还剩25页未读，继续阅读
本资源只提供10页预览，全部文档请下载后查看！喜欢就下载吧，查找使用更方便
文本内容:
Policy Research Working Paper10966Perceptions of Economic Mobilityand Support for Education ReformsAlexandru CojocaruMiduiel LokshinIvanTorreWORLD BANKGROUPEurope and Central Asia RegionPoverty and Equity GlobalPractice November2024EBRD
2023.The surveycovers thetransition economiesof Europe and Central Asia andseveralcomparator countries in WesternEurope,the Middle East,and NorthAfrica.2The2010roundincludes35countries,the2016round34countries,and the2023round39countries.The surveyis basedon a nationallyrepresentative sample of around1,000households percountry in the secondand fourthrounds and around1,500households percountry in the thirdround.We capturesupport for tax-financed investments in educationby examiningresponses to the questionwouldyou be willing to give up part ofyour income or pay more taxesif youwere surethat theextramoney wasspent oneducationThis dummyvariable is the maindependent variable of ouranalysis.Our measure of expectations of future mobility is based on the twosocial ladderquestions:“Pleaseimagine a10-stepladder whereon thebottom,the firststep,stand the poorest10percent peoplein[COUNTRY],and on the higheststep,the10th,stand therichest10percent peoplein[COUNTRY].On whichstep of the ladder isyour householdtoday andWhere on the ladderdo youbelieve yourhousehold willbefour years from nowWe takethe difference between theexpected future and current income ladderpositionsas ourmeasureof expectations of future mobility,which rangesfrom-9to+
9.We alsoconduct asensitivity analysis thatrelies onlyon thedirection of the expectationsby creatinga variablewith threecategoriesdownwardly mobile,static,and upwardly mobile based on the difference betweenthe ladderposition today and the ladderposition four yearsfromtoday.The LiTSdata alsoallow forassessment of the expectation of intergenerational mobility,which is capturedby responses to the following statement:Children who are born now will have abetter lifethan mygeneration//with respondentsanswering on a five-step Likertscale rangingfrom stronglydisagree tostronglyagree.We constructa dummyvariable that takes thevalue of1if theanswer iseither agreeorstrongly agreeand zerootherwise.An individualsrisk aversioniscapturedby assessingresponses to the followingquestion:Please rateyourwillingness totake risksin general,onascale from1to10,where1means thatyou are not willing to takerisksat alland10means thatyou arevery muchwillingtotake risks//To measurerisk aversion,wereverse thecategories sothat highervalues implyhigher risk aversion andlower valuesimply lowerriskaversion.To capturebeliefs about the fairness of socioeconomic mobility,we relyon responsesto thefollowingquestion:z7n youropinion,which of the followingfactors is the mostimportant tosucceed inlife2We usethe second,third,and fourthrounds of the LiTS,for whichcomparable informationon parentaleducation backgroundisavailable.The classificationof parentaleducation in the firstround of the LiTSis notcomparable to that of the followingrounds,which iswhy weexcluded itfrom ourstudy.in[COUNTRY]now”Answer optionsare effortand hard work,intelligence andskills,political connections,breaking thelaw,and others.We collapsethese optionsinto abinary variablethattakesa valueof1ifrespondents think that successstems frommeritocratic reasonseffort/hard workor intelligence/skills andzeroif itrelies onpolitical connections,breaking thelaw,or other.Table lpresents the summarystatistics.
5.Main resultsInthis section,we firstexamine the relationship between expectations of future mobility and willingness topay for education investmentsacross countries in thelatest2023UTS round.Respondents averageexpectations of theirown mobility and that of children born todayvary significantlyacross countries.Incountries wherethe populationis more optimistic about the future,the share of thepopulation willingtopay part of itsincome orpaymoretaxes to support investments in educationis alsolarger panela ofFigure
1.Positive expectations about theprospects ofchildrenborn today with respect tothe currentgenerationare also correlated withgreater willingness to pay for education investments panelb ofFigure
1.We thenestimate therelationship betweensupport for tax-financed education reforms andexpectationsof onesown future mobility conditionalon the respondents characteristics.Table2reports themarginaleffects fromthis baselineprobit estimationin Specification
1.The resultsindicate that the respondentseducational levelis a strong predictorof support for investmentsin education,consistent withearlier findingsBusemeyer and Garritzmann2017;Garritzmann,Busemeyer,and Neimanns
2018.The shareof thepopulation withprimary educationsupporting education reforms is23percentage pointslower thanthat of respondents withpost-graduate education;for people withupper-secondary education,support is19percentage pointslower.The associationsbetween willingness topay for education reformsand age,gender,religious affiliation,area ofresidence,household composition,and incomearenotstatistically significant.Turning toour primaryvariableof interest,more positiveexpectations of future mobilityare associatedwithincreased willingness to supporteducation reforms:Moving upone stepon the income ladderisassociated with a
1.4percentage pointincrease in the probability of supportinginvestments in education.3At firstglance,a greaterwillingness to bear higher taxes for investments in education may seem tocontradictthe canonicalPOUM hypothesis,which positsthat itwill bein theself-interest of upwardly3We alsoestimate specificationswith polynomialsof the expectations of mobility variablesquadratic andcubic.Theseestimations failedto producesignificant coefficientson nonlinearparameters.mobile individualsto supportlower taxrates.However,as wehave arguedin section3,if reachinghigheducational attainmentrequires private investments on the partof parents,then investments in improvingeducationwould actuallybe consistent with self-interest forthose whoexpect upward mobility,becausetheir children will be more likely to benefit from improvededucation.We furtherhypothesized that people whothink successreflects hard work andskills about70percent ofthe sample are more willingto support reforms thatreward meritocraticefforts thanpeople who believethat successreflects othernonmeritocratic factors.The perceptionof fairnesscould beinterpreted astheeffectiveness ofconverting skillsand education into future incomes.Beliefs about fairness are alsoassociated with expectations of future mobility.The meandifference betweenfutureandcurrent ladderpositionsin our sample istwice ashigh amongthose whoperceive thepath tosuccess to be determinedbymeritocratic factors.People with a loweraversion torisk4may also have a lower discountrate thanpeoplewitha higherriskaversion andmay,therefore,bewillingto investmore inreforms thatbring payoffsin the future.Risktolerance is alsocorrelatedwith expectations of mobility-the meandifferencebetweenthe futureandcurrent incomeladder positionsis threetimes higher in oursample forthose withrisk toleranceabove7compared tothose withrisk tolerance below
4.To testthese hypotheses,we re-estimate the baseline model,adding controlscapturing fairnessSpecification2inTable2,risk aversionSpecification3,and bothjointly Specification
4.The estimationsshowthatpeople who believethat success in theirsociety ismeritocratic determinedby hard work andintelligence/skills are9percentage pointsmore likely togiveuppartof theirincomeorpay additionaltaxesto support investments in education thanpeople whothinkthatsuccessinlife reflectspersonalconnections orother reasonsunrelated tohardworkor skills.Risk aversionis negativelycorrelated withsupport for education reforms.This resultis consistentwith thefinding thatrisk-averse individualsdiscountthe future more,making themless likely to makeinvestments whosepayoffs are uncertain.On average,aone-step increaseon therisk aversionscale is associated witha2percentage pointdecrease in thewillingness to pay foreducation reforms.The estimationsalso reveal that introducingfairness andriskaversioninto themodel reduces the size ofthe coefficienton the mobility expectationsfrom
1.4to
1.1percentage points,equivalent toone-fifth of4In oursample,the averagedegree ofrisk tolerancefrom alevel of1being completelyrisk averseto alevel of10beingcompletely risklover isabout
4.8,suggesting onaverage amild riskaversion.Of the individuals in thesample,
12.3percent have arisk toleranceabove7,and
36.1percent have a risktolerancebelow
4.thebaselinepartial correlation-the resultconsistentwitha mediatingeffect of these variables.Our focuson expectations of ownfuture mobilitystems from observed intergenerationalpersistence,where children of upwardly mobile parentsmay bemore likely to benefit from animproved educationsystem,increasing theprobabilityof their parentssupporting educational investments.We can also useexpectationsof intergenerational mobility directlyby relyingon the question Childrenwho areborn nowwillhaveabetter lifethan mygeneration,”While thequestion refersto children in general,it isplausibleto assumethat respondentsbeliefs aboutchildrens fortuneswill beinformed,to a larger extent,by theexperiencesof andexpectations5for theirown childrenand thechildrenof relatives orfriends whowilllikely sharecommon sociodemographic characteristics with therespondent.Table3reports theestimates withan alternative,intergenerational measureof expectedmobility.As in thecase ofown mobility,among respondents who agreethat childrenwhoarebornnowwillhavea betterlifethan mygeneration/support foreducationinvestmentsis
5.3percentage pointshigherinthe baselinespecificationSpecification
1.As before,we findevidence ofpartial mediationof therelationship betweenintergenerational mobility and support foreducation reformsthrough beliefs aboutfairnessand attitudestoward risk.Accounting forboth variablesreducesthecorrelation betweenintergenerational mobilityexpectations and support foreducation reforms by20percent Specification
3.How doesthe positiveassociation betweenfuture upward mobility andwillingness to pay forimprovededucation squarewith evidence from previousstudies thathave foundupward mobility expectations to beassociated witha reducedpreference for redistribution Alesina and La Ferrara2005;Alesina andFuchs-Schuendeln2005;Cojocaru2014Upwardly-mobile parentsare more likely to invest in education iftheyassociate this mobility witha higherlikelihood that their childrenwill benefitfrom education,and,inthe caseof preferencesfor redistribution,upwardly mobileindividuals willmore likelybe disadvantagedbyhigher taxes,thus makingit in their self-interest to oppose suchpolicies.But thereturns oneducation aresohigh comparedto anyother investments,that theydominate thedecrease inwillingness to pay highertaxes.65Children fromwealthy householdsmay optfor privateeducation,reducing their parents incentivesto investin public education.Private educationis commonin severalLatin Americancountries.It is not widespreadin Europe and Central Asia,which has alegacy ofuniversal provisionof education and well-established stateuniversities tendto bemore prestigiousthan therecentlyestablished privateinstitutions.If thetop of the socioeconomic ladder wereto optout ofpublic education,the coefficientin ourmodelwould beattenuated.6Deming2022demonstrates thateach yearof eductionyields about10percent risein annualearnings.
