ANNALSOF ECONOMICS AND STATISTICS NUMBER97/98,JANUARY/JUNE 2010 AssetHoldings, Savings, andTemporary Migration ChristianDUSTMANNand JosepMESTRES (CReAM) University CollegeLondonand Centre forResearchandAnalysisofMigration inrelation totheirreturn ofimmigrants Thispaperanalyzessavingsandassetholdings plans.We Further, byreturn plansofimmigrants. maybe affected arguethatsavingsandassetaccumulation return thewaysavingsandassetsareheldinthehomeandhostcountry mayalso be relatedtofuture andnativesmayleadtoserious savingsandassetsbetweenimmigrants plans.Thus,comparing with We showthatimmigrants whenneglecting thehomecountry underestimation component. In in the home of their a return addition, country. savings proportion temporary plansplace higher inthehomecountry are andtheshareofassetsandhousingvalueaccumulated boththemagnitude andlowerthevalueofassetsand theirmigration as temporary, whoconsider largerforimmigrants we findno on observable heldinthehostcountry. characteristics, Finally,andconditional property with more than withtemporary save evidencethatimmigrants immigrants plans migration plans.* permanent migration I. Introduction is important forassessmentof thewelfareimpliof immigrants The economicperformance the a largenumberof papersinvestigate Not surprisingly cationsof immigration. therefore, and for different over the of countries, cycle, migration (relative)earnings position immigrants butis data1.An area thathas receivedless attention, and time-series usingbothcross-section in thereceiving forassessingtheeconomicpositionof immigrants perhapsequallyimportant assetsandsavingsarelikelyto andsavings.Like earnings, is theirassetaccumulation country, inthepastmayhaveaffected Returnintentions return. be affected past byplansabouta future ofcurrent assetholdingsand the and therefore as wellas pastexpenditures, magnitude earnings where return intentions Furthermore, maydetermine pastandcurrent savings.2 pastandcurrent assetsand savingsare held.For instance,whenconsidering housingand otherinvestments, rather as temporary areintended inthecountry oforiginifmigrations thesemaybe undertaken thanpermanent.3 1. From the early work of CHISWICK [1978] and BORJAS [1985] to DUSTMANN [1993], FRIEDBERG [1993], BORJAS [1995], BARTH et al. [2004], BRATSBERG and al [2006] and LUBOTSKY [2007] among others. 2. See DUSTMANN [1995, 1997] for a theoreticalanalysis of the interactionbetween returnmigrationand immigrant savings. 3. See WOODRUFF and ZENTENO [2007] and YANG [2008] forevidence on the creationof enterprisesof immigrants in theirhome countrieswhile abroad. DUSTMANN and KIRCHKAMP [2002] and MESNARD [2004] provide evidence of activitiesafterreturn. immigrantsundertakingentrepreneurial * JEL:F22, D31, R21 /KEY WORDS: International Migration,Wealth Accumulation, Housing Demand. 289 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions CHRISTIAN DUSTMANN AND JOSEP MESTRES A number ofpapersanalysethewealthgapbetweenimmigrants andnatives.Mostofthese a even conditional on observable characteristics4. However,mostmeasures report persistent gap, ofimmigrant wealthdo notconsider in wealth the home andhostcountry immigrants' separately. wealthaccumulation without betweenlocations Hence,an analysisthatconsiders differentiating mostofthese mayonlygivean incomplete pictureof assetholdingsand savings.In addition, studiesdo notallow eitherforheterogeneity acrossimmigrants due to differences in pastand current return plans5. In thispaper,we providean analysisofimmigrant to savingsandassetholdingsin relation and current return We also consider the that and assets are held past plans. possibility savings notonlyinthehostcountry, butalso inthecountry oforigin.Ouranalysisis basedon a unique datasourcethatprovidesinformation on assetholdings, itscomposition andlocation,as wellas return assetaccumulation and savings,andhowit immigrants' plans.We describeimmigrants' relatestoreturn as wellas individual andhouseholdcharacteristics. Thepapermakes intentions, twocontributions. it betweenreturn First, providesanalysisoftherelationship plans,on theone ontheother. theimportance ofconsiderhand,andsavingsandassetholdings Second,itillustrates assetholdings notonlyinthehostbutalso thehomecountry. ingmigrants' Ourresults showthattheoveralllevelofsavingsandassetaccumulation ofimmigrants would be severely ifhomecountry underestimated wealthis nottakenintoconsideration. In addition, return The totalvalueof theyshowhowimmigrants' plansarerelatedto wealthaccumulation. assetshelddoesnotdiffer between with households intentions significantly immigrant temporary andthosewithpermanent ones.However, thedistribution oftheseassetsbetweenhost-andhome locationdoesdiffer. whoplantoreturn do allocatea higher oftheir country Migrants proportion intheirhomecountry. savings,assetsandproperty The structure ofthepaperis as follows:inthenextsectionwe discussconceptual considerin III Section we ourdataandexplainthedescriptive ations, present evidence,in SectionIV we showourresults andfinally in SectionV we concludeanddiscusspotential implications. II. ConceptualConsiderations and Estimation II.7 A SimpleModel We start witha simplemodelthatfocuseson thewaytemporary vs permanent relate migrations to savingsbehavior. A moredetailedanalysisoftheinterplay betweensavingsandreturn (both exogenousandoptimally chosen)canbe foundinDUSTMANN [1995]. As inGALORandSTARK[1990],supposethatthelifetime oftheimmigrant canbe divided into2 sub-periods: 1 is the time to be in the host and period spent country, period2 is thetime tobe spentinthehomecountry aftera possiblereturn. inperiod2 takesplacewithprobReturn 4. See for example BLAU and GRAHAM [1990], COULSON [1999], BORJAS [2002], AMUEDO-DORANTES and POZO [2002], PAINTER et al. [2003], OSILI and PAULSON [2004], COBB-CLARK and HILDEBRAND [2006], SINNING [2007, 2009] or BAUER et al. [2009]. 5. An exceptionis BAUER and SINNING[2009], who foundthatthesavings behaviourof migrantsis relatedto theirreturn measures of migrantsavings,assuming eitherthatno remittancesare saved or that plans. The analysis considersdifferent all remittancesare saved, withoutdifferentiating between remittancepurposes. It does not investigateimmigranthomeownershipor asset holdings. 