Working papersYear 2008 Paper 9Migration and developmentA theoretical perspectiveHein de HaasIMI does not have a view as an Institute and does not aim to present one. The views expressed in thisdocument are only those of its independent author.2AbstractThis paper aims to put the debate on migration and development in a broader historical perspective of migration theory in particular and social theory in general. The scholarly debate on migration and development has tended to swing back and forth like a pendulum, from developmentalist optimism in the 1!"s and 1#"s, to structuralist and neo$Marxist pessimism and scepticism over the 1%"s and1&"s, to more nuanced views influenced by the new economics of labour migration, 'livelihood( approaches and the transnational turn in migration studies as of the 1"s. )uch discursive shifts in the scholarly debate on migration and development should be primarily seen as part of more generalparadigm shifts in social theory. The shift that occurred over the 1"s was part of a more general shift away from grand structuralist or functionalist theories towards more pluralist, hybrid andstructuralist approaches attempting to reconcile structure and actor perspectives. *owever, attempts to combine different theoretical perspectives are more problematic than sometimes suggested due to incommensurability issues and associated disciplinary divisions.)ince 2""", there has been a remarkable, and rather sudden, renaissance of optimistic views, in particular in the policy debate, as well as a boom in empirical work on migration and development. This has coincided with the rediscovery of remittances as a 'bottom up( source of development finance and the celebration of the transnational engagement of migrants with the development of theirorigin societies. *owever, such optimism has tended to go along with a striking level of amnesia of decades of prior research. Migration and development is anything but a new topic. The accumulated empirical and theoretical evidence stress the fundamentally heterogeneous nature of migration$ development interactions as well as their contingency on spatial and temporal scales of analysis and more general processes of social and economic change, which should forestall any blanket assertions on migration$development interactions.+urrent policy and scholarly discourses naively celebrating migration, remittances and transnational engagement as self$help development 'from below( also shift attention away from the relevance of structural constraints and the important role states and other institutions play in shaping favourable general conditions for social and economic development to occur. This raises the fundamental ,uestion whether the recent shift towards optimistic views reflects a veritable change in-increasingly transnationally framed. migration$development interactions, the use of other methodological and analytical tools, or is rather the deductive echo of a general paradigm shift from dependency and state$centrist to neoliberal and neoclassical views in general.The lack of theoretical rootedness and largely descriptive nature of much empirical work has haunted the improvement of theories. /s a result of the general lack of a common theoretical thread, most empirical work 0 especially from outside migration economics 0 remains isolated, scattered, andtheoretically underexplored. 1eal progress in the understanding of the factors determining the fundamental heterogeneity of migration and development interactions is only possible if more empirical work is designed to test theoretically derived hypotheses and, hence, to improve the generali2ed understanding of migration$development interactions.Keywords: Migration theory, development theory, social theory, remittances, neo$Marxism,developmentalism, transnationalism.Author: *ein de *aas is 1esearch 3fficer at IMI, 4niversity of 3xford. 5mail 0he i n .dehaa s6 ,eh.ox.ac.uk7Table of contents1. Introduction .......................................................................................................................... 12. General migration theories .................................................................................................. 42.1. The neo$classical e,uilibrium perspective ....................................................................................... 82.2. *istorical$structural theory and asymmetric growth........................................................................ %2.7. 