6.Sensitivity androbustness analysisWetest severalalternative specificationsof ourmain modelto assessthe sensitivityof the results todifferentdefinitions of our variablesofinterest.First,we assumethat thedifference inutility ofrespondentswho movefrom thefirst tothe secondrung of the ladderis equaltothatfor someonewhomoves from the seventhtotheeighth rung.To relaxthis cardinalityassumption,we collapsethe variablerepresentingtheexpectationsof future mobility into a variablewith categoriescorresponding topossibledirections ofthemobility expectations:downward,static,and upward.7Specification1inTable4shows thatcomparedwiththestatic reference category,upward mobility expectations are associated witha higherwillingness to supporteducation reformsand downward mobilityexpectations are associatedwithalowerwillingness todo so.The magnitudeoftheeffect islarger thanthatofthe originalvariable Specification4inTable2;the sizeand statistical significance ofother controlsare similarinthetwo estimations.We alsoanalyze theimpact ofrespondents longer-term economictrajectories on their support for thereforms.These variablesare constructedfrom theresponsestoquestions aboutthe respondentspositionon the economic ladderfour yearsago,now,and four years inthe future.The risingincome trajectorycorrespondsto asituation inwhich theexpected ladderposition four years inthe futureis higherthan theladderposition four years priorto today,and there isnodownwardmobility in eitherofthetwo four-yearwindows betweenfouryearsearlier andtoday andbetween todayand fouryears inthe future.In otherwords,this measuredocuments animprovement ina respondentseconomicladderposition overthe8-year period.Such alonger-term upwardmobility measureaddresses theconcern that4years istoo shorttorecord meaningfulchanges inmobility andguards againstsharp short-term discontinuities,and inparticularagainst the possibility thatreported expectedupwardmobilitycould besimply anexpectation ofrecoveryafter somenegative economicshock.Specification2of Table4shows thatrespondentswhoexpect their futureincome to improve in fouryearsrelative tofouryearsearlier aremore likelyto supporteducation reforms,and theconditional correlationbetweenthismobilityvariable and support forreforms iseven strongerthan in Specification
1.In Specification3,we conductsensitivityanalysisfor theintergenerational mobilitymodels reported inTable
3.We restrictoursampleto respondentsfrom households with childrenroughly30percent ofthefull LiTSsample,to increasethe salienceofrespondents/own childrenwhen answeringthequestionof7Individuals whoexpect to be onalowerstep ofthe ladderin fouryears are categorized as having downwardexpectations ofmobility.Individuals whoexpect tobe ona higherstep ofthe ladderfouryearsfrom noware categorizedashavingupwardexpectations of mobility.Individuals whoexpect tobe on the samestepoftheladderfouryearsfrom nowarecategorizedashaving staticexpectations.whether childrenborntodaywill farebetter thanthe currentgeneration.This restrictiondoes not appearto affectthe magnitudeor thestatisticalsignificanceofthechildrens mobilityvariable.Finally,we explorethe possibilitythat ourestimates sufferfrom endogeneitybias stemmingfromunobserved individualtraits such as pessimism,which maybe simultaneouslycorrelated withexpectationsof mobility and individualwillingness tosupportreforms.As notedin section3,we employ10-year cohort measures of mobility matchedtotherespondents country-age cohortas instrumentsforindividual expectationsof mobility.For eachcohort,we derivedshares of individuals withmore orfeweryears ofschooling thantheirparentsi.e.,the sharesof upwardand downwardmovers.We derivetheseshares based on boththe father/sandmothers educationfromTorre,Lokshin,and Foster
2024.To testthe robustnessof ourinstruments,we reportIV estimatesfrom twospecifications thatuse differentcombinationsof IVs.In Specification4of Table4,expectationsoffuture mobilityare instrumentedby theshareof upward movers withrespect tothe fatherand downwardmovers withrespect tothe mother;Specification5shows theshareof upwardmoverswithrespectto mothersand downwardmovers withrespectto fathers.8The testsfor weakinstrument andoveridentification testspass atconventional levels,but endogeneitytests onlymarginally rejectthe nullhypothesis ofexogeneity inSpecification5and failtoreject itinSpecification
4.9These findingssuggest that the casefor relyingon the IV resultsover baselineprobitestimates isnot strong.Nevertheless,the second-stage IVestimates ofthe coefficienton theexpectationsof mobilityvariable remainstatistically significantand increase.Given theabove concerns,however,we do not want to overinterpretthe largermagnitudes oftheIVcoefficients.