290 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS, ASSET HOLDINGS, AND TEMPORARY MIGRATION = 0, themigration In thecase that/? is permanent. Considerthefollowing inter-temporal ability/?. function: utility U = u'cl) + p$ü2(cE2)+ (l-p)u2(c12) 1 (1) 2 9 in period1 in thehostcountry and и and и arethe In equation(1), и is thesub-utility whichwe assume inperiod2 inthehomeandinthehostcountry samesub-utilities respectively, first and second J2 and (pare in concave consumption. as beingstrictly c1, Further, periodconTheparameter inimmigration respectively. (indexE) countries (indexI) andemigration sumption in the home to the a return attaches to the g is the country possible probability migrant p [0, 1] secondperiod. thanconinthesecondperiodinthecase ofa return mayinducemoreutility Consumption etc.Thisis capduetocomplementarities inthehostcountry, climate, friends, through sumption > and a of level has a If the turedbytheparameter highermarginal utility higher ß. ß 1, migrant ifhe/she consumesinthehomecountry. utility for forthefirst Thebudgetconstraint periodis givenbyw1= c1 + s. The budgetconstraint = andw^2+ rs= cE2inthe thesecondperiodis w72+s J2 inthecase ofa permanent migration inperiod1 aredenotedbyw1,andinperiod2 byw^2andw72inhome case ofa return. Earnings in the The purchasing and hostcountriesrespectively. powerof thehostcountrycurrency is higher is givenbyr. Ifr > 1,thepurchasing homecountry currency powerofthehostcountry homecountry6. inthemigrant's itfixesconsumption Thechoicevariableinperiod1 is savingss. Giventhebudgetconstraint, is givenby: ordercondition inthefirst period(c1) andinthesecondperiod(c^2,c72).The first u'=p$üir+ ('-p)u¡ (2) derivative. wherethesubscript j denotesthefirst levelofsavings.Savingswillbe setsuchthatthemarginal the determines optimal Equation(2) inperiod2. return in offorgone costinterms utility period1 is equalizedtotheexpectedmarginal = ofconsumption is permanent), utility savingswillequalizethemarginal If/? 0 (themigration in savings a to leads a in thetwoperiodsin thehostcountry. change in/? 1), change If/?e(0, the to relationship: following according ds_ -№u?-uh dp rìi+pfì^SiiHl-pWl (3) is alwaysnegative.Assumefirstthatß = 1 and r = 1: in thedenominator The expression as is thepurchasing arethesameinbothcountries, forconsumption powerofthehost preferences < In thiscase,savingswillbe increasing in/?as longas w^ v/,duetothestrict currency. country themarginal is thatan increaseinsavingsincreases Theintuition function. oftheutility concavity between ratesweredifferent ifinterest ratesareequalinthetwocountries; we haveassumedthatinterest 6. Forsimplicity assetaccumulation. sourceofdifferential thenthiswouldbe an additional homeandhostcountry 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions 291 CHRISTIAN DUSTMANN AND JOSEP MESTRES ofconsumption duetolowerwages.Thus,an increaseinthe Utility bymoreinthehomecountry, return leads to Now toconsume probability p higher savings. supposethatß > 1: individuals prefer at homerather thanabroad.Thiswillreinforce theeffect ofan increaseinthereturn probability on savings.Finally,supposethatthepurchasing is higherat powerofthehostcountry currency homeso thatr > 1. In thiscase,theoveralleffect on savingsis ambiguous, as itis nowunclear whether an increaseinsavingsincreases themarginal of moreinthehomeor utility consumption thehostcountry. A wagedifferential between homeandhostcountry thanthegainobtained larger > on savingsthrough thepurchasing differential m/-w^ is sufficient fortheeffect power (r l)s ofan increasein/?on savingstobe positive.Thus,according to thissimplemodel,savingsmay be positively or(ifpurchasing arelarge)negatively affected powerdifferentials byan increasein thereturn orthetwoeffects probability, maycanceleachotherout. Ourmodelhas nothing to say howeveraboutwheresavingsareheld.It maywellbe that who a to a return aremorelikelyto transfer someof their immigrants assign highprobability Ifthisis thecase,an empirical savingstothehomecountry. analysisofimmigrants' savingsmay leadto an underestimate whenonlyconsidering in the host savings economy. Now considerassetholdings, likehousingassetsorlong-term Iftheseareproinvestments. to the levelofpastsavings,thenoursimplelife-cycle modelshouldsuggestthat,in portional therelationship to thetemporariness ofa migration is ambiguous. The modeldoes not general, where these assets are held. it is not that who assigna explain However, unlikely immigrants to a later return have a foraccumulation inthehomecountry. Thisis highprobability preference thecase forassetsthathavethecharacter ofdurableconsumption particularly goods,andthatcan notbe movedfromoneplaceto another, likehousingassets. II.2 EmpiricalImplementation In theempiricalanalysiswe regressthevariousoutcomevariableson a vectorof individual oforigindummies, anda measurefortheprobability to return. specificcharacteristics, country Thegenericregression hastheform Yi=ai+Xia2+iri+ui (4) whereYtis therespective outcome,Xtis a vectorof background characteristics, щ is an error and7} is a measureforthetemporariness ofa migration. As we explainbelow,inourdata term, we observeforeachyearan indicator ornottheindividual wouldliketoreturn questionwhether homeat somepointinthefuture. Theseintentions affect maychangeovertime,andaccordingly thesavingsandassetholdingdecision.In ouranalysis, we willusetheaverageintention toreturn, from information overthelastfiveyears,as a measureoftemporariness whenanalyzing computed assetholdings, andthecurrent intention toreturn whenanalyzing current savings. We wouldliketo emphasizethatwe do notinterpret ourestimates as causal.Whilein our is exogenously simplemodel,thereturn probability given,immigrants maywellchoosewhether wish to and this choice not be to or decireturn, they may exogenous savings assetaccumulation sions.Further, ourmeasureforthetemporariness ofa migration witherror, maywellbe measured whichwouldbiasthecoefficient estimate towards zero.We believehoweverthattheassociations betweenthetemporariness ofa migration on theone hand,and savingsandassetaccumulation 292 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS, ASSET HOLDINGS, AND TEMPORARY MIGRATION onbackground as wellas thechoiceofwherethesearetobe held,whichareconditional behavior, in The overall are and characteristics, interesting important7. comparison savingsand assetacof andnativesis ofcoursenotaffected betweenimmigrants cumulation bypossibleendogeneity return migrations. III. Backgroundand Data III.7 Background a strong The WestGermaneconomyexperienced by a upwardswingafter1955,accompanied ratefellfrom rate.Between1955and 1960,theunemployment sharpfallin theunemployment of FÜRARBEIT[2009]). At thesametime,thepercentage 5.6% to 1.3% (BUNDESAGENTUR in West and countries from Southern workers born Germany Turkeyemployed European foreign increasedfrom0.6 percentin 1957to 5.3 percentin 1965,to 11.2 percentin 1973 (see BLITZ were Suchagreements recruitment was regulated agreements. bybilateral [1977]).Immigration setup withItaly,Spain,Greece,Turkey,Portugaland Yugoslaviain the 1950's and 1960's. thesecountries from labourstopped. offoreign After1973,recruitment Nevertheless, immigration The reunification dueto family [1996]formoredetails). immigrant continued, (see DUSTMANN overthis Labormigration movement. from that stems in this we migration paper study population the and countries the both as