9uestioning the push$pull framework.............................................................................................. &3. Towards a transitional migration theory ......................................................................... 127.1. The mobility transition ................................................................................................................... 127.2. The migration hump ....................................................................................................................... 1#4. Internal dynamics and feedbacs: networs and migration systems ............................ 1!8.1. )ocial capital, chain migration and network theory....................................................................... 18.2. Migration systems theory ............................................................................................................... 21". #igration and de$elo%ment o%timists $s. %essimists ...................................................... 23!.1. The dawning of a new era: developmentalist views ...................................................................... 28!.2. The migrant syndrome: cumulative causation and structuralist views .......................................... 2#!.7. Towards a pluralist perspective...................................................................................................... 71&. 'luralist $iews on migration and de$elo%ment interactions .......................................... 34#.1. ;ew economics of labour migration -;5 1&&., in which he formulated his 'laws of migration(.*e saw migration as an inseparable part of development, and he asserted that the maIor causesofmigrationwereeconomic.Migrationpatternswerefurtherassumedtobeinfluenced byfactorssuchasdistanceandpopulationdensities-)keldon1%:1..Thisperspective,inwhich people are expected to move from low income to high income areas, and from denselyto sparsely populated areas, that is, the general notion that migration movements tend towardsa certain spatial$economic e,uilibrium, has remained alive in the work of manydemographers, geographers, and economists ever since -+astles C Miller 2""7:22., and, as wewill see, is also the underlying assumption of push$pull theories./lthoughtheissueofmigrationhasnotattractedsubstantialattentionwithinmainstreameconomictheoryitself-GauerCLimmermann1&:!>Dassaris1&$%.,economicexplanationshavenonethelessdominatedpopularandscholarlythinkingonmigration./tthemacro$level,neo$classicaleconomictheoryexplainsmigrationbygeographicaldifferencesinthesupply anddemandforlabour.Theresultingdifferentialsinwagescauseworkerstomovefromlow$wage,labour$surplus regionstohigh$wage,labour$scarceregions.Migrationwillcauselabourtobecomelessscarceatthedestinationandscarcer at the sending end. +apital is expected to move in the opposite direction. In a perfectlyneo$classicalworld,thisprocessof'factorpricee,uali2ation(-the*eckscher$3hlin model.will eventually result in growing convergence between wages at the sending and receiving end!-*arrisCTodaro1%">1anisCKei1#1>)chiff18>TodaroCMarus2ko1&%.. In the long run, this process would remove the incentives for migrating./tthemicro$level,neo$classicalmigrationtheoryviewsmigrantsasindividual,rationalactors,whodecidetomoveonthebasisofacost$benefitcalculation./ssuming freechoiceand full access to information, they are expected to go where they can be the most productive,thatis,areabletoearnthehighestwages.Thiscapacityobviously dependsonthespecificskills a person possesses and the specific structure of labour markets.;eo$classical migration theory seesrural$urban migration asanconstituent part of the wholedevelopmentprocess,bywhichsurpluslabourintheruralsectorsuppliestheworkforce fortheurbanindustrialeconomy -1#.wasthefrontrunner ofthe'dependency(theory,whichhypothesi2ed that global capitalism -and migration as one of its manifestations. contributed tothe 'development of underdevelopment( -see also Garan 1%7.. The dependency school viewsmigration not Iust as detrimental to the economies of underdeveloped countries but also as oneof the verycauses of underdevelopment,ratherthan asa pathtowardsdevelopment./ccording tothisview,migrationruinsstablepeasantsocieties,undermines theireconomiesanduprootstheirpopulations. 