7.Changes in support foreducation reformsover timeWenext investigatewhether theregularities weuncover byanalyzing thesampleofLiTS2023existed intheearlier rounds ofthesurvey.10Expectations offuture mobilityor support for longtermreforms mayhavefallen inthe wake oftheCOVID-19pandemic,which hadprofound andlong-lasting impactsoneconomic activityand householdwelfare World Bank
2022.We usedata from the2010and2016roundsof the LiTS toexamine thestability ofthe linkagesbetween expectationsof mobility and support foreducation reformsover time.8We usemixtures offathers andmothers education as overidentification tests,because whenthe sharesofupwardmovers anddownwardmovers are basedonfathers,one ofthe instrumentsbecomes insignificantinthefirst-stage regression.9Table4reports theCragg and Donald1993tests ofweak instrumentsand theSargan1975overidentificationtests.The p-valuefor theWu-Hausman testHausman1978is
0.0417for Specification4and
0.0585for Specification5;thep-value for the Durbinscorechi-square seeNakamura1981is
0.0414for Specification4and
0.0582for Specification
5.10The ChowI960test onstructural breakindicates thatfor allpairwise comparisonsof threerounds ofLiTS,estimates basedonpooled samplesare rejectedin favorof separateestimates.The2010and2016survey rounds offer twocontrasting datapoints.Like the2023round,the2010roundwas conductedinthewakeofthe GlobalFinancial Crisisand theEuropean debtcrisis of2009/
10.Incontrast,the2016data werecollected duringthe periodof relativeeconomic stabilityinthe region.Table5presents estimationsfor threerounds ofthe LiTS2010,2016,and
2023.11The firstthree setsofresults demonstrate that,as in2023,respondents withmoreoptimisticexpectations abouttheir economicmobilitywere more likelytosupport tax-financed educationreforms in2010and
2016.The strengthof thisrelationshipdoubled from
0.007in2010to
0.014in2016,remaining virtuallyunchanged in
2023.Theeffect onsupport forreforms of expectationsofintergenerational mobilityincreased from
0.054in2010to
0.073in2016and decreasedto
0.041in
2023.In allthree surveyrounds,beliefs aboutthe fairness of theunderlyingeconomic environmentand attitudes towardriskremain significantcorrelates of support fortax-financed educationreforms.Two findingsemerge fromthese intertemporalcomparisons.First,the positiveassociation betweenintra-and intergenerationalexpectationsofupwardmobilitypersists forover a decade andduring periodsofeconomic turmoiland relativecalm.Second,therelationship betweenexpectationsof upwardmobility—particularly intergenerationalmobility—andsupportfor tax-financed educationreforms appearsstrongerduring periodsofrelativeeconomic stability.
8.ConclusionsTo moveto steepergrowth trajectories,middle-income countriesof Europeand Central Asia mustimplementsystemic structural reforms.The successfulimplementation ofthese oftencostly policiesrequiresbroad publicsupport.Earlier studies,focusing primarilyon preferencesfor redistribution,have found that supportfor reformstendstobehigher amongpeople whoexpect tobenefitfromsuch reformsand people whobelievetheoutcomes ofthese reformstobefair.Misperceptions aboutthe impactsof reformsadd frictionsbyreducing the publics willingness to adjustto newrealities andcan evenjeopardize achievingthe reformobjectives.Managing peoplesexpectations abouttheirfuturewelfare iscrucial toconsolidating politicalsupport.This paperconfirms the importance of self-interest andbeliefs aboutfairness forsupport oftax-financed11The resultsfor the2023round ofLiTS in Table5differ fromthe resultsshown inTable2,because herewe usethe monthlyconsumptionaggregate as a controlvariable whereas theresultsinTable2arebasedonaspecification withmonthly incomeascontrol variable.Information aboutmonthly incomewas collectedonly inthe2023round.educationreforms.People whoexpect themselvesor their children tobe upwardlymobile aremore likelytosupport tax-financed educationreforms thanpeoplewhoexpect theirpositiontodeteriorate.Weconjecture thatthis highersupport stemspartly fromtheeconomicself-interest ofupwardly mobileindividualstoinvestin improvededucation thatwill benefittheir children.This relationshipis partiallymediatedby normativebeliefs aboutthe fairnessof economic opportunities in society and individualswillingness totake onrisks.These resultsare robust to different formulations of mobilityexpectationsandpersist acrossthree surveyrounds spanningboth stable and post-crisis economic environments.Our resultspoint tothe criticalrole expectationsof socioeconomic mobility andbeliefs aboutfairness playindriving supportfor costlyreforms.In addition,our findingshighlight thechallenges facingthe countriesofEuropeand Central Asia,where beliefsthat educationaland professionaladvancement arenot basedonmeritocracy remainwidespread.Indeed,data fromthe2016round ofthe LiTSrevealthatmore thantwo-thirds ofadults intheregionreport thathaving informalconnections isimportant tosecure agoodgovernment job,and almost40percent reportthat connectionshelp obtainuniversity admission.12Estimates ofintergenerationalmobilityacross agecohorts pointto deterioratingprospects ofsocioeconomic mobilityin Europeand Central Asiafor peoplewho reachedadulthood afterthe fallof theBerlinWall Cojocaru
2021.Governments canincrease supportfor structural reforms bymanaging expectationsofsocioeconomic mobility and investing in institutionalreforms aimedat addressinginequality ofopportunity.Clearly definingthe futurecourse ofthe reforms,building anarrative aroundthem,andsignaling thegovernments determinationto followthrough mightreduce individuals/uncertaintiesand improve their perceptionsof theirprospects,easing theadoption ofthe newpolicy regime.Thepolitical legitimacyof reformscritically depends onthegovernments abilityto demonstratethat theywill be carriedout fairlyand inclusively,providing alevel playing field forall segmentsof thepopulation.12Questions aboutthe importanceof informalconnections werenot askedinthe2023LiTS round.AbstractThis paperinvestigates therelationshipbetweenthesupport tax-financed educationreforms.This correlationisexpectations of economicmobilityandsupportfortax-financedrobust todifferentformulationsof mobilityexpectations andeducationreforms usingdata fromthe Life in Transition Survey,persists overadecade,encompassing bothstable andpost-crisiswhich covers39countriesinEurope,CentralAsia,and theeconomicenvironments.The relationshipis partiallymediatedMiddle Eastand NorthAfrica.The analysisdemonstrates thatby beliefs aboutthe fairnessofeconomic opportunitiesinindividuals whoexpect themselvesor their children tobe societyand individuals,readiness toembrace risks.upwardlymobilearemorelikely toPolicy ResearchWorking Paper10966This paper isaproduct ofthe OfficeoftheChief Economist,Europeand CentralAsiaRegion and the Povertyand EquityGlobalPractice.It ispartofalargereffort by the World Bank toprovide openaccess toits researchand makea contributionto developmentpolicydiscussions aroundthe world.Policy ResearchWorking Papersai*ealsoposted onthe Webat http://.The authorsmay becontactedat,,and.ThePolicyResearchWorkingPaper Seriesdisseminatesthefindingsofworkinprogresstoencouragetheexchangeofidecisaboutdevelopmentissues.Anobjectiveoftheseries islothejmdin^soulquickly,evenifthepresenluliunsarelessthanfullypolished.Thepaperscarryhenamesoftheauthorsandshouldbecitedaccordingly.Thefindings,intetpretations,andconclusionsexpressedinthispaperareentirelythoseoftheauthors.TheydonotnecessarilyrepresenttheviewsoftheInternationalBcmkforReconstructionandDevelopment/WorldBankanditsaffiliatedorganizations,orthoseoftheExecutiveDirectorsoftheWorldBankorthegovernmentstheyrepresent.Produced by the Reseai-ch SupportTeamReferencesAgranov,M.,T.R.Palfrey.
2020.“The effectsof incomemobilityandtax persistenceon incomeredistributionand inequality/European EconomicReview,
123.Aiyar,S.,R.Duval,D.Puy,Y.Wu,and L.Zhang.
2013.Growth Slowdownsand theMiddle-Income Trap.IMF Working Paper13-
71.Washington,DC:International Monetary Fund.Alesina,A.,and E.La Ferrara.2OO
5.zzPreferences for Redistribution inthe Landof Opportunities.”Journalof PublicEconomics895-6:897-
931.Alesina,A.,and N.Fuchs-Schuendeln.
2005.zzGoodbye Lenin.Or NotThe Effectof CommunismonPeoples Preferences.NBER WorkingPaper11700,National Bureauof Economic Research,Cambridge,MA.Alesina,A.,and P.Giuliano.
2009.Preferences forRedistribution.NBER WorkingPaper14825,NationalBureau of EconomicResearch,Cambridge,MA.Alesina,A.,and E.Glaeser.
2004.Fighting PovertyintheUS and Europe:A Worldof Difference.Oxford:Oxford UniversityPress.Alesina,A.,S.Stancheva,and E.Teso.
2018.z1ntergenerational Mobilityand Preferences forRedistribution.,American EconomicReview1082:521-
54.Benabou,R.,and E.Ok.
2001.Social Mobilityand theDemand forRedistribution:The POUMHypothesis/Quarterly Journal of Economics116:447-
87.Boeri,T.,A.Boersch-Supan,andG.Tabellini.
2002.Pension Reformsand theOpinions ofEuropeanCitizens/American EconomicReview922:396-
401.Busemeyer,M.R.and J.L.Garritzmann.