considered was emigration immigration by temporary period initially has beenquiteconsiderable return countries. (see BOHNING[1987]),a Still,although migration settled bornworkers offoreign permanently8. largefraction III.2 Data andSample Panel(GSOEP).The GSOEPis a householdThe datasetwe use is theGermanSocio-Economie based panelsurvey,similarto theUS Panel Studyof IncomeDynamics(PSID) or theBritish HouseholdPanel Study(BHPS). Initiatedin 1984,theGSOEPoversamplesthethen-resident we havedescrimovement whichstemsfromthemigration in Germany, population immigrant bed above.In thefirst wave,about4500 householdswitha Germanbornhouseholdheadwere head.Thedataareunique bornhousehold witha foreign andabout1500households interviewed, over on a boostsampleof immigrants a longperiodoftime. in providing repeatedinformation as well as fortheforeignbornfromtheover-sample, For ouranalysis,we use observations forthenativebornfromthestandard observations sample. The householdhead in a householdandovertheage of 16 is interviewed. Each individual the andbelow interviewing inthehousehold aboutall otherindividuals information age. provides areincludedinthepanel. andformtheirownhouseholds wholeavehouseholds Individuals decitotemporary andtheirrelationship andMESTRES[2010]wherewe analyzeremittances 7. In DUSTMANN migration access we have In that IV with an estimator fixed effects a these we address strategy. paper, sions, bycombining problems are estimates effects oftimeperiods.We findthattheIV-fixed fora largenumber information forremittances torepeated sizetotheupward error measurement biasthrough duetothedownward OLSestimates, closetotheoriginal beingofsimilar Assetsandsavings- whichwe analyzeinthispaper- areonlyobserved unobservable biasinducedthrough heterogeneity. onceortwiceoverthecourseofthepanel. from the in2004was 3.7 millionpeople,ofwhicharound60 percentoriginated laborinGermany 8. Thestockofforeign here considered countries (OECD [2006]). sending 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions 293 CHRISTIANDUSTMANNAND JOSEPMESTRES TheGSOEPdataprovidesinformation on assetholdingsinboththehomeandhostcountries for 1988. For that there is detailed information onthetypeofassetholdings, their only yearonly, - whether theassetis heldin thehostor sourcecountry9. valuesand- importantly Assetholdcash,savings,home ingsreferto thetotalamountofassetholdingsofthehousehold(including of financial in net each location Home refers etc.) ownership, obligations separately. ownership or anyotherproperty in to all houses,apartments ofthehouseholdat market both home prices andhostcountries. locationsonlyfor1992 and 199410. Savingsare declaredin bothhomeand hostcountry in the host to the net Savings country correspond monthly savingsofthehouseholdtransformed toa yearlylevel.Savingsinthehomecountry to correspond theindividual yearlyamountsaved inthehomecountry andtransformed tohousehold level.We construct thetotalamountofhouseholdsavingsas thesumoftheyearlyamountsthehouseholdsavedin bothlocations.We will use thosetwoyearswherewe observesavingsin bothlocations(1992 and 1994) to studythe allocationof savings.All monetary variablesareat thehouseholdlevelin realamounts, where thereference yearis 2002. A further ofourdatais thatimmigrants ineachwaveofthe uniquefeature provideinformation whether intend to remain in or whether wish toreturn home panel they permanentlyGermany, they atsomestageinthefuture. We usethisinformation todifferentiate between thosewhodo andthose whodo notplantoreturn tothehomecountry. Ifeconomicdecisionsareinvolved, itis likelythat thesearebasedon intentions ofthissort,rather thanonpossiblerealizations ata laterstage. In addition, we observeindividual andhouseholdcharacteristics inthehostcountry, as well as information on family members whoarelivinginthecountry oforigin. IV. Results IV.7 Descriptive Evidence Individual Characteristics: As we mention on above,we measuresavingsandassetaccumulation thelevelofthehousehold. Whenwe refer to characteristics ofindividuals within we households, refer totheheadofhousehold. Entriesin TABLEI showthattheaverageage ofhousetypically holdheadsinoursampleis 45, andthatmigrants residedslightly lessthan22 yearson averagein Almost90 percent oftheheadofhouseholds aremale,and78 percent areemployed. The Germany. net household income is around Euros 2002 Around 93 25,000 yearly average (in prices). percent ofhousehold headsarenotsingle;however, havenativepartners. Almost40 percent only7 percent ofall headsofhouseholds thattheygrewup ina ruralarea.The lastvariablemeasures the report return intention ofthehousehold head.On average, 51percent ofthehousehold headsinoursample thattheywouldwishtoreturn totheirhomecountry at somepointinthefuture. report We the amount of for 1992 and 1994.Forimmigrants, Savings: study yearly savings savings refer tothetotalamount savedas wellas theamounts savedinhostandhomecountries. As a referHeresavingsrefer tothetotalamount saved. ence,wealsoreport savingsfornativebornindividuals. 9. See theAPPENDIX fora more detaileddescriptionof thedata construction. 10. The amountof savings in Germanyis declared from1992 onwards,while the amount saved in the home countryis declared only for1984-1990,1992 and 1994. See thedata constructionAPPENDIX forfurther details. 294 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS,ASSET HOLDINGS,AND TEMPORARYMIGRATION TABLE I. - SummaryStatistics Mean Std. Dev. Age 45.46 11.29 Age AtArrival 23.73 9.16 YearsSinceMigration 2 1.76 5.87 NumberYearsEducation HouseholdIncome inHousehold NumberChildren 9.43 1.92 25,186 0.85 12,157 1.07 NumberAdultsinHousehold 2.66 1.21 inHousehold NumberEmployedIndividuals Sex 1.59 0.94 0.88 0.32 Employed Non Single 0.78 0.41 0.93 0.25 NativePartner 0.07 0.26 SpouseAbroad Abroad Children 0.04 0.19 0.07 0.26 RuralChildhood 0.39 0.49 Temporary NumberofObservations 0.51 0.50 2,456 Note:Calculationsbasedon GSOEPdata,1988,1992,1994.Individualinformation totheheadofhousehold.HouseholdIncomein 2002 Euros. corresponds inthefirst IntheupperpanelofTABLEII, we describesavingsforall immigrants pairofcolumns. andpermanent with betweenimmigrants temporary In nexttwopairsofcolumnswe distinguish betweenimmigrant themeandifference return pairofcolumnswe report plans.In thefollowing of household11. the head refers to return on information The its t-statistic. and plans groups in thehostcountry. save that households of all 48 About percent report they immigrant which on savinga positiveamount), The averageamountsavedis 2,046Euros(notconditional withpermanent to 7.4 percentof overallhouseholdincome.Immigrants migration corresponds with those than than in the host to save less are plans,andthey temporary country likely plans to one percent in host in the save a loweramount.The difference savings corresponds country The