5mmanuelFallersteinBs-1%8>1&".world$systemstheoryclassifiedcountriesaccording totheirdegreeofdependency,anddistinguished betweenthecapitalist 'core( nations, followed by the 'semi$peripheral(, 'peripheral(, and isolated nationsinthe'external(area,whichwerenot-yet.includedinthecapitalistsystem.Inthisperspective, the incorporation of the peripheries into the capitalist economy is associated withputting a -migration. drain on them, exactly the opposite of factor price e,uali2ation presumedby neo$classical theory. Instead of flowing in the opposite direction of capital as predicted byneo$classical category, the idea is that labour follows where capital goes.*istorical structuralists have critici2ed neo$classical migration theory, stating that individualsdonothaveafreechoice,becausetheyarefundamentally constrainedbystructuralforces.1ather thanamatteroffree choice, people are forced to move because traditional economic&structureshavebeenunderminedasaresultoftheirincorporation intotheglobalpolitical$economic system. Through these processes, rural populations become increasingly deprived oftheirtraditionallivelihoods,andtheseuprootedpopulations becomepartoftheurbanproletariat to the benefit of those core areas that rely on cheap -immigrant. labour.*istoricalstructuralistshavebeencritici2edforbeingtoodeterministandrigidintheirthinkinginviewingindividualsasvictimsor'pawns(thatpassively adapttomacro$forces,thereby largely ruling out individual agency. Moreover, rigid forms of historical structuralismhavebeenrefutedbyrecenthistory,asvariousformerlydevelopingandlabourexportingcountries haveachievedsustained economicgrowth inthepastdecadesdespite0or perhapsthanks to 0 their firm connection to global capitalism -)en 1.. Kor most southern 5uropeancountriesandsome'/sianTigers(,theincorporationintoglobalcapitalismand,possibly,highlabourmigrationhaveapparentlyworkedoutwell,despitegloomypredictionssomedecades ago -/lmeida 1%7> Dapademetriou 1&!..Thereisincreasingconsensus thatcapitalismassuchcannotbeblamedfortheproblemsofunderdevelopment, but thatthespecific developmental effects of incorporation of a region orcountry into the global capitalist system seems to depend much more on the conditions underwhichthistakesplace,thatis,howtheincorporation isembeddedintowiderinstitutionalstructuresaswellastheinternalsocio$political cohesion andeconomic strength ofcountriesand regions. Thus, depending on these circumstances, the incorporation into global capitalismcan have both positive and negative effects in different areas of development and on differentgroups of people within society. In the same vein, -labour. migration cannot automatically beinterpretedasadesperateflightfrommisery,notonly becauseitisseldomthepoorestwhomigrate,butalsobecausewecanatleastnotlogically ruleoutthe possibility thatmigrationfacilitatesdevelopmentthroughreverseflowsofcapital-remittances.,knowledge,ideas,attitudes, and people -return migration..2.3. Questioning the push-pull frameworGoth neo$classical and historical$structural theories of migration generally fail to explain whysomepeopleinacertaincountry orregionmigrateandothersdonot-Massey etal17>1eniers1:#&".,andwhy peopletendtomigratebetweenparticularplacesinaspatiallyclustered,concentrated,typically non$randomfashion.Itcanthereforebeusefultolookatsome of the spatial models developed by mainly geographers and demographers.factorsassociated withtheareaofdestination>so$calledintervening obstacles-suchasdistance,physicalbarriers,immigrationlaws,andsoon.>andpersonalfactors.3lesen,2""2../lthoughthereisroomtocritici2etheportrayalofregionssuchassub$)aharan/fricaasweaklymobile,itistruethatmostmigration inthisregion isregional and thatmigration toindustrialised countries isrelativelyweak-GakewellCde*aas2""%..ItseemsnottobecoincidentalthattheworldBsmaIorlabouremigrationcountriesinrelative$to$population terms-e.g.,Mexico,Morocco,Turkey,Dhilippines. are typically upper$lower to middle income countries,3.2. The migration humpFhathasremainedunobserved sofaristhatmigrationeconomistshave0apparentlyunintentionally 0providedadditionaleconomicevidenceforandfurtherrefinementofthetransitionalmigrationperspectivebyuncoveringtheanatomyofthe'migrationhump(.Martin -17. and Martin and Taylor -1#. argued that a temporary increase in migration 0 amigration hump 0 has been a usual part of the process of economic development.In the earlystagesofdevelopment,anincreaseinwealthtendstoleadtoariseinmigration,sinceacertainthresholdofwealthisnecessary toenablepeopletoassumethecostsandrisksofmigrating.Fithincreasingwealthandtheestablishmentofmigrantnetworks,anincreasingproportion of the population is able to migrate, selectivity of migration tends to decrease, andthisprocessof'development(initially tendstoleadtoanincreasingdiffusionofmigrationacross communities.3nly atlaterstagesofdevelopment,doesemigrationtendtodecreaseanddoregionsandcountriestendtotransformfromnetlabourexporterstonetlabourimporters-cf.Gohning18:1#> Martin 12:8%1> Martin C Taylor 1#> 3lesen 2""2:181> 1otte et al 1%.