2017.Public opinionon policyand budgetarytradeoffs inZEuropeanwelfare states:evidencefroma newcomparative survey,Journal ofEuropean PublicPolicy,24:6,871-
889.Cancho,C.,M.Davalos,andC.Sanchez.
2015.Why SoGloomy Perceptions ofEconomic Mobility inEuropeandCentralAsia.Policy ResearchWorkingPaper7519,World Bank,Washington,DC.Carter,E.C.,M.E.McCullough,J.Kim-Spoon,C.Corrales,and A.Blake.
2012.“Religious PeopleDiscountthe FutureLess/7Evolution and Human Behavior333:224-
31.Cava川巳C.
2023.Fair EnoughSupport forRedistribution inthe Age of Inequality,CambridgeUniversity Press.Checchi,D.,and A.Filippin.
2004.“An ExperimentalStudy ofthe POUMHypothesis//Research inEconomicInequality11:115-
36.Chow,G.I
960.“Tests ofEquality betweenSets ofCoefficients inTwo LinearRegressions/Econometrica283:591-
605.Cojocaru,A.
2014.Prospects ofUpward Mobilityand PreferencesforRedistribution:Evidence fromtheLife InTransition Survey/7European Journal of PoliticalEconomy34:300-
14.-------.2021/zlnequality andWell-Being in Transition:Linking Experienceand Perceptionto PolicyPreferences.In ThePalgrave Handbookof Comparative Economics,ed.E.Douarin andO.Havrylyshin.London:Palgrave Macmillan.-------.
2023.Z/Flaunt ThemIf YouveGot ThemInformal Connectionsand Beliefs about ProspectsofUpward Mobilityin TransitionEconomies.ComparativeEconomicStudies653:416-
41.Corneo,G.,andH.P.Gruner.
2002.Individual Preferencesfor PoliticalRedistribution//Journal ofPublicEconomics83:83-
107.Cragg,J.G.,and S.G.Donald.
1993.“Testing Identifiabilityand Specificationin InstrumentalVariableModels/Econometric Theory9:222-
40.Deming,D.2022,Four Factsabout HumanCapital//Journal ofEconomic Perspectives,363:75-
102.Denisova,I.,M.Eller,T.Frye,and E.Zhuravskaya.
2012.Everyone HatesPrivatization,But WhySurveyEvidence from28Post-Communist Countries/Journal ofComparative Economics401:44-
61.Doner,R.F.,and B.R.Schneider.
2016.“The Middle-Income Trap:More PoliticsThan Economics.WorldPolitics684:608-
44.Duch,R.M.
1993.Tolerating Economic Reform:Popular Support for TransitiontoaFree Marketin theFormerSoviet Union//American PoliticalScience Review873:590-
608.EBRD European Bank forReconstruction and Development.
2023.Life in Transition Survey.London.Flechtner,S.,and S.Panther.
2016.Economic Inequality,Political Power,and PoliticalDecision-Making:The CaseoftheMiddle-Income Trap/In VarietiesofEconomicInequality,ed.S.Fadda andP.Tridico.London:Routledge.Fisher,S.
1986.Issues inMedium-Term MacroeconomicAdjustment/7World BankResearch Observer12:163-
82.Garritzmann,J.L.,Busemeyer,M.R.,andE.Neimanns.
2018.Public demandfor socialinvestment:newsupporting coalitionsfor welfarestate reformin WesternEurope,Journal ofEuropean PublicPolicy,25:6,844-
861.Hanushek,E.A.,and L.Woessmann.
2008.“The Roleof CognitiveSkills inEconomic Development/7Journalof EconomicLiterature463:607-
68.Hausman,J.A.1978/zSpecification Testsin Econometrics.Econometrica466:1251-
71.Heinemann,F.,I.Bischoff,G.Hennighausen,and T.Hennighausen.
2009.Xhoosing fromthe ReformMenuCard:Individual Determinantsof LabourMarket PolicyPreferences/In LabourEconomics,Jahrbucherfur Nationaldkonomieund Statistik,ed.Bernd Fitzenberger,Werner Smolny,and PeterWinker,180-
97.Berlin:De GruyterOldenbourg,https://doi.org/
10.1515/9783110508284-
006.IMF InternationalMonetaryFund.
2019.G-20Report onStrong,Sustainable,Balanced,and InclusiveGrowth.Washington DC.Kaufman,R.R.,and L.Zuckermann.
1998.Attitudes towardEconomicReformin Mexico:The RoleofPolitical Orientations.American PoliticalScience Review922:359-
75.Kelley,S.M.C.,andC.G.E.K.Kelley.
2009.Subjective Social Mobility:Data from30Nations//In ChartingtheGlobe:The InternationalSocial SurveyProgramme1984-2004,ed.Max Hallier,Roger Jowell,and TomW.Smith,106-
24.London:Routledge.Luttmer,E.,and M.SinghaL
2011.Culture,Context,and theTaste forRedistnbution//American EconomicJournal:Economic Policy31:157-
79.Margalit,Y.
2019.Political Responsesto EconomicShocks/Annual Reviewof PoliticalScience22:277-
95.Meltzer,A.H.,andS.F.Richard.1981,/ZA RationalTheory ofthe Sizeof Government/,Journal ofPoliticalEconomy89:914-
27.Mengel,F.,andE.Weidenholzer.
2023.PreferencesforRedistribution.JournalofEconomic Surveys37:1660-
77.Nakamura,A.
1981.zzOn theRelationships amongSeveral SpecificationError TestsPresented byDurbin,Wu,and Hausman/7Econometrica.496:1583-
88.Narayan,Am bar,Roy Van der Weide,AlexandruCojocaru,Christoph Lakner,Silvia Redaelli,DanielGerszon Mahler,Rakesh GuptaN.Ramasubbaiah,and StefanThewissen.
2018.Fa/r ProgressEconomicMobility acrossGenerations aroundthe World.Equity andDevelopment.Washington,DC:World Bank.OECD Organisationfor EconomicCo-operation andDevelopment.
2023.PISA2022Results.Volume I:TheState ofLearning andEquity inEducation.Paris.Piketty,T.
1995.“SocialMobilityand RedistributivePolitics.Quarterly Journalof Economics1103:551-
84.Powdthavee,N.,and AJ.Oswald.
2014.“Does MoneyMake PeopleRight-Wing AndInegalitarian ALongitudinalStudy ofLottery Winners/7IZA DiscussionPaper7934,Institute ofLabor Economics,Bonn,Germany.Restuccia,D.,C.Urrutia.
2004.Intergenerational persistencein earnings:The roleof earlyand collegeeducation/,American EconomicReview,945:1354-
1378.Rovelli,R.,and A.Zaiceva.
2013.“Did Supportfor Economicand PoliticalReforms IncreaseDuring thePost-Communist Transition,and IfSo,Why”Economics ofTransition212:193-
240.Rueda,
0.,andD.Stegmueller.
2019.Who WantsWhat RedistributionPreferences inComparativePerspective,Cambridge UniversityPress.Sargan,J.D.
1975.“Testing forMisspecification afterEstimating UsingInstrumental Variables/7Contributions toEconometrics.New York:Cambridge UniversityPress.Torre,I.,M.Lokshin,and J.Foster.Forthcoming.zzDoes SocialMobility AffectEconomic OutcomesCross-Country AnalysisUsing DifferentMobility Measures.PolicyResearchWorkingPaper,WorldBank,Washington,DC.VanderWeide,R.,C.Lakner,D.Gerszon Mahler,A.Narayan,and R.Gupta.
2024.IntergenerationalMobility aroundthe World:A NewDatabase/JournalofDevelopment Economics166:103-
67.Wiese,R.S.Eriksen.2O
24.zzWillingness topay forimproved public education andpublic healthcarezsystems:the roleof incomemobility prospects.Fiscal Studies,45:55-
76.WorldBank.
2018.World DevelopmentReport2018:Learning toRealize EducationsPromise.Washington,DC.doi:
10.1596/978-l-4648-1096-l.-------.