proportion ofthehouseholdincome.The nextrowshowsthesavingsinthehomecountry. whosave is morethan4 percentage intentions withtemporary ofimmigrants pointshigherthan withtheamountsavedbeinghigheras well.Both withpermanent thatofimmigrants intentions, different are significantly differences Finally,thelastthreerowsreportthetotal statistically. thetotalamountofsavingsofnatives we report ofyearlysavings.As a pointofreference, amount householdsreportsavings,as compared in thelasttwocolumns.One in twoof all immigrants The totalaverageamountofsavingsis equalto 2,199Euros ornativehouseholds. to 65 percent to 10% fornatives), householdincome,as compared to 8.1% ofimmigrants' (whichcorresponds whichis lowerthantheaveragesavingsfornativesbothin absolutevalueandrelativeto their householdincomes12. andreturn forwhichbothsavingsorassetinformation 11. We useall observations plansarereported. datafromtheGermanCentralBank's(Bundesbank) 12. Thehousehold savingsratioinourdatais inlinewithaggregate onaverageto 12%ofhousehold wherehousehold FinancialAccounts, disposableincomefor1992and savingscorrespond 1994 (BUNDESBANK [2008]). 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions 295 CHRISTIANDUSTMANNAND JOSEPMESTRES TABLE II. - Savings,HomeOwnershipandAssets JtlT All Immigrants Temporary Permanent Immigrants Immigrants жг . Natives Mean Mean N Mean N Mean N Difference TemporaryPermanent t Mean N Migrants Savings In HostCountry thatsaves Proportion AverageAmount 48.3% 1871 50.4% 1,871 2,218 782 782 46.7% 2,046 1,089 1,922 1,089 3.6% 297 1.55 1.74 7.4% 1809 8.0% 754 7.0% 1,055 1.0% 1.89 4.8% 155 1,902 1,902 7.3% 230 797 797 3.0% 101 1,105 1,105 4.3%** 129* 4.35 2.31 HH 0.7% AverageValueas Percentage Income TotalSumHomeandHostCountry 49.8% Proportion 2,199 AverageAmount 1,838 0.9% 769 0.5% 1,069 0.3% 1.26 1871 1871 52.8% 782 782 47.6% 2,448 2,021 1,089 1,089 5.2%* 427* 2.24 2.32 8.1% 1,809 8.8% 754 7.5% 1,055 1.3%* 2.17 8.4% 860 857 5.5% 577 575 14.1% 27,049 283 282 -8.6%** 7,324 -19,726** -4.31 -4.91 283 275 20,408** 5.37 4.69 283 283 11.8%** 188 3.26 0.03 229 229 1.5% -29,786** 0.041 -3.76 172 172 18,706** 155 155 -2,595 (1655) (Std.Dev.) HH AverageValueas Percentage Income (3,585) (3,700) In HomeCountry Proportion AverageAmount 0,207) (Std.Dev.) (1921) (Std.Dev.) HH AverageValueas Percentage Income HomeOwnership In HostCountry thatHoldsProperty Proportion AverageValue (Std.Dev.) In HomeCountry thatHoldsProperty Proportion AverageValue 13,814 (55,896) 44.2% (1,127) (3,835) (38,995) 30,043 859 835 36,764 TotalSumHomeandHostCountry thatHoldsProperty 49.9% Proportion 44,381 AverageValue 859 832 44,443 629 629 20,805 (Std.Dev.) (59,823) 50.5% (66,816) 53.8% (Std.Dev.) (81,167) In HostCountry thatHoldsAssets Proportion AverageValue 31,649 In HomeCountry thatHoldsAssets Proportion AverageValue 48,723 595 595 54,130 TotalSumHomeandHostCountry thatHoldsAssets 83.7% Proportion 66,777 AverageValue 486 486 65,949 AssetHoldings (Std.Dev.) (StdDev.) (Std.Dev.) 73.4% (96,379) 71.6% (75,382) (103.651) (78,167) 74.0% (126,877) 75.7% (81,975) 84.3% (96,608) 576 560 (1,259) (3,974) (78,417) 31.4% 16,356 (38,792) 576 558 44,255 42.0% 400 400 50,591 (87,108) 72.5% (128,105) 423 423 35,424 61.6% 331 331 68,544 (53,899) 82.6% (117,606) 19.1%** 14.0%** 1.7% 65.4% 6,901 2,888 6,901 (5,101) 9.9% 6,688 43.6% 3,329 60,973 3,194 (111,939) 3.46 2.76 0.48 0.25 79.8% 2,959 104,966 2,959 (162,877) * meandifference at 5%; ** significant meandifference at 1% significant Note:Calculations basedonGSOEPdataonhouseholdlevel.Averageamount(in 2002 Euros)notconditional on reporting anypositive amount.Property includeshouse,apartment oranyotherproperty. Assetholdingsrefertothetotalamountofassetholdings Ownership netoffinancialobligations, etc.Savingsinthehostcountry tothenetmonthly cash,savings,property, including corresponds savings ofthehouseholdtransformed to annualamount.Savingsin thehomecountry to theyearlyamountremitted to thehome corresponds and is that saved. Both and Asset country Property Ownership holdingsrefertotheyear1988.SavingsFlowsreferto years1992and 1994.We use all observations forwhichrespective information is available. 296 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS, ASSET HOLDINGS, AND TEMPORARY MIGRATION Whenwe distinguish withpermanent betweenimmigrants andtemporary there intentions, is a cleardifference betweenthetwogroups,withthosewithtemporary intentions savingmore, inabsoluteas wellas inpercentage oftheirhouseholdincome. terms, HomeOwnership andAssets:A setof questionsaskingaboutassetholdingswas included in thesurveyin 1988.Forimmigrants, and assetholdings, bothin questionsrelateto property andinthehomecountry. Fornatives, relatetototalproperty andassetholdings. Germany questions inthesecond(homeownership) andthird We report (assets)panelsofTABLEII. As descriptives whilethenextcolumnsdistinguish thefirst twocolumns before, report averagesforallimmigrants, We includethe andtemporary betweenimmigrants withpermanent migration plansrespectively. inthelasttwocolumns. natives'averageas reference in householdsreportowninghousingproperty of about 8.4 percent all immigrant Only intentions reveals with and between immigrants temporary permanent Germany. Distinguishing witha permanent however.While14 percentof immigrants remarkable differences, migration do so. in Germany, intention ownhousingproperty only5 percentofthosewhowishto return latter is much lower for the stock the value of the Likewise, category. housing in thehomecountry. About44 percentof all In thenextrowwe reporthomeownership thelastrowcombines inthehomecountry. households Finally, report owningproperty immigrant in Germany whoholdproperty ofimmigrants withtheproportion thisnumber (firstrow).Half 44 percent of with This contrasts in home or host holdproperty either oftheimmigrants country. is about The averagevalueof immigrants' thatholdproperty. nativebornhouseholds property lowerproperty lowerthanthatofnatives.However,thismaypartlyreflect 30 percent pricesin thecountries oforigin. andpermanent withtemporary Inthenextcolumnswe distinguish againbetweenimmigrants with return households in two One return housingstock owning plans reports immigrant plans. In addition, intentions. with 1 of those to 3 inthehomecountry, permanent compared just percent who wish toreturn. for those twice as is morethan inthehomecountry thevalueofproperty high aremorelikelyto ownproperty. However,thetotalvalueofhome Overall,temporary migrants andthosewhodo not. wish toreturn who households is similar between immigrant ownership refer tothetotalamount Asset on asset information Thenextpanelreports holdings holdings. the Forimmigrants ofassets(including cash,savings,property, etc.)netoffinancial obligations. in the home assets held in and held between assets a distinction draw country. Germany, questions