#. Thishashappenedinrecentdecadeswithcountriesasdiverseas)pain,Italy,Ereece,Ireland,Malaysia,Taiwanand)outh=orea,whichusedtobelocatedonglobal'labourfrontiers(,which have shifted further on since then.5conomicdevelopmentanddecreasingincomedifferentialswithdestinationcountriestherefore tend to have a P$curve or inverted 4$curve effect on emigration, steeply increasing intheinitialphasesofeconomicdevelopmentandonlylatergradually decreasing-MartinCTaylor1#..This'migrationhump(seemstoreflectratherade,uately past-Massey 11.and contemporary -3lesen 2""2:181. global migration patterns and shows a striking degree ofcorrespondence withLelinskyBs and)keldonBs models. Inmany ways, LelinskyBs modelcanbeseenasamigration humpavantlalettre.Thisappears tobeanother ground forreIectingrelativistclaimsthatnogeneralinferenceswhatsoevercanbemadeaboutthehistorical,spatio$temporalevolution ofmigrationindirectorindirectrelationtobroader developmentprocesses.Infact,LelinskyBs,)keldonBsandMartinandTaylorBsapproachescanbeintegratedintoonesingle,spatio$temporal@transitionalBmigrationperspective,whichrecognises the complex,non$linear interlinkagesbetweentheoccurrenceofvarious formsofmigrationandgeneralsocial,technological,economicanddemographictransformationprocesses embodied in the term 'development(.# Drior research seems to have indicated that emigration tends to decrease significantly if the income differentialbetweensendingandreceivingcountriesreachesvaluesbetween 1:7and1:8.!,providedthattheemigrationcountry is growing fast -Martin 18> Kaini and Menturini 18> 3lesen 2""2..1%Thus,itispossible toincorporate thecriticismonLelinksyBs original theoreticalperspectivewithoutreIectingitbutbymodifyingit,providedthatsomeconditionsaremet-Ae*aas2""%a..Kirst,suchamodifiedmodelshouldallowforhistorical$regional,time$spatialvariationsviewedasavarietyofpatternswithinageneralsystem.)econd,arealistic@transitionalB perspective onmigration should perceive therelation between development andtheoccurrence ofspecifictypesofmigrationasanindirect andprobabilistic ratherthanas adirect and deterministic one. Moreover, this relation is not unidirectional,because adecreasingrelativedevelopmentlevelmayalsotransformanimmigrationcountry intoanemigrationcountry.Intheend,itisthelevelofdevelopmentrelativetootherplacesandcountriesthatdeterminesmigrationpropensities./swewillsee,thisfitsintotheidea-particularly advanced by thenew economics of labour migration. thatrelative, not absolute,deprivation is one of the main determinants of migration.Third,afinalnecessarymodificationtothistransitionalmigrationperspectiveseemstobroadenourviewof@developmentB beyondanarrowfocusoneconomicanddemographicvariables.Themigration$development nexusalsocomprisesimportantsocialandculturaldimensions.Gesidesrisingincomesandimprovedtransportationandcommunicationinfrastructure, also factors like education and access to information and social capital initiallytend to increase peopleBs aspirations and abilities to migrate -Ae *aas 2""!..Therefore, itisnecessary to conceive @developmentB as the more general increase of peopleBscapabilities. )uch a more comprehensive approach to development is offered by /martya )en-1.,whoconceivesdevelopmentas'theprocessofexpandingthesubstantivefreedomsthatpeopleenIoy(.Inordertooperationali2ethese@freedomsB,heappliestheconceptofhuman capability, which relates tothe ability of human beings to lead lives they have reasonto value and to enhance the substantive choices they have 0 including the option to migrate asastrategytoachievelivelihoodimprovements.)enBscapabilitiesapproachcontrastswithnarrower views of development that are largely restricted to income indicators.Tosummarise,thefundamentalideaisthatsocio$economic developmenttendstoincreasepeopleBs capabilities and aspirations to migrate. *owever, while the effect of development oncapabilities to migrate is more or less linear, the effect on peopleBs aspirations to migrate is ismorelikelytoresembleaPorinverted4$curveasaconse,uence ofdecreasing levelsofrelative deprivation.Kigure 1depicts how, according to transitional models, 'development( and different forms ofmigration are supposed to be generally related to each other. Fithin an integrated transitionalmigrationperspective,@developmentBinitsbroadersenseisinitiallylikelytoincreasepeopleBs propensity tomigrate, first predominantly internally and in later development stagesincreasingly internationally. 