2022.World DevelopmentReport2022:Finance foran EquitableRecovery.Washington,DC.doi:
10.1596/978-l-4648-1730-
4.Figure1Expectations ofintra-and intergenerationalmobilityandwillingness topay for investments ineducationNote:Authors estimates,basedon data fromthe2023LifeinTransitionSurvey.a.Intragenerational expectationsb.Intergenerationalexpectations0Expecledfuturemobity ShareagreethatchildrenwillhaveabetterMevis-a-viscurrentgeneration——Linearfit-------Lxiearfitopf42uoqeonp£sulauHSuwddnsoxel2OGAs.0-6£=-s£T able1Summary statisticsfor thefull sampI eMinimuMaximu Number ofVariable MeanStd.dev.observationsm mWillingness topay fortax-financed investments in
0.
4790.
50001.025,268education yes/no
0.
3971.627-
9.
09.025,268Expectations ofown mobilityincome ladderExpectationsofupwardintergenerationalmobility
0.
4900.
50001.025,268childrenwilllive betterthan parents[yes/no]
0.
7050.
45601.024,820Belief aboutfairness
0.
3060.
46101.025,256Upward trendin incometrajectory
0.
2360.
42501.025,268Downward expectationsof mobility
0.
3850.
48601.025,268Static expectationsof mobility
0.
3800.
48501.025,268Upward expectationsofmobility
4.
8002.
8641.
010.025,151Risk aversioncategoricalGender andageFemale
0.
5800.
49401.025,
26847.
72317.
11718.
095.025,268AgeofrespondentEducationNo degree/no education
0.
0290.
16701.025,268Primary education
0.
0480.
21301.025,268Lower-secondary education
0.
1440.
35101.025,268Upper-secondary education
0.
4380.
49601.025,268Post-secondary non-tertiary education
0.
0770.
26701.025,268Tertiary educationnotauniversity diploma
0.
0470.
21201.025,268Bachelors degreeor more
0.
1630.
37001.025,268Masters degreeor PhD
0.
0540.
22701.025,268Marital statusSinglenever married
0.
2230.
41701.025,200Married
0.
5330.
49901.025,200Widowed
0.
1310.
33701.025,200Divorced
0.
0970.
29601.025,200Separated
0.
0160.
12501.025,200Religion
0.
1050.
30701.024,100Atheist,agnostic,or none
0.
5250.
49901.024,100Christian
0.
0190.
13801.024,100Buddhist
0.
0010.
02701.024,100Jewish
0.
3400.
47401.024,100MuslimOther
0.
0110.
10301.024,
1000.
6110.
48801.025,268Respondent livesin urbanarea
2.
8151.
7021.
020.025,268Household size
0.
1350.
22400.87525,268Share ofchildren inhousehold under
150.
2270.299010,51420,944Monthly householdincome000s USDat PPP
20170.
0970.
14104.00222,676Monthly householdexpenditure000s USDat PPP2017Note:The householdexpenditure variableincludes thefollowing categories:food andconsumption staples,transportation,education,health,clothing,and durablegoods.Table2Willingness topay for investments ineducation asa function ofexpectationsabout economicmobilityand respondent characteristics.Probit marginaleffect ME.________________________________Specification1Specification2Specification3Specification4Variable ME Std.err.ME Std.err.ME Std.err.ME Std.err.
0.014**
0.
0130.
0110.011*
0.
0040.
0040.
0040.004Expected ownmobility*……*Beliefs aboutfairness
0.
0860.
0170.087”
0.017…Risk aversion-
0.01r*
0.
0030.003-
0.019***Gender andage
0.
0080.
0080.
0080.008Female-
0.007-
0.
0090.
0050.
0030.
2000.
1930.
1990.194Age/100-
0.210-
0.171-
0.182-
0.
1510.204-
0.
0640.
1970.
0100.
2050.198Age squared-
0.028-
0.021Education reference category:masters degreeor
0.053-
0.233*”
0.055PhD-
0.
21610.
0550.057No degree/no education-
0.230*“-
0.21r*Primary education-
0.231***
0.035-
0.231*M
0.033-
0.216i
0.035-
0.215…
0.
0330.033-
0.252*”
0.032-
0.244*”
0.
0330.033Lower-secondary-
0.256…-
0.241”.educationUpper-secondary-
0.187“*
0.025-
0.181***
0.025-
0.180***
0.026-
0.
173.
0.026educationBachelors degreeor
0.026-
0.
0870.025-
0.086…
0.
0260.026--
0.082”*』moreMarital status
0.091***referencecategory:single[nevermarried]
0.
0170.
0170.
0170.017Married
0.
0140.
0120.
0170.
0160.
0180.
0180.
0180.018Widowed-
0.008-
0.009-
0.003-
0.
0030.
0190.
0190.
0180.018Divorced-
0.007-
0.007-
0.007-
0.
0060.
0330.
0050.
0330.
0110.033Separated
0.
0090.
0080.033Religion referencecategory:atheist,agnostic,ornone
0.346**
0.343**
0.
0450.
0440.
0470.045Jewish
0.
33510.336***Christian-
0.
0220.033-
0.
0240.034-
0.
0190.033-
0.
0210.
0340.
0520.
0520.
0510.052Muslim-
0.018-
0.017-
0.019-
0.
0170.
0430.
0490.
0440.051Other-
0.013-
0.003-
0.017-
0.009Other variablesHousehold size
0.
0150.
0050.
0170.
0050.016”，
0.
0050.01r-
0.005……
0.
0300.
0310.
0220.
0310.
0260.
0310.
0190.031Share ofchildren-
0.
0250.020-
0.
0230.020-
0.
0250.020-
0.
0230.020Urban
0.
0630.
0470.
0590.
0450.
0530.
0460.
0500.044Monthly incomeNumberof observations21,16020,80021,04820,704Note:Marginal effectsof probitestimations arereported inthis table.Marginal effectson38county dummiesare omitted.Standard errorsare clusteredat thecountry level.Significant at the1percent level,*significant at the5percent level,*significant at the10percent level.*Age andage-squared arejointly significant at the1percent level.Table3Willingness topay forinvestmentsineducationasa function ofexpectationsabout intergenerationalmobilityand respondentcharacteristics.Probit marginaleffect ME.________________________________Specification1Specification2Specification3Specification4Variable ME Std.err.ME Std.err.ME Std.err.ME Std.err.Children willlive betteryes/no
0.
0530.
0150.
0460.
0140.049”*
0.
0140.042**
0.013』』*Beliefsaboutfairness
0.
0920.
0160.093**
0.015小*Risk aversion-
0.020”
0.003-
0.019”.
0.003Gender andageFemale
0.008-
0.
0080.
0090.
0070.
0080.009-
0.
0060.004Age/
1000.215-
0.086*
0.213-
0.101*
0.
2120.210-
0.110*-
0.081♦
0.
2170.
2140.
2150.212Age squared-
0.150-
0.170-
0.085-
0.104Education referencecategory:masters degreeor
0.047-
0.229”.
0.049PhD-
0.206”*
0.
0500.052No degree/no education-0,222”*-
0.215…Primary education-
0.222…
0.032-
0.223…
0.031-
0.202”.
0.032-
0.203”*
0.032Lower-secondaryeducation-
0.249…
0.032-
0.246…
0.032，
0.032-
0.232”*
0.032-
0.234”Upper-secondaryeducation-
0.182”*
0.025-
0.176***
0.025-
0.173”*
0.026-
0.167”.
0.026Bachelors degreeor
0.
0270.026-
0.077**
0.
0280.028-
0.083“*-
0.077*-
0.071more*Marital statusreferencecategory:single[nevermarried
0.
0170.
0170.
0180.018Married
0.
0100.
0080.
0130.
0130.
0170.
0180.
0170.018Widowed-
0.011-
0.011-
0.005-
0.
0050.017-
0.
0100.017-
0.
0090.
0170.017Divorced-
0.010-
0.
0080.
0300.
0310.
0300.031Separated
0.
0180.
0150.
0150.013Religion referencecategory:atheist,agnostic,Jewish
0.
0440.323**
0.042ornone
0.31T*
0.
0450.
0430.322**
0.315*“**
0.035-
0.
0360.036-
0.
0330.
0360.036Christian-
0.035-
0.
0330.056-
0.
0250.056-
0.
0280.
0550.055Muslim-
0.025-
0.
0270.089-
0.
0960.097-
0.
0900.
0810.088Other-
0.095-
0.091Other variables
0.
0160.017**
0.016**
0.
0050.
0050.
0050.005Householdsize
0.017-…**
0.
0220.
0290.
0170.
0300.
0210.
0300.
0170.030Share ofchildren
0.
0190.
0190.
0190.019Urban-
0.021-
0.019-
0.023-
0.
0200.
0470.
0450.
0460.045Monthly income
0.
0660.
0620.
0560.052Numberofobservations22,00021,56921,87621,461Note:Marginal effectsof probitestimations arereported inthis table.Marginal effectson38county dummiesare omitted.Standard errorsare clusteredat thecountry level.…Significant atthe1percent level,**significant atthe5percent level,*significant atthe10percent level.*age andage-squared arejointly significant atthe1percent level.Table4Willingnesstopayforinvestmentsineducationinfive modelspecifications^Probit marginaleffects ME.______________________Probitmarginaleffects2SLSsecondstageSpecification1Specification2Specification3Specification4Specification5Variable MEStd.err.MEStd.err.ME.Std.err.ME.Std.err.ME.Std.err.Expectedmobilityreferencecategory:staticDownward-
0.028”
0.
0140.042…
0.012UpwardUpwardtrendinownmobility
0.052…
0.
0110.040-
0.019Childrenwilllivebetterthanparentsyes/noExpectedownmobility
0.096”
0.
0430.103”Beliefinfairness
0.086…
0.
0170.088-
0.
0160.096,
00.
0180.070-
0.
0080.070-Riskaversion-
0.018…
0.003-
0.
019.
0.003-
0.015…
0.004-
0.014…
0.002-
0.013-GenderandageFemale
0.
0030.
0080.
0040.
0080.
0120.013-
0.
0030.007-
0.003Age/100-
0.
1440.193-
0.
0530.
1910.799,
0.
4410.413”
0.
1900.432”AgesquaredEducationlevel-
0.
0130.198-
0.
1110.195-
0.990,
0.493-
0.380”
0.151-
0.389”referencecategory:mastersdegreeorPhDNodegree/noeducation-
0.220”
0.057-
0.217…
0.054-
0.161**
0.075-
0.24#
0.031-
0.246^Primaryeducation-
0.215…
0.033-
0.
210.032-
0.186-
0.040-
0.214，”
0.024-
0.216*”Lower-secondaryeducation-
0.24T
0.033-
0.239…
0.032-
0.149”’
0.039-
0.226”
0.018-C.22”Upper-secondaryeducation-
0.172,
00.026-
0.171,
0.026—
0.107…
0.033-
0.168”
0.016-
0.161BachelorsdegreeormoreMaritalstatus-
0.0810,
0.026-
0.077”，
0.026-
0.
0390.034-
0.138F
0.020-
0.138***referencecategory:SinglenevermarriedMarried
0.
0170.
0170.
0150.017-
0.
0020.
0340.033…
0.
0120.034…Widowed-
0.
0030.018-
0.
0020.
0170.088”
0.
0420.
0090.
0150.009Divorced-
0.
0050.018-
0.
0020.018-
0.
0470.
0360.
0070.
0150.
0070.
0090.
0330.
0220.
0300.
0280.032-
0.
0330.
0420.027SeparatedReligion referencecategory:atheist,agnostic,ornoneBuddhistJewish-
0.
0190.019-
0.
0110.
0230.
0190.
0140.
0050.
0360.
0060.334-
0.
0460.339”’
0.
0410.366…
0.
1350.369-Christian-
0.
0200.034-
0.
0220.035-
0.059,
0.034-
0.028”
0.014-
0.029”Muslim-
0.
0180.051-
0.
0200.055-
0.
0770.071-
0.
05800.024-
0.060Other-
0.
0110.051-
0.
0850.080-
0.
1790.120-
0.
0020.035-
0.002OthervariablesHouseholdsize
0.017…
0.
0050.017”
0.
0050.019”
0.
0080.015^
0.
0030.015…Shareofchildreninhousehold
0.
0200.
0310.
0170.030-
0.119”
0.
0550.
0140.
0220.012Urban-
0.
0240.020-
0.
0210.020-
0.
0170.026-
0.026***
0.008-
0.026…MonthlyincomethousandUSDatPPP
20170.
0490.
0440.
0490.
0450.
0120.
0370.039…
0.
0130.039…SarqanscoreW
0.001p=
0.
97040.608p=
0.4355⑵CraqqandDonaldF19,438___________________________
28.635p=
0.
000021.287p=
0.0000Numberofobservations20,70422,0956^161949519,495tNote:Marginaleffectsinspecifications1,2,and3arecalculatedatthemeansofindependentvariablesandcontrols.Marginaleffectson30countydummiesinspecifications1,2,and3are omitted.Standard errorsare clusteredatthecountry level.***Significant atthe1percent level,**significant atthe5percent level,*significant atthe10percent level.♦Age andage-squared arejointly significantatthe1percent level.29055006700/4Vn/XQoQ38373o0oo2oo0Qoo55Table5Intertemporal comparisonsof willingnesstopayforinvestmentsineducationasafunction ofexpectationsaboutintra-andintergenerationa Imobilityandresp ondentcharacteristic.Pro bitmarg inaleffect ME._____________________________________________201020162023201020162023ME Std.err.ME Std.err.ME.Std.err.MEStd.err.ME.Std.err.MEStd.err.Variable
0.007-
0.
0030.014…
0.
0040.013-
0.004ExpectedmobilityChildrenwilllivebetteryes/no
0.054-
0.
0160.073…
0.
0140.041***
0.012Beliefinfairness
0.067…
0.
0190.108i
0.
0110.083…
0.
0150.062-
0.
0180.106…
0.
0100.088…
0.015-
0.014…
0.002-
0.019…
0.002-
0.018-
0.
0030.016-
0.002-
0.019-
0.002-
0.018…
0.003RiskaversionGenderandageFemale
0.
0130.014-
0.
0020.
0080.
0030.
0080.
0130.013-
0.
0050.
0080.
0050.009Age/
1000.227*
0.
1440.054*
0.162-
0.217*
0.196-
0.314-
0.
1390.085*
0.143-
0.245*
0.199Agesquared-
0.300-
0.145-
0.
1990.172-
0.
0360.192-
0.382…
0.140-
0.
2530.
1560.
0370.193Education referencecategory:mastersdegreeorPhDNodegree/noeducation-
0.247***
0.034-
0.235-
0.032-
0.234-
0.
0580.255-
0.034-
0.232,
00.028-
0.234…
0.052Primaryeducation-
0.169”
0.035-
0.205…
0.024-
0.239…
0.033-0,172…
0.034-
0.194***
0.024-
0.
23020.032Lower-secondaryeducation-
0.183…
0.030-
0.195-
0.025-
0.257…
0.032-
0.178…
0.026-
0.192…
0.023-
0.250-,
0.034Upper-secondaryeducation-
0.131-
0.027-
0.129-
0.020-
0.186-
0.026-
0.132-
0.024-
0.123…
0.019-
0.178***
0.027Tertiarynon-university-
0.106-
0.029-
0.117-
0.020-
0.179-
0.030-
0.111…
0.025-
0.112-
0.018-
0.165,
0.029Bachelorsdegreeormore-
0.065***
0.024-
0.112-
0.033-
0.064-
0.
0220.036Maritalstatus referencecategory:single[nevermarried]Married
0.020,
0.
0110.
03140.
0160.
0230.
0160.
0120.
0120.026,
0.
0150.
0230.016Widowed-
0.03T
0.017-
0.
0000.
0200.
0020.015-
0.035-
0.017-
0.
0030.
0190.
0010.015Divorced-
0.
0040.015-
0.
0290.018-
0.
0020.019-
0.
0110.018-
0.