in thehostcountry, The numbers suggestthatifwe consideronlyassetholdingsof immigrants thisdifference lowerthanthoseofnatives.However, is considerably theamount ofassetholdings holdassetsalso in thehome reducedwhentakingintoaccountthatimmigrants is significantly betweenimmigrants inthedistribution ofassetholdings Thereis againa starkdifference country. oftheirassetsin hold most Whilepermanent intentions. andpermanent withtemporary migrants The totalamount holdassetsmostlyin thehomecountry. thehostcountry, migrants temporary ofassetholdingsis slightly migrants. higherforpermanent andnative betweenimmigrant showthat- foranycomparison thesefigures To summarize, in oforigin. the and assets consider it to is households country savings immigrants' important andassetholdingsbetween in savings,totalproperty also showdifferences thefigures Further, There are also starkdifferences and with plans. migration permanent immigrants temporary withtemporary thesetwogroupsas towheretheseassetsareheld.Immigrants between migration 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions 297 CHRISTIAN DUSTMANN AND JOSEP MESTRES andassetsthanimmigrants withpermanent plansholdless property plansin thehostcountry, butmoreinthehomecountry. Thispointsat different wealthallocationprofiles betweenthose who want to return and those who do It not. also that the migrants suggests way immigrants maypossiblyaffectthehousingmarketin thehostand thehomecountrydependson their re-migration plans. IV.2 Conditional Results We nowfocuson thedifferences betweenimmigrants withtemporary andpermanent migration inTABLEII do notaccountfordifferences we report inhouseholdandindiplans.Thenumbers vidualcharacteristics. inassetholdings todifferences incontemporaTheyalsorelatedifferences neousintentions abouta possiblereturn. We nowprovidesomefurther wherewe condition results, on differences in householdcharacteristics, anduse information aboutcontemporaneous return plans(inthecase ofsavings),andaveragepastreturn plans(inthecase ofassetholdings). Forbothsavings, homeownership andassetholdings, we estimate linearprobability models for the outcome and OLS and Tobit models for the amount of variable, (LPM) binary savings. IN.2.1 Savings As we discussin SectionII, it is generallyambiguouswhetherimmigrants withtemporary withpermanent migration planssavemorethanimmigrants plans.Thisis inlinewiththefigures in TABLEII, whichshowthattemporary aremorelikelyto save bothin thehostand migrants thehomecountry. Someofthesedifferences be in composition between may dueto differences thetwogroups.To investigate thisfurther, we nowpresent someconditional wherewe estimates, use datafortwoyearsofourpanel(1992 and 1994)thatprovideinformation on theamountof in each location. We construct a measure for total and the ratio ofsavingsinthe savings savings, homevs thehostcountry. If immigrants withtemporary intentions havea higherpropensity to intentions overall.If(in save,we shouldobservethattheysavemorethanthosewithpermanent havea preference forshifting then addition) temporary immigrants savingstothehomecountry, theratioofhometohostcountry should be related to return savings positively plans. Resultsarereported inTABLEIII. In thefirst column,we use thetotalamountofsavingsas thedependent variable.Columns2 and3 distinguish betweensavingsinthehomeandhostcounColumn 4 the ratio of in the home andtotalsavings.We report LPM try. reports savings country inthefirst results OLS in results the second and Tobit in results the third The panel, panel panel13. onthetemporary variableincolumns1 suggest thatoverall,immigrants pointestimates migration withtemporary withpermanent migration planssave morethanimmigrants migration plans. Estimates arehowevernotstatistically When into significant. splitting up savings savingsinthe homeandhostcountry, are and associatedwith temporary migration plans positively significantly for theratioofsavingsinhomevs savings thehomecountry only.In thelastcolumn,we report hostcountry, whichis positively andsignificantly relatedtoreturn are plans.Thus,theestimates 13. Total amountof savings correspondsto the amountreported.See the APPENDIX (TABLE I) forfullregressionresults withall the additionalcontrolvariables forTotal Savings. These resultssuggestthatboth household income and employmentof thehead of thehouseholdaffectsavingspositively.Age, yearssince migration,and educationof thehead of household do not seem to be significantly associated withhousehold savings conditionalon household income and employment of thehead of thehousehold. 298 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS, ASSET HOLDINGS, AND TEMPORARY MIGRATION withthehypothesis withtemporary consistent thatimmigrants migration planshavea preference rather thaninthehostcountry. forholdingtheirsavingsintheorigincountry Theydo notpoint withtemporary return withpermanent at immigrants planssavingmorethanimmigrants plans, conditional on otherobservable characteristics. TABLE III. - Savings - Home and Host Country LinearProbabilityModel Total Temporary (Std.E.) Savings(=1 Yes,=0 No) HostCountry HomeCountry 0.024 (0.028) 0.044 (0.028) 0.034*** (0.013) OLS Temporary (Std.E.) Total Amount Savings HostCountry 14.09 (233.296) -89.944 (221.917) Total Amount Savings HostCountry 179.181 (166.524) U>59 51.145 (151.934) U>59 RatioHomevs Total HomeCountry 99.476 (60.442) 0.029*** (0.009) Tobit Temporary (Std.E.) Observations RatioHomevs Total HomeCountry 116.343*** (41.013) 1^685 0.020*** (0.006) 1,680 at 1% at 5%; *** significant at 10%; ** significant significant includetimeand country Note: GSOEPdata (1992 and 1994). Householdlevel. All specifications dummiesand conditionon age, yearssince migration (and its square),education,gender,marital statusof thehead of householdas well as householdincomeand numberof statusand employment household.Standarderrorsare clusteredby household.Tobit adultsand childrenin thehostcountry on the tothecoefficient resultsshowunconditional correspond Reportedcoefficents marginaleffects. intention variable. temporary contemporary * andAssetsHoldings IV.2.2 Property betweenassetholdings Distinction ofimmigrants. andassetholdings We nowturntotheproperty is onlyavailableforoneyear(1988).Assetsmeasurethestock inthehomeandinthehostcountry overprevious ofassetsaccumulated years, upto 1988.As thestockofassetshasbeenaccumulated for1984-1988as a regressor14. intention we use theaveragereturn in TABLEIV and foroverallassetholdings We showtheresultsforproperty ownership forthe estimates to above.We reportthecoefficient in TABLEV, usingsimilarspecifications the inthefirst column total We report intentions overtheperiod1984-198815. averageofreturn location betweentheproperty whilecolumns2 and3 differentiate ofproperty amount ownership, in home the theratioofproperty Column4 reports in homeandhostcountry. country holdings onthe LPM estimation the results of a shows Thefirst tototalproperty. withrespect simple panel haveanyproperty holdings. migrants binaryoutcomewhether on returnplans since the startof thepanel in 1984. 14. We onlyhave information 15. See theAPPENDIX (TABLE II) forthe fullset of regressionresults.Household income and household size are, respecconditionalon household income and tively,positivelyand negativelyassociated with asset accumulation.Furthermore, household size, households witholder and bettereducated heads hold more wealth. 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions 299 CHRISTIAN DUSTMANN AND JOSEP MESTRES TABLE IV. - PropertyOwnership - Home and Host Country LinearProbabilityModel Temporary (Std.E.) (=1 Yes,=0 No) Property Ownership Total HostCountry HomeCountry 0.122* -0.079** 0.190** (0.055) (0.03) (0.054) OLS Total Temporary (Std.E.) 7,456.369 (8,832.433) Amount Property HostCountry HomeCountry RatioHomevs Total 0.214** -2,2919.058** 30,939.940** (5,565.475) (6,773.723) (0.054) Tobit Total Temporary (Std.E.) Observations 11,678.86 (7,608.978) 739 Amount Property HostCountry HomeCountry RatioHomevs Total 0.243** -4,573.934** 28,297.509** (1,520.867) (5,849.031) (0.061) 738 719 739 * at 5%; ** significant at 1% significant Note:GSOEPdata(1988). Householdlevel.All specifications includetimeandcountry dummiesand condition on age,yearssincemigration maritalstatus,householdincome, (and itssquare),education, statusandnumberofadultsandchildreninthehostcountry household.Property employment includesthepurchaseofhouse,apartment oranyotherproperty, inthehostand inthe ownership homecountry. Tobitresultsshowunconditional to marginaleffects. Reportedcoefficents correspond theaverageintention to return up to 1988(1984-1988). Theresultsshowthaton averagetemporary migration plansareassociatedwitha 12.2pertoholdproperty. return associated centagepointshigherlikelihood Further, plansarenegatively withowninghousing in the host but with associated property country, positively owningproperty at home.The intention to return to thehomecountry is associatedwithan 8 percentage points lowerlikelihood to owna housein thehostcountry, butan almost20 percentage pointshigher likelihood to owna houseinthehomecountry. In thesecondand thirdpanelof theTable,we reportOLS and Tobitresultsforthevalue oftheproperty held(in 2002 Euros).Theseresultsindicatethatthetotalvalueofproperty that with hold is not different from that of those immigrants temporary migration plans significantly withpermanent difference intheproperty wealth plans.However,thereis a starkandsignificant allocation between hostandhomecountries. Thoseimmigrants withtemporary return plansholda loweramount ofproperty inthehostcountry anda higher amount ofproperty inthehomecountry, inbothOLS andTobitspecifications. In thelastcolumn,we againreporttheimpactof temporary migration planson theratio between heldinthehomecountry, andtotalproperty As forsavings,thisratio property holdings. is strongly skewedtowards forimmigrants withtemporary holdingsinthehomecountry migrationintentions, inbothspecifications. We showtheresultsforassetholdingsin TABLEV. Again,we report thecoefficient estimatesfortheaverageofreturn intentions overtheperiod1984-1988.The structure ofthetableis identical totheprevious one.Theresults thatthetotalvalueofassetshelddoesnotdiffer suggests between and location significantly temporary permanent However,thegeographical immigrants. as towhereassetsareheldis different: households withtemporary intentions holdmoreassetsin 300 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS,ASSET HOLDINGS,AND TEMPORARYMIGRATION thehomecountry, andless inthehostcountry, aftercontrolling forhouseholdincomeandother heldassetsto characteristics. As theresultsin thelastcolumnshow,theratioofhomecountry to return totalassetsis positively related plans. TABLE V. - AssetHoldings- HomeandHostCountry LinearProbabilityModel AssetHoldings(=1 Yes,=0 No) HomeCountry HostCountry Total Temporary (Std.E.) -0.036 (0053) 0.002 (0.057) 0.096 (0.059) OLS Total Temporary (Std.E.) 6,718.334 (15,522.831) AssetHoldings Amount HomeCountry RatioHomevs Total HostCountry -29,173.987** (10,146.918) 28,606.677** (10,310.560) 0.133* (0.060) Tobit AssetHolding Amount HomeCountry RatioHomevs Total HostCountry 0.140* 25,472.782** -17,238.395* 3,510.74 (0.068) (8,603.728) (6,714.548) (12,167.750) 432 531 546 432 Total Temporary (Std.E.) Observations * at 1% at 5%; ** significant significant includetimeand countrydummies Note: GSOEPdata (1988). Householdlevel,All specifications and conditionon age, yearssince migration (and its square),education,maritalstatus,household household.Asset statusand numberof adultsand childrenin thehostcountry income,employment tothetotalamountofassetholdingsnetoffinancial cash,savings, including obligations, holdingsrefer Tobitresultsshowunconditional marginaleffects. etc.,inthehostand inthehomecountry. property, to return totheaverageintention up to 1988(1984-1988). correspond Reportedcoefficents V. Conclusions inrelation totheirreturn ofimmigrants Inthispaper,we analyzesavingsandassetholdings plans. in thehomeandinthe held and assets stock between Ouranalysisdistinguishes savings,housing of distribution We findevidencethatreturn hostcountry. plansare associatedwitha different locations. home host and between and assets country savings,property and intotalsavings,property difference thatthereis no significant Ourresultsshowfurther on conditional andtemporary withpermanent betweenimmigrants assetholdings migration plans, toreturn withintentions characteristics. household observable However,immigrants background in thehome andmorelikelyto ownproperty inthehostcountry areless likelyto ownproperty is quitesubstantial. andthisdifference Thus,ourstudypointsat immigration policies country, than a different thatfavorpermanent impacton thedomestichousingmarket having migrations policies. policiesthatfavortemporary of savings arealso associatedwithholdinga higherproportion plans migration Temporary in thehome held assets of for both andassetsinthehomecountries. groups immigrants Finally, of is that an assessment of our another Thus arequitesubstantial. paper finding important country inthe needstotakeaccountofwealthandassetsaccumulated wealthaccumulation immigrants' nature. areofa temporary themoremigrations Thisis moreimportant, homecountries. 