3nthelonger run,decreasing spatial-internalandinternational.opportunity gapsmightdecrease peopleBs incentives tomigrate,which islikely toresultinaslow$down in both rural$to$urban and international out$migrationbut, most importantly, in anincreasingdiversificationofmigrationpatterns,alongwithincreasingimportanceofimmigration, rather than a decrease in mobility as such.international migrationinternal migration1&/igure 1. The general effect of de$elo%ment on migration %atterns according to transitional modelsimmigrationurban$urbanSintra$urbancircular migrationemigration rural$urbanmigration'development('development(+ource: Ada%ted from 0elinsy 11!(1:23323 see also #artin and Taylor 11!!&2+ertainly, there is a dangerin explainingmigration bysinglingout factors such asdemographic and economic development. /fter all, countries with roughly the samepopulationgrowthratesandlevelsofdevelopmenttendtoshowhighly divergingmigrationcharacteristics. Towhatextentmigration willoccur, and where migrants willgo, depends ontheinterplaybetweenmanyvariables-cf.1ussell1!.,suchasgeographicaldistance,historicalpoliticalandculturallinks,pastandpresentimmigrationpolicies,income,unemployment,leisure,education,thenatureofpoliticalregimes,safety,environmentalfactors,accesstoinformation, socialcapital-amongwhich accesstomigrantnetworks., andsoon.Pointly,they determinethegenerallevelanddistributionofwealthandotheropportunities-cf.freedomofexpression,employment,education.,peopleBsperceptionsof'here( and 'there(, and, by that, the propensity and capability of people to migrate. /lthoughcountriesorregionsresemblingLelinskyBscategories-b.and-c.seemtohaveahigherpropensitytoexperiencehighinternationalout$migration,theextenttowhichthisoccursseems to vary greatly within a possibilistic range.Theaddedvalueoftransitionalmodelsistheircapacitytolinktheoccurrenceofspecificforms of migration to broader processes of social and economic development into one spatio$temporalperspective.Thisoffersinsightintothecomplexanddynamiclinksbetweenmigrationanddevelopment,andtacklesidesreuessuchas'poverty breedsmigration(.*owever,transitionalmodelsconcentrateonthechangingspatio$temporalmorphology ofmigration,andalmostuni,uely interpretchangingmigrationpatternsasaresultofbroaderdevelopmentprocesses.Theytherebyignorethereciprocaleffectsofmigrationondevelopment processes both at the destination and origin./shasbeenargued,itisthisvery inability tointegrate bothcausesandeffectsofmigrationwithinabroadertheoreticalperspectiveondevelopmentthathashauntedmuchmigrationresearch so far. /s a constituent component of broader development processes, migration is tobesimultaneously consideredasbothadependentandindependentvariable.InordertodosufficientIusticetothelatterobservation, the remainder ofthis paper willexamine the maintheoreticalperspectivesthathavebeendevelopedonthevariousfeedbackmechanismsthroughwhichmigrationpatternsarebothmodifiedandperpetuatedovertime,andthecontroversialissueofhowmigrationaffectsdevelopmentinmigrants$sendingsocieties.)ubse,uently, we will attempt to integrate both sets of theories within one single perspectiveon migration and development.4. Internal dynamics and feedbacks: networks andmigration systems!.1. "ocial capital# chain migration and networ theory)choorl1&..Thisdrawsourattentiontotheroleofnationstates,geographicalproximity,institutions,socialnetworks,andculturalandhistorical factors in creating new migration patterns.3nceacertaincriticalnumberofmigrantshavesettledatthedestination,however,otherforcescomeintoplay.Theoftencoincidentalchoicesmadeby pioneermigrantsorlabour$recruiting employers tend to have a great influence on subse,uent migration patterns. There isAe*aas2""7.,beinghesitantorunwilling toassistprospectivemigrants.EurakC+aces12..Migrationsystemslinkpeople,families,and communities over space in what we nowadays would refer to as transnationalcommunities-cf.Mertovec1..Thisresultsinaratherneatgeographicalstructuring andclustering of migration flows, which is far from a random state, in whichformalandinformalsubsystemsoperate to perpetuate and reinforce the systematic nature ofinternationalflowsbyencouragingmigrationalongcertainpathways,anddiscouragingitalongothers.Theendresultisasetofrelativelystableexchanges....yieldinganidentifiable