0270.017-
0.
0030.018Separated-
0.047,
0.
0260.
0190.
0250.
0110.033-
0.
0280.
0260.
0150.
0240.
0160.028Religion referencecategory:atheist,agnostic,ornoneBuddhist-
0.
0120.
0420.118”，
0.015-
0.
0000.
0210.
0080.
0510.116”*
0.
0150.
0280.021Jewish
0.
1000.091-
0.
0410.
0730.294”’
0.
0470.
1150.095-
0.
0110.
0720.273…
0.040Christian-
0.
0080.027-
0.
0220.020-
0.
0300.031-
0.
0110.027-
0.
0210.020-
0.
0390.031Muslim-
0.
0500.
0380.
0130.046-
0.
0110.047-
0.
0480.
0360.
0180.049-
0.
0090.048Other
0.
0090.038-
0.
0120.
0280.
0090.
0480.
0230.036-
0.
0050.
0280.
0020.034OthervariablesHouseholdsize
0.021”’
0.
0050.
0030.
0050.015”’
0.
0050.02T
0.
0050.
0040.
0050.015”
0.005Shareofchildreninhousehold
0.
0250.
0290.
0310.
0280.
0130.
0300.
0430.
0300.
0320.
0270.
0190.028Urban
0.
0310.022-
0.
0160.013-
0.
0220.
0200.
0230.021-
0.023,
0.012-
0.
0210.019o.oooMonthlyexpenditure000sUSD
0.000…
0.
0000.5504-
0.
1470.
0840.
0820.
0000.55V
0.
1290.
0720.081Numberofobservations15,13533,45922,41515,73235,41323,387Note:Marginal effectsof probitestimations arereportedinthis table.Marginal effectsfor30county dummiesare omitted.Standard errorsare clusteredatthecountry level.Significant atthe1percent level,**significantatthe5percent level,*significantatthe10percent level.*Age andage-squared arejointly significantatthe1percent level.PerceptionsofEconomicMobilityand SupportforEducationReformsAlexandru Cojocaru,Michael Lokshin,and IvanTorreiJEL:D3J62,01,015,P2Keywords:Economic mobility,equality ofopportunity,supportforeducationreform,Europe andCentralAsia,Middle Eastand NorthAfrica1The authorsare atthe WorldBank.This papersfindings,interpretations,and conclusionsare entirelythose ofthe authorsand donotnecessarily representthe viewsofthe WorldBank,its ExecutiveDirectors,orthecountries theyrepresent.Miguel Purroyprovidedexcellent assistancein collecting,processing,and analyzingdata forthis paper.Send correspondenceto.
1.IntroductionStructural reformscan raiseincomes,but thecosts ofreforms mightfall disproportionatelyon somepopulationgroups,increasing inequalityand fuelingarguments opposingchange.The riskof lossesforlarge segmentsof societycould bea powerfulshaping forceofeconomicand socialpolicies.Identifying whowill likelylose outfrom reformsand managingpeoples perceptionsabout theirfuturewelfare iscrucial forconsolidating politicalsupportforcostly measuresand movingcountriesto highergrowth trajectories.It mightbe particularlyimportant tounderstand thefactors thatshapesupport forstructuralreformsfor themiddle-income countriesinEurope,CentralAsia,andelsewhere thatare facingchallenges inaccelerating economicgrowth andraising theirpopulationsstandards ofliving.This paperinvestigates whethersupportforeducationreformsis affectedby expectationsaboutfuture economicmobility.The focusonthesubset of structuralreformsrelated toeducation stemsfromthe evidencethat investmentsin human capital playa crucialrole indriving economicgrowthHanushek andWoessmann2008and thateducationreformsare apowerful policychannel throughwhichcountries canimprovetheinclusiveness ofgrowth byleveling theplayingfieldand enhancingtheupward economicmobility ofdisadvantaged groupsIMF
2019.The analysisis basedondataon39countriesinEurope,CentralAsia,and theMiddleEastand NorthAfricafromthe2010,2016,and2023roundsofthe LifeinTransitionSurvey EBRD
2023.Thissample ofcountries isanalytically interestingbecause,despite progressmade overthe lastseveraldecades,these countrieslag thehigh-income countriesin educationalachievement WorldBank
2018.Reforms aimedat improvingthe qualityof educationare alsomore costlyand politicallychallengingthan effortsto expandaccess toeducation,which helpedcountries movefromlow-income tomiddle-income statusAiyar andothers
2013.Profound educationreforms thattakeyears toachieve lastingeffects requireastrong,broad-based,durable coalitionatall levels of society.The paperfinds thatindividuals whoexpect themselvesor theirchildren tobe upwardlymobile aremorelikelytosupport tax-financed educationreforms than individuals whodonot.This relationshipisrobusttovarious formulationsofmobilityexpectationsandpersists acrossseveral surveyroundsspanning bothstableandpost-crisis economicenvironments.This relationshipis partiallymediatedby beliefsaboutthefairnessofeconomic opportunitiesinsocietyandindividualswillingnesstotakerisks.We usethe instrumentalvariable approachto mitigatethepossibilitythat ourestimates arebiasedbythepresence ofunobserved individualtraits correlatedwithexpectationsofmobilityandindividual willingnesstosupportreforms.The restofthepaperisstructured asfollows.Section2reviews the literature.Section3presents ourestimationframework.Section4describes thedata andpresents somedescriptive results.Section5presents themain results ofourmultivariate analysis.Section6presentstheresults ofour sensitivityanalysis.Section7examines changesinsupportforeducationreform overtime.Section8summarizes thepapers mainfindings.
2.Literature reviewEconomictheory suggeststhat self-interest drivespolicy preferencesand votingbehavior.The poorwouldbemorelikelytosupport higherlevelsofredistribution,for instance,whereastherich wouldbemorelikelytoopposethem Meltzerand Richard
1981.Having experiencedan adverseeconomicshock,such asjob lossor insecureemployment,heightens apersons senseof vulnerabilityand leadstostronger preferencesfor moreexpansive safetynets Margalit2019;Heinemann et al.2009;Boeri,Boersch-Supan,and Tabellini
2002.On theother hand,lottery windfallsare associatedwith supportformore conservativeand lessegalitarian politicalviews Powdhaveeand Oswald
2014.In transitioneconomies,self-reported upwardmobility wasreported tobe related to broaderacceptance oftheoutcomes ofprivatization Denisova etal.2012and greatersupportfor the marketeconomy Cancho,Davalos,and Sanchez
2015.Earlier studiesfoundthatother factors—including thedegree ofunderstanding ofwhat the reformsentail,political affinities,and normativebeliefs,such asattitudestowardfairness—also shapepreferences.Alesina andGlaeser2004argue thatthedifferenceinthesizeofthe welfarestate intheUnited Statesand Europepartly reflectsdifferences inbeliefsaboutthefairnessunderlyingpoverty and social mobility,as wellas ethnicfractionalization,which influencespreferences forredistribution.Luttmer andSinghal2011demonstratethatimmigrants politicalpreferences areinfluencedby beliefsabout redistributionintheircountry oforigin.Future expectationscanalsoinform economicself-interest.Benabou andOk2001show thatexpectationsof socioeconomicmobility canaffect preferencesfor redistribution.Individuals withbelow-average incomeswho wouldbenefitfromredistributive policiesfinanced byhigher taxesmayoppose suchpolicies ifthey expect tobeupwardlymobile.This prediction,known asthe ProspectsofUpward Mobilityhypothesis,has beenborne outbythedata inexperimental settingsChecchi andFilippin2004;Agranov andPalfrey2020as wellas inobservational dataAlesina andLaFerrara2005;Alesina andFuchs-Schuendeln2005;Cojocaru2014;Rueda andStegmueller
2019.The literatureon preferencesforredistributionpoints totheimportanceofself-interest,culturalfactors,andsocialmobility,both pastand futureCorneo andGruner2002;AlesinaandGiuliano2009;Alesina,Stancheva,and Teso2018;Cavaille2023;see alsoMengeland Weidenholzer2022for adetailedsurvey ofthis literature.Less isknown aboutthe rolethese factorsplay inshapingpreferences forpolicy choicesunrelated toredistribution.This paperaims tocontribute toclosingthis knowledgegap.Several earlier studies considerthe driversofsupportfor marketreforms atthebeginning ofthe post-Soviet transitionDuch1993;Denisova etal.2021;Rovelli andZaiceva2013orin LatinAmerica Kaufmanand Zuckermann
1998.Denisovaetal.2012find thateconomichardships duringthe post-Soviet transition—such asthe experienceof downwardmobility,the lackofhumancapitalsuitable fora marketeconomy,andthelack ofprivate assets—significantlyincreased thedesire forrevising theresultsofearlier privatization.Several earlierstudies haveexamined theindividual determinantsof preferencesforinvestmentsineducation fora smallsubset ofEuropean countriesBusemeyer andGarritzmann2017;Garritzmann,Busemeyer,and Neimanns
2018.Wiese andEriksen2024investigate linkagesbetweenexpectations ofmobilityandwillingnesstopayforimprovements inpubliceducationand health careusing data fromthe2016round oftheLiTS,framing itasatest ofthe POUMhypothesis withrespectto indirectrather thandirect redistribution,which wasthe focusonthemajority ofearlierstudies.However,willingnesstopay additional taxes forinvestmentsineducation orhealthcaremight notnecessarilybe drivenby redistributionmotives,as respondentscan viewsuch investmentsasbenefiting themselvesand theirfamilies;thus supportfor suchinvestments canbe entirelyconsistentwith self-interest.The latteris indeedthe argumentthis studyproposes toinvestigateempirically.