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SOI CHRISTIANDUSTMANNAND JOSEPMESTRES We shouldemphasize between thetemporariness ofmigrations and againthattherelationship in and asset accumulation behavior that we show this should not be as causal. savings paper interpreted ourstudy substantial inthelocation differences ofsavingsandasset Nevertheless, pointsatpossibly between withdifferent intentions aboutthepermanency oftheir Further, holdings immigrants migration. we findnoevidence thattotalsavingsandassetsheldaredifferent between thesetwogroups. ACKNOWLEDGEMENTS Thisresearch was financially German, Korean,andNorwegian supported bytheAustrian, the Multi-donor Trust Fund on Labor JobCreation, and Markets, governments through EconomicGrowth administered andLaborunit.Dustmann bytheWorldBank'sSocial Protection fromtheNORFACE Wewouldliketothankthe acknowledges support Migration Programme. HillelRapoport, andtwoanonymous referees forconstructive comments. editor, Correspondence: Department ofEconomics,University CollegeLondon,GowerStreet, LondonWC1E6BT,U.K.Email: c.dustmannt&ucLac.uk Appendix Data Construction We usedatafrom theGermanSocio-Economic Panel.Assetholdings arereported for1988,based on a specialsurveymodule.Savingsare reported for1992 and 1994. Our sampleconsistsof households whoseheadwas borninTurkey, immigrant Greece,Yugoslavia,Italyor Spain. Information on return are in each wave ofthepanel.Individuals wereasked plans provided whether toremainpermanently inGermany, orwhether homeat theyintend theywishto return somestageinthefuture. We construct a binary variablethatdefinesas temporary thosewhoplan inthefuture. toreturn As return is relatedto pastreturn we plansmaychange,andassetaccumulation intentions, construct an averagereturn intention variableforthelastfiveyearsbeforeassetsaremeasured (thatis, 1984-1988). All ourincomevariablesarereported inrealterms(inEuros,deflated tothebaseyear2002), andathousehold level.Householdincomecorresponds tothenetmonthly incomeofthehousehold transformed toannuallevel.Theexactwording ofthequestionis "Ifeverything is takentogether: howhighis thetotalmonthly incomeofall thehouseholdmembers at present? Pleasegivethe netamount, theamountafterthededuction oftaxandnationalinsurance contributions. monthly suchas rentsubsidy, childbenefit, subsistence Regularpayments allowances, government grants, Ifnotknownexactly, estimate the amount". etc.,shouldbe included. please monthly Information on household is availablefor1984-1990,1992and savingsinthehomecountry to theyearlyamountsavedin thehomecountry 1994,andcorresponds bythehousehold.The todeclaretheamountsentortakentothehomecountry forthepurpose questionasksindividuals of"savingsforlater".We transform thisvariabletothehousehold level.Information onhousehold in the host is available for 1992 onwards and to the netmonthly savings country corresponds of the household transformed to level. The the asked is "Do you savings yearly question survey have an amount of left formajorpurchases, orsavusually money overeachmonth emergencies, on savingsinboththehomeandthehost ings?Ifyes,howmuch?".Thisimpliesthatinformation 302 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS, ASSET HOLDINGS, AND TEMPORARY MIGRATION thetotal is availableonlyfortwoyears(1992 and 1994).Forthoseyears,we construct country amountofsavingsas thesumofsavingsinbothlocations. ina specialsurvey in 1988whereimmigrants drawnfromquestions We useassetinformation bothinthehomeandinthehostcountry wereaskedfortheirassetholdings separately. refer tothetotalamountofassetholdings(including Assetholdings cash,savings,property, ofthequesThewording bothinthehomeandhostcountries. etc.)butnetoffinancial obligations, tionis "Ifyoucouldaddup all thewealthofthishousehold(including cash,goodsandproperty total value? Please makesureto its what will be without own but furniture), approximate you includesthe thatyoucouldhaveon them".Property loansandcredits all themortgages, subtract at market or anyotherproperty houses,apartments prices,bothin thehomeandhostcountries. ofthequestionis "Areyoutheownerof(specifictypeof thewording Foreachtypeofproperty, valueis,thatis,howmuchmoney itscommercial Ifyes,howmuchdo youestimate property)? householdamounts to the All entriescorrespond willyougetifyou sold itnow?" aggregated tothebase year2002. declaredin 1988,inEuros,deflated TABLE I. - Total Savings TotalSavings (=1 Yes, =0 No) LPM Age/10 (Std.E.) 10 YearsSinceMigration/ (Std.E.) YSM-Squared/100 (Std.E.) Log HH Income (Std.E.) NumberAdultsHH Host (StdE.) HH Host NumberChildren (Std.E.) Head HH Employment (Std.E.) I NumberEmployedHH (Std.E.) NumberYearsEducation (Std.E.) Male Head HH (Std.E.) NonSingle (StdE.) NativePartner (Std.E.) SpouseAbroad (StdE.) Abroad Children (Std.E.) RuralChildhood (Std.E.) Temporary {Std.E.) NumberofObservations Amount TOBIT OLS 15.043 (10.888) -44.678 (579.814) -92.917 (133.413) 20.828 (20-377> -72.385 (954.475) -131.676 (197.844) -0.001 (°°°2) 0.063 (0100) -0.019 (°024) 0.286*** (°089) -0.047** (0021) -0.073*** (°014) 0.152*** (°038) 0.002 (°029) 0.003 (О-007) 0.019 (O-040) -0.107** (0049) 0.072 (°053) 0.084 (°083) -0.067 2,691.025*** (935.280) -280.394 (219.850) -352.268*** (118.826) 62.768 (292.701) 236.447 (282.581) -57.785 (51.418) 167.13 (469.024) -829.918** (357.352) 397.2 (403.186) 947.252 (623.977) 529.332 4,072.837*** (255.809) -469.503*** (107.843) -433.709*** (79.531) 642.021*** (233.220) -212.843 (129.628) -57.937 (38.930) 97.344 (234.473) -889.294*** (288.561) 369.201 (269.934) 0.047* (0°28) 0.044 (°028) 238.197 (230.128) 14.09 (233.296) 354.596** (154.448) 179.181 (166.524) (O-064) (665.530) 1,127.213*** (432.572) 213.764 (343.230) 1,659 1,659 1,659 0.024 0.151 0.142 R-squared * at 1% at 5%; ♦♦♦significant at 10%; ** significant significant areclustered errors dummies.Standard includetimeandcountry Note:GSOEPdata(1992 and 1994).Householdlevel.All specifications byhousehold. 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions 3Q3 CHRISTIANDUSTMANNAND JOSEPMESTRES TABLE IL - TOTALPROPERTYANDASSETHOLDINGS TotalProperty 1 (= Yes,= ONo) LPM Age/10 0.065* 6,992.39 0.439* (0026) (Std.E.) YearsSinceMigration/ 10 (31,944.141) (0.228) -40,474.467 -1,594.633 -9,021.404 -0.035 12,131.839 57,212.046** 58,354.315** -0.028 -6,545.521* -6,026.319** -0.03 -0.027 -932.913 -1,940.823 -0.007 (0.017) (4,933.143) 1,709.037 0.028 14,874.005 13,933.759 -15,353.334* -13,211.824* (Std.E.) (0.061) (0.016) (Std.E.) Head HH Employment 19,354.011** 0.084 0.344** (Std.E) 20,423.945** 42,990.847 (0.055) HH Host NumberChildren 0.066** 13,363.174 (Std.E) NumberAdultsHH Host 8,167.744* (0.025) -0.104 Log HH Income (4,123.640) TotalAssetHoldings Amount (= 1 Yes,= ONo) LPM OLS TOBIT (3,567.110) (0-223) (Std.E.) YSM-Squared/100 Amount OLS TOBIT (0.018) (35,502.003) (8,818.495) (9,691.445) (2,666.964) (2,948.062) (7,884.666) (8,676.121) (2,269.345) (2,504.003) (0.057) 0.342** (0.062) (0.016) (7,165.884) (66,282.065) (16,414.831) (5,616.612) -21,355 (52,034.042) 5,675.163 (12,929.400) 108,771.151** 100,564.981** (18,119.010) (14,485.252) -5,194.61 -5,626.295 -79.446 -266.035 (4,546.187) (3,557.839) (3,833.999) 0.021 -1,801.02 (7,806.949) (0.056) NumberEmployed HH -0.013 -2,358.308 -3,104.786 -0.032 NumberYearsEducation -0.014 2,321.072 (1,568.693) 366.429 (1,330.714) -0.012 0.172 18,773.464 25,832.981 -0.007 29,624.491 19,496.307 0.015 25,354.86 17,984.492 -0.025 66,389.392* 51,424.168* -0.068 -9,697.935 -8,631.509 -0.084 13,884.182 15,725.938* 5,437.249 (0.058) (Std.E.) (0.029) (Std.E.) (0010) (Std.E.) Male Head HH (0.094) (Std.E.) NativePartner (0.081) (Std.E.) SpouseAbroad (0.087) (Std.E.) Children Abroad 0.128* (0.057) (Std.E.) RuralChildhood -0.01 (0.037) (Std.E.) 0.122* Temporary R-squared (4,635.975) (15,077.112) (13,105.281) (13,894.105) (9,335.730) (5,901.585) (3,859.390) (15,119.807) (11,312.488) (11,827.956) (0.027) (0.009) (0.127) (0.101) (0.081) (16,136.560) (7,779.869) 7,329.063** (2,737.697) (36,803.571) (29,301.068) -19,941.21 (23,635.383) (12,670.834) (6,112.197) 4,813.823* (2,151.789) (28,689.586) (22,861.355) -19,590.228 (18,863.518) (7,599.675) 0.006 (0.052) 25,068.459 19,478.381 3,562.898 -0.006 6,965.387 4,994.18 6,718.334 3,510.74 (4,932.460) (0.034) (15,167.549) (9,993.592) (11,837.909) (7,791.598) 7,456.369 11,678.86 -0.036 (0.053) (15,522.831) 738 718 718 432 432 432 0.138 0.117 0.011 0.131 0Л8 Q.Q1 (0.055) (Std.E.) NumberofObservations (9,181.844) (8,832.433) (7,608.978) (12,167.750) * at 10%; ♦* significant at 5%; *** significant at 1% significant Note:GSOEPdata(1992 and 1994).Householdlevel.All specifications includetimeandcountry dummies.Standard errors areclustered byhousehold. References AMUEDO-DORANTES C. and S. POZO(2002): "Precautionary andYoung SavingsbyYoungImmigrants EconomicJournal, Vol. 69,No. 1,pp.48-71.[290] Natives,"Southern BARTHE., В. BRATSBERG and О. RAAUM(2004): "Identifying of Immigrants EarningsAssimilation underChanging Macroeconomic Scandinavian Journal Vol. Conditions," 106,No. 1, ofEconomics, pp. 1-22.[289] BAUERТ.К. andM. SINNING(2009): "The SavingsBehaviorofTemporary and Permanent in Migrants Vol. 24,pp.421-449.[290] ,"Journaloj Population Economics, Germany BAUERТ.К.,D.A. COBB-CLARK, V. HILDEBRAND andM. SINNING (2009): "A Comparative Analysisof theNativity WealthGap,"EconomicInquiry, [290] forthcoming. BLAU F.D. and J.W. GRAHAM(1990): "Black-White Differences in Wealthand AssetComposition," TheQuarterly JournalofEconomics, Vol. 105,No. 2, pp.321-39.[290] BLITZR.C. (1977): "A Benefit-Cost in West-Germany, AnalysisofForeignWorkers 1957-1973,"Kyklos, Vol. 30,No. 3,pp.479 - 502. [293] 304 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions SAVINGS,ASSET HOLDINGS,AND TEMPORARYMIGRATION BOHNING W. (1987): StudiesinInternational St.Martin'sPress,New York.[293] Migration, andtheEarnings ofImmigrants," Journal BORJAS G.J.(1985):"Assimilation, ChangesinCohortQuality, of LaborEconomics, Vol. 3,No. 4, pp.463-489.[289] - WhatHappenedtoImmigrant andChangesinCohortQualityRevisited BORJAS G.J.(1995):"Assimilation Vol. No. in the Journal Labor Economics, 13, 2,201-245.[289] 1980s," of Earnings Journalof UrbanEconomics, BORJASG. J. (2002): "Horneo Population," wnershipin the Immigrant Vol. 52,No. 3,pp.448-476.[290] and theRelativeWages of BRATSBERG В., E. ERLINGand O. RAAUM(2006): "Local Unemployment The Review Evidence from the Current of Economicsand PopulationSurveys," Immigrants: Vol. 88,No. 2, pp.243-263.[289] Statistics, imZeitverlauf," FÜRARBEIT(2009): "Arbeitslosigkeit BUNDESAGENTUR http://www.pub.arbeitsamt.de/ hsťservices/statistik/detail/z.html. [293] bank.de/down1991to2007,"http://www.bundesAccounts forGermany BUNDESBANK (2008):"Financial 1991_2007.en.xls.[295] load/statistik/fìnanzierungsrechnung/anlageverhalten_ ofForeign-born ontheEarnings B.R. (1978):"TheEffect ofAmericanization CHISWICK Men,"TheJournal Vol. 86,No. 5, pp. 897-921.[289] ofPoliticalEconomy, COBB-CLARKD.A. and V. HILDEBRAND(2006): "The WealthAnd AssetHoldingsOf U.S.-BornAnd Households:EvidenceFromSipp Data," ReviewofIncomeand Wealth,Blackwell Foreign-Born 52, Publishing, Vol. No. 1,pp. 17-42.[290] RatesSo Low?ImmigraCOULSONN. E. (1999):"WhyAreHispanicandAsian-American Homeownership Vol. 45,No. 2, pp.209-227.[290] JournalofUrbanEconomics, tionandOtherFactors," JournalofPopulationEconomics, ofTemporary C. (1993): "Earnings DUSTMANN Migrants," Adjustment Vol. 6, pp. 153-168.[289] - A LifeCycleAnalysis," Workers DUSTMANN С (1995): "SavingsBehaviorofMigrant Zeitschrift fuer 51 1 533. Vol. und 4, [289,290] pp. Socialwissenschaften, WirtschaftsDUSTMANNС (1996): "ReturnMigration:The EuropeanExperience,"EconomicPolicy,Vol. 22, pp.215-250.[293] Journal andPrecautionary С (1997):"Return DUSTMANN ofDevelopment Savings," Uncertainty Migration, Vol. 52,pp.295-316.[289] Economics, Choiceafter DurationandActivity DUSTMANN C. andO. KIRCHKAMP (2002): "The OptimalMigration Vol. 67,pp.351-72.[289] JournalofDevelopment Economics, Remigration," JournalofDevelopment andTemporary DUSTMANN C. andJ.MESTRES(2010): "Remittances Migration," No. 62-70. 92, 1,pp. Economics, [293] intheUnitedStates:TheRoleof ofImmigrants FRIEDBERG R.M. (1993): "TheLaborMarketAssimilation BrownUniversity [289] manuscript. Age atArrival," andMigrants' ofReturnMigration GALORO. andO. STARK(1990): "Migrants'Savings,theProbability International EconomicReview,Vol. 31,No. 2, pp.463-67.[290] Performance," Journalof LUBOTSKYD. (2007): "Chutesor Ladders?A Longitudinal Earnings," Analysisof Immigrant Vol. 115,No. 5,pp. 820-867.[289] PoliticalEconomy, EconomicPapers, andCapitalMarketImperfections," MESNARDA. (2004): "Temporary Oxford Migration Vol. 56,pp.242-262.[289] Outlook:SOPEMI 2006Edition.[293] OECD (2006):International Migration WealthAssimilation:EviOSILI U.O. and A. PAULSON(2004): "ProspectsforImmigrant-Native Federal ReserveBank of Chicago Working dence fromFinancialMarketParticipation," Paper No. 04-18. [290] 2010 JANUARY/JUNE © ANNALSOF ECONOMICSAND STATISTICS- NUMBER97/98, This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions 3Q5 CHRISTIANDUSTMANNAND JOSEPMESTRES inAsian-American PAINTER theImpactof G.,L. YANGandZ. Yu (2003):"Heterogeneity Homeownership: HouseholdEndowments andImmigrant status,"UrbanStudies,Vol. 40,pp. 505-530.[290] in Germany," IZA DiscussionPaper SINNINGM. (2007): "Wealthand AssetHoldingsof Immigrants No. 3089.[290] andEconomicPerformance in Germany," Urban SINNINGM. (2009): "Home-Ownership of Immigrants Vol. 47,pp.387-409.[290] Studies, WOODRUFF С andR. ZENTENO(2007):"Migration Networks andMicroenterprises inMexico,"Journal of Vol. 82,No. 2, pp. 509-528.[289] Economics, Development YANGD. (2008): "International Remittances andHouseholdInvestment: EvidencefromPhilipMigration, Vol. 118,No. 528,pp.591-630.[289] pineMigrants' ExchangeRateShocks,"TheEconomicJournal, 306 © ANNALSOF ECONOMICS AND STATISTICS - NUMBER97/98,JANUARY/JUNE 2010 This content downloaded from 140.247.212.176 on Tue, 02 Jun 2015 00:54:54 UTC All use subject to JSTOR Terms and Conditions