geographical structure that persists across space and time -MabogunIe 1%":12.22;etwork theory canalready explainwhy,once amigration system has developed, ittends tooperaterelativelyindependently ofgovernmentpolicy intervention.Migrationsystemtheoryaddstothat,inlinewith=rit2etal12..Thereareclearparallelsbetweenmigrationsystemtheory, worldsystemtheory and)keldonBs regionalisation of global development tiers, as they alldraw onthenotionofmigrationconnectingsemi$peripheral@migrationfrontierBcountries-)keldon1%. to wealthy @core countriesB. Kawcett -1&. stressed the relevance of both national andtransnational networks, which tend to be closely interwoven, blurring the distinction betweeninternalandinternationalmigration-Martin12:8!&>Mc=eeCTisdell1&&:81&.. Miaaprocessofso$calledleapfrogging,internationalmigrationisoftenprecededbyinternalmigrantmoves,andreturning migrantsmay settleinotherthantheirplacesoforigin.Inaprocesscalledrelaymigration-/ri2pe1&1.,returnmigrationmaybefollowedbythemigration of another family member.Thefactthattheinitialcircumstances atboth the receiving and sending end aremodified bythemigrationprocessimpliesthatthecausesandconse,uences ofmigrationshould notbestudiedseparately,butaspartofthesamesystemandprocesses. Migration simultaneouslyreshapes the socio$economic 'development( context at both the origin and destination, whichintheirturn,arelikely toinfluencesubse,uent migrationpatterns.Korexample,remittancessent back to family members could alter the social and economic context in the areas of originandencouragesubse,uentmigration-vanAalenetal2""!.. Galdwin 1%"..Fhereas at the national level the feelings were mixed, the disappointment seemed particularlyhighconcerning theeffectsofmigration ondevelopment atthe regional and locallevel.Thedominantvisionwasthat,althoughremittancesweresentback,they wererarely investedinsuchawaythattheycouldcontributetodevelopmentintheregionsandcommunitiesoforigin. This coincided with the renaissance in historical$structuralist Marxist thinking in socialsciencesanddependencythinkingindevelopmenttheoryandpractice-Krank1##b>Krank1#..*istoricalstructuralistandrelatedcentre$periphery theoriesseemedapplicabletothestudyofmigrationanddidnotbodewell-AeMasCMermeulen17..Infact,theseapproachesturnedtheargumentofneo$classicalanddevelopmentalistapproachesupsidedown:migrationdoesnotdecrease,butincreasespatialandinter$personaldisparitiesindevelopmentallevels./lsoinasocio$culturalrespect,theeffectsofmigrationwereincreasinglyseenasdetrimental,astheywouldprovokeconsumerist,non$productiveandremittance$dependentattitudesofnonmigrants.Kromtheearly1%"s,numerousacademicpublications have supported thehypothesis that migration contributes to the 'development ofunderdevelopment(insteadofthereverse-/lmeida1%7>Ginford2""7>Qasin1&%..5venintheheyday ofneo$Marxism,therehavebeenempiricalstudiesthatstressed thenon$uniform, differentiated impactof migration-/badan$4natetal1%#>*einemeiIeretal1%%>Denninx1&2..Thispartlyreflecteddisciplinary divisions./nthropologistsandsociologistsseemedthemostpessimistic,whilegeographers andeconomists tendedtowards amore moderatestance-AeMasCMermeulen17..Thestreamofplainlypessimisticpublicationsseemedtodecreasetowardstheendofthe1&"s,andthegeneraltoneofthescholarlydebatehasbecomemoremoderatesincethen.This timing seems not coincidental, as this has corresponded with a general paradigm shift incontemporary socialtheory,awayfromgrandtheoriestowardsmorepluralist,hybridapproaches.Moststudiesfromthelate1&"sand1"s seemtohavedepartedfromastructuralist stance,andincreasingly acknowledge the differentiated, non$deterministic natureofthemigrationimpacts-cf./dams11>/hlburg1!>Pones1&>=eely CTran1&>Mc=ee C Tisdell 1&&..*owever, the influence of pessimistic views on migration and development has long remainedenormous,andmany ofitsviews0inparticularonmigrantsBallegedinclinationtowastemoneyonconspicuousconsumption0stillpervadescholarly and,particularly,policyviewsonmigrationanddevelopment.Gothexplicitlyandimplicitly,manyviewsderivedfromcumulativecausationtheoryOwhichseesmigrationasade$developing, destabili2ing,and,hence,undesirable,productofpoverty,asaproblemwhichcanandshouldbe'solved(throughclosedborderpoliciesincombinationwithaidanddevelopmentprogramsOstillretain currency among academics, politicians, and the media.QettherearegoodtheoreticalargumentstoreIectthepropositions ofthemigrantsyndromeview.ThiscriticismisnotIustbasedonanotherinterpretationofempiricaldataduetoageneralparadