfor12months.ForMISrecipients,morethan70%ofwhomhavebeenunemployedformorethanayear,wewouldneedlonger longitudinal information. Secondly, we would like to obtain more preciseinformationonwhattrainingprogrammesMISrecipientsreceive,soastolearnmoreaboutwhat typesof trainingprogrammearemost successful in termsof job-findingrates.Thiswouldenableustobemoreprecisewithregardtopolicyadvice.

References

Alvarez,J.,&Arellano,M.(2003).TheTimeSeriesandCross-SectionAsymptoticsofDynamicPanelDataEstimators.Econometrica,71(4),1121-1159.Angrist,J.D.,&Pischke,J.S.(2008).Mostlyharmlesseconometrics:Anempiricist'scompanion.Princetonuniversitypress.Austin,P.C.(2011).Anintroductiontopropensityscoremethodsforreducingtheeffectsofconfoundinginobservationalstudies.Multivariatebehavioralresearch,46(3),399-424.Austin,P.C.,&Stuart,E.A.(2015).Movingtowardsbestpracticewhenusinginverseprobabilityoftreatmentweighting(IPTW)usingthepropensityscoretoestimatecausaltreatmenteffectsinobservationalstudies.Statisticsinmedicine,34(28),3661-3679.Ballester,R.(2012).LosprogramasdegarantíaderentasenEspaña:larentamínimadeinsercióncatalanaysuscomponentesdeinserciónlaboral(No.2012-333).DepartamentodeEconomía-PontificiaUniversidadCatólicadelPerú.Brunori, P., Chiuri, M. C., & Peragine, V. (2009).The Economic effects of a LocalMinimumIncomeSupportProgram(No.029).SERIESWorkingPaper.Caliendo,M.,Künn,S.,&Schmidl,R.(2011).Fightingyouthunemployment:Theeffectsofactivelabormarketpolicies.Card,D.,Kluve,J.,&Weber,A.(2010).Activelabourmarketpolicyevaluations:Ameta-analysis.TheEconomicJournal,120(548),F452-F477.Chemin,M.,&Wasmer,E.(2011).Ex-anteandex-postevaluationofthe1989Frenchwelfarereformusinganaturalexperiment:the1908sociallawsinAlsace-Moselle.underreviewattheJournalofPoliticalEconomy.Clavet,N.J.,Duclos,J.Y.,&Lacroix,G.(2013).FightingPoverty:AssessingtheEffectofGuaranteedMinimumIncomeProposalsinQuébec.(No.7283).IZADiscussionPaper.Colombino,U.,Locatelli,M.,Narazani,E.,&O’Donoghue,C.AlternativeBasicIncomeMechanisms:AnEvaluationExercise.Cordazzo,P.(2005).MinimumIncomeAllocationSystem(RMI)ALongitudinalView.Evaluationreview,29(5),407-424.delaRica,S.(2015).Políticasactivasdeempleo:unapanorámica.FedeaPolicyPapers-2015/01.

Ehlert,C.R.,Kluve,J.,&Schaffner,S.(2012).Temporaryworkasanactivelabormarketpolicy:evaluatinganinnovativeprogramfordisadvantagedyouths.Ferrera,M.(2005).WelfarestatesandsocialsafetynetsinSouthernEurope.WelfarestatereforminsouthernEurope.FightingpovertyandsocialexclusioninItaly,Spain,PortugalandGreece,1-32.Freire, G. B. O. (1997). La renta mínima garantizada: a caballo de la Asistencia, laSeguridadSocialylaPolíticadeEmpleo.Revistadetreball,economiaisocietat,(7),33-56.Gardeazabal,J.,&VegaBayo,A.(2016).Theeconomiccostofarmedconflict.Gouveia,M.,&Rodrigues,C.F.(1999).Theimpactofa"MinimumGuaranteedIncomeProgram"inPortugal.Heckman,J.J.,LaLonde,R.J.,&Smith,J.A.(1999).Theeconomicsandeconometricsofactivelabormarketprograms.Handbookoflaboreconomics,3,1865-2097.Imbens,G.W.(2004).Nonparametricestimationofaveragetreatmenteffectsunderexogeneity:Areview.TheReviewofEconomicsandStatistics,86,4–29.Imbens,G.W.,&Wooldridge,J.M.(2009).Recentdevelopmentsintheeconometricsofprogramevaluation.Journalofeconomicliterature,47(1),5-86.Kluve,J.,Card,D.,Fertig,M.,Gora,M.,Jacobi,L.,Jensen,P.,...&Schmidt,C.M.(2007).ActiveLaborMarketPoliciesinEurope.Kluve,J.(2010).TheeffectivenessofEuropeanactivelabormarketprograms.Laboureconomics,17(6),904-918.Marchal,S.,Marx, I.,&VanMechelen,N.(2016).Minimumincomeprotectionintheausteritytide.IZAJournalofEuropeanLabourStudies,5(1),1.Matsaganis,M.,Ferrera,M.,Capucha,L.,&Moreno,L.(2003).MendingNetsintheSouth:Anti-povertyPoliciesinGreece,Italy,PortugalandSpain.SocialPolicy&Administration,37(6),639-655.Ravagli,L.(2015).AminimumincomeinItaly(No.EM16/15).EUROMODattheInstituteforSocialandEconomicResearch.Rodrigues,C.F.(2001).Anti-povertyeffectivenessandefficiencyoftheGuaranteedMinimumIncomeProgrammeinPortugal.RosenbaumPR,RubinDB.(1983).Thecentralroleofthepropensityscoreinobservationalstudiesforcausaleffects.Biometrika;70:41–55.

RosenbaumPR,RubinDB.(1984).Reducingbiasinobservationalstudiesusingsubclassificationonthepropensityscore.JournaloftheAmericanStatisticalAssociation;79:516–524.Rosenbaum,PR,Rubin,DB(1985).Constructingacontrolgroupusingmultivariatematchedsamplingmethodsthatincorporatethepropensityscore.TheAmericanStatistician,39,33–38.RubinDB.(2001)Usingpropensityscorestohelpdesignobservationalstudies:applicationtothetobaccolitigation.HealthServices&OutcomesResearchMethodology;2:169–188.RubinDB.(2004)Onprinciplesformodelingpropensityscoresinmedicalresearch.PharmacoepidemiologyandDrugSafety;13(12):855–857.Whitton,T.(1993).Does'insertion'work?France'sminimumincome.Citizen'sIncome,(19),pp-11.Wooldridge,J.M.(2015).Introductoryeconometrics:Amodernapproach.NelsonEducation.Wooldridge,J.M.(2010).Econometricanalysisofcrosssectionandpaneldata.MITpress.

Figures

Figure1

Figure2

Figure3

Figure4

Tables

Table1.Probabilityoffindingajob.

Dependentvariable:exitprobability

UnemployedMISrecipients

UnemployedNon-MISrecipients

Women -0.0008 -0.0014**

(0.00067) (.0006315)

Foreignnationals 0.0002 -0.0106**

(0.00076) (0.00107)

Disabledpersons -0.0074*** -0.0173***

(0.00197) (0.00243)

MISrecipients 0.0013* -

(0.0007) -

Socialservicesderivation -0.0227*** -0.0554***

(0.00505) (0.01238)Benefits contributory 0.0104*** .0292***

(0.00151) (0.00089) attendance 0.0091*** 0.0245**

(0.00080) (0.00112)

ex-contributory - 0.0406***

- (0.00090)

ex-attendance - 0.0207*** - (0.00105)

Activationservices guidance 0.0056*** 0.0054***

(0.00056) (0.00082)

monitoring 0.0068*** 0.0048

(0.00176) (0.00504)

self-employmentinfo 0.0164*** 0.0237***

(0.00358) (0.00426)

training 0.0195*** 0.0403*** (0.00126) (0.000148)

Age 25-30 0.0018 0.0001

(0.00151) (0.00152)

30-35 0.0022 -0.0144***

(0.0015) (0.00154)

35-40 0.0012 -0.0206***

(0.00144) (0.00153)

40-45 0.0015 -0.0200***

(0.00146) (0.00154)

45-50 -0.0006 -0.0208***

(0.00148) (0.00156)

50-55 -0.0028* 0.0265*

(0.00153) (0.00158)

55-60 -0.0083*** -0.0473***

(0.00159) (0.00159)

60-65 -0.0178*** -0.0785***

(0.00166) (0.00159)

Education primary 0.0026*** 0.0025*

(0.00098) (0.00144)

uncompletedsecondary 0.0001 0.0041***

(0.00095) (0.00141)

secondary 0.0053*** 0.0151

(0.00102) (0.00140)

highschool 0.0080*** 0.0157***

(0.00135) (0.00155)

Medium-levelvocationaltraining 0.0111*** 0.0289***

(0.00148) (0.00158)

High-levelvocationaltraining 0.0175*** 0.0284***

(0.00177) (0.00158)

Undergraduate 0.0253*** 0.0301***

(0.00317) (0.00187)

Bachelor’sdegreeorhigher 0.0176*** 0.0300***

(0.00230) (0.00170)Unemployment

duration3-6months -0.0524*** -0.0796***

(0.00189) (0.00091)

6-12months -0.0662*** -0.1045***

(0.00172) (0.00087)

1-2years -0.0819*** -0.1297***

(0.00163) (0.00084)

2-3years -0.0857*** -0.1392***

(0.00164) (0.00091)

3-4years -0.0891*** -0.1480***

(0.00164) (0.00092)

4yearsormore -0.0943*** -0.1566***

(0.00160) (0.00081)

baselineprob. 0.0291 0.0617

averagepred.prob. 0.0304 0.0750

Observations 431,773 1,297,683Robuststandarderrorsinparentheses***p<0.01,**p<0.05,*p<0.1Additionalvariablesareincludedintheestimation:experienceinrequestedoccupations,activityinpreviousfieldofwork,languageskills,geographicalscopeofjobsearch,provinceofregistrationandmonthsinwhichtheindividualisobservedasunemployed.Baselineprofile:men,native,nodisabilities,notreferredtosocialservices,under25,illiterate,unemployedforlessthan3months.

Table2.Compositionofthetreated,non-weightedandweightedcontrolgroupsintheanalysisoftheimpactofMISontheprobabilityoffindingajob(%)

Treatment

Non-weightedControl

Weightedcontrol

Gender Men 49.6 42.19 48.3

Women 50.4 57.81 51.7Age

<30 16.27 20.13 14.130-44 45.73 39.32 50.5>44 37.99 40.55 35.4

Education Primary 59.82 32.7 61.3

Secondary 26.83 29.72 26.3Tertiary 13.35 37.58 12.4

Unemploymentduration <3months 12.29 33.73 11.5

3-6months 7.04 10.8 6.26-12months 11.03 11.98 11.3

1-2years 17.42 13.55 18.8>2years 52.21 29.94 52.1

Treatedgroup:UnemployedMISrecipients.Controlgroup:Unemployedpeoplewithoutbenefits.

Table3.Assessmentresults:impactofMISontheprobabilityoffindingajob.

IPW AIPW

ATT 0.000135 -0.000690

(0.000823) (0.000510)

Observations 724,141 724,141Robuststandarderrorsinparentheses***p<0.01,**p<0.05,*p<0.1IPW:InverseProbabilityWeighting.AIPW:AugmentedInverseProbabilityWeighting.Treatedgroup:UnemployedMISrecipients.Controlgroup:Unemployedpeoplewithoutbenefits.

Table4.Assessmentresults:impactofMISontheprobabilityoffindingajobpergroup. IPW AIPWGender MenATT 0.00308** 0.000745

(0.00120) (0.000835)

Obs. 324,751 324,751

WomenATT -0.00165** -0.00190***

(0.000681) (0.000506)

Obs. 399,393 399,390

Age <30ATT -0.0128*** -0.0108***

(0.00124) (0.00109)

Obs. 190,570 190,570

30-44ATT 0.00343*** 0.00127

(0.00122) (0.000883)

Obs. 272,115 272,115

>44ATT 0.00228** 0.00223***

(0.00102) (0.000598)

Obs. 261,456 261,456

Education PrimaryATT -0.00144* -0.00202***

(0.000804) (0.000592)

Obs. 371,111 371,111

SecondaryATT 0.00368*** 0.00276***

(0.00119) (0.000895)

Obs. 196,226 196,226

TertiaryATT 0.00641*** 0.00545***

(0.00180) (0.00129)

Obs. 156,807 156,807Robuststandarderrorsinparentheses***p<0.01,**p<0.05,*p<0.1IPW:InverseProbabilityWeighting.AIPW:AugmentedInverseProbabilityWeightingTreatedgroup:UnemployedMISrecipientsbelongingtothespecificgroup.Controlgroup:Unemployedpeoplewithoutbenefitsbelongingtothespecificgroup.

Table5.CompositionofMISrecipientspertypeofactivation(%) Noactivation Guidance TrainingGender

Men 48.0 51.7 64.5Women 52.0 48.3 35.5

Age <30 18.6 12.9 16.0

30-44 43.1 49.4 54.3>44 38.3 37.7 29.7

Education Primary 60.8 59.1 41.1

Secondary 26.6 27.0 36.4Tertiary 12.6 13.9 22.5

Unemploymentduration <3months 13.2 10.7 18.2

3-6months 7.9 5.8 4.56-12months 11.3 10.6 11.0

1-2years 16.7 18.5 19.5>2years 50.9 54.4 46.8

Table6.Assessmentresults:impactofactivationontheprobabilityoffindingajob. Guidance Training IPW AIPW PSM IPW AIPW PSM

ATT 0.00543*** 0.00475*** 0.00760*** 0.0297*** 0.0258*** 0.0298***

(0.000601) (0.000453) (0.000772) (0.00233) (0.00204) (0.00292)

Observations 431,773 431,773 420,482 431,773 431,773 292,816Robuststandarderrorsinparentheses***p<0.01,**p<0.05,*p<0.1IPW:InverseProbabilityWeighting.AIPW:AugmentedInverseProbabilityWeighting.PSM:PropensityScoreMatchingTreatedgroup:UnemployedMISrecipientswhohavereceivedactivationservicesinthelastsixmonths.Controlgroup:UnemployedMISrecipientswhohavenotreceivedanyactivationservicesinthelastsixmonths