3.Theoretical frameworkand empiricalstrategyThe effectsofstructuralreforms taketime—possibly aslong asa generation—to materialize.Unlikelow-income countries,where educationpolicies focusmainly onexpanding accessto primaryand/orsecondary education,middle-income countries,which oftenalready haveuniversal accessto basiceducation,require harder-to-implement,longer-term investments,suchasupgrading thevocationaland tertiaryeducation thatis criticalfor researchand developmentDoner andSchneider
2016.In decidingwhether tosupport tax-financed investmentsineducation,individuals comparethediscounted valueoffuture benefits fromthe reformsfor themselvesor membersof theirfamiliesto thelosses incurrent consumptionbecause ofhigher taxes.Perceptions aboutsocioeconomicmobility arean importantfactor inthis decision.Intergenerational mobilityineducationand incomeischaracterized bya significantdegree ofpersistence.Children ofbetter-educated parentsand/orparents with higher socioeconomicstatus aremorelikelythan otherchildren tohave highlevels ofeducationand incomeas adultsvan derWeide etal.
2024.One ofthe keyfactors drivingthisintergenerational persistence,in botheducationandincomes,is parentalprivate investmentsin theeducationoftheirchildren Restucciaand Urrutia
2004.Individuals whoexpecttoimprove theirsocioeconomicstatus ormaintain highstatus wouldwanttotransfer thatstatus totheirchildren,partly byinvestingineducational attainment for them.Assume theeducational outcomesareafunctionoftwo complimentaryinputs-publicly-providedschooling andprivate parentalinvestment.Then,individuals whoanticipate theireconomic situationtoimproveinthe futurewould expecthigher returnsfrom betterpubliceducationbecause theycancomplement itwithhigherprivateinvestment.In contrast,peoplewhobelieve theirsituation willdeterioratemayseeless valueforgoing current consumption tofinance improvementsineducation,because theymay notbe ableto affordthe privateeducationalinvestmentsnecessary tosecurehigher educationalattainmentfor theirchildren.More risk-averse individualswill discount the futuremore heavily,reducing theirsupportforreformswhich outcomesareuncertain.Belief inmeritocracy,ontheother hand,willbeexpected toincreasewillingnesstofinance investmentsin policiesthat promotemeritocratic advancement.These considerationsmotivate our empirical model.The corehypothesis wetest isthat supportforeducation reformsfinanced byhighertaxesdependsonindividual expectationsoffuturemobility.This relationshipcould beexpressed asthefollowingmodel:；probWj*0=FpM+x Y4-v,1icccwhere wisthelatent willingnesstopayadditionaltaxesforinvestmentsineducationby individual/icfrom countryc withthe observedwillingnesstosupport tax-financed reformsw=1if w10and0icotherwise;F.isastandard normalcumulative distributionfunctionof—8;icM isa variablereflecting theexpectationoffuturemobility;X isa vectorofindividual-level controls,icincluding age,gender,education,area ofresidence,currentincome,household size,and presence ofchildren inthe household;v arecountry fixed effects;and记N0,1istheidiosyncratic errorterm.c£〜Given thetradeoff betweencurrentconsumptionandthediscounted valueoffuturebenefits fromimprovededucation,the controlvariables inourempiricalspecification areincluded tothe extentthatthey affecttheindividualscost-benefit decisions.We expectindividuals withhigher educationorincometosupportinvestmentsin humancapital.Wealthier householdswould alsohaveagreatercapacity tosupport tax-financed educationreforms.Younger peoplemay havea differentdiscountrate forfuturebenefitsthan olderones.We controlforthepresenceofchildrenina householdto capturedifferences inbenefits fromimprovededucation betweenhouseholdswithand withoutchildren.Area ofresidence isincluded toaccountforthedifferences ineconomicopportunities—and thusdifferent mobilityprospects—between urbanand ruralresidents.Socioeconomic mobilityprofiles alsovary bygenderNarayan etal.2018,which may affect expectationsofmobilityif suchexpectationsareformed bylearningfrom experiencePiketty
1995.Religious peoplewere shownto discountthe futureless andmay alsohavedifferent preferencesrelatedtoinvestmentsinthepubliceducation systemCarter etal.
2012.More risk-averse individualsdiscountthefuturemorebecause oftheir aversiontouncertainty.To theextent thatmoving uptheincomeladder canbe seenas indicativeof successinlife,beliefsaboutwhat underpinssuch success—hardwork,skills,and politicalconnections—isimportant Alesina,Stancheva,and Teso
2018.Supportforpolicies promotingmeritocraticadvancement advancementsachieved throughhardworkand intelligence/skills dependsonwhether humancapital investmentsare seentobeinstrumental forsocioeconomicmobility.Themodel alsoincludes countryfixedeffectsto accountfor heterogeneityacross countriesthat maybecorrelated withindividual beliefsabout futuremobilityandwillingnesstosupport tax-financedreforms.Unobserved factorsmayalsoaffect ourresults.For instance,anindividualwithapessimisticdisposition mayhave morenegative expectationsoffuturesocioeconomicmobilityand holdpessimisticviews aboutthereforms,and thusbe lesslikelytosupport suchreforms,biasing thecoefficientsonthemobilityexpectationsvariable.To addresssuch endogeneity,we instrumentindividualexpectationsoffuturemobilitywith cohort-level estimatesof pastintergenerationalmobility forthe10-year country-birth cohortcorresponding toa givensurvey respondent,obtainedfrom theestimates byTorre,Lokshin,and Fosterforthcoming.Our choiceof pastintergenerationalmobilityas aninstrument forfuturemobilityexpectations drawsontheliteratureshowing thatpast mobilityexperiences informpeoples perceptionsandexpectations ofthefuturePiketty1995;Kelley andKelley2009;Cojocaru
2023.Our exclusionrestrictionisbasedontheargument thatthereisno theoreticalreason whymeasures of pastintergenerational mobilityatthebirth-cohort levelshould affectindividual supportfor educationreformsdirectly beyondtheir effecton mobilityexpectations.We shouldnote,however,that wecannotexclude thepossibility ofcohort shocks,namely thatcohortmeasuresofpastmobility maystillbe correlatedwith otherfactors thatmayaffectsupportforeducation spending.As such,the IVestimatesshould similarlybe viewedwith thesecaveats inmind.
4.DataWe usedatafromthe2010,2016,and2023roundsofthe LifeinTransitionSurvey LiTSconductedby theEuropeanBankof ReconstructionandDevelopmentEBRD andtheWorldBank。
个人认证
优秀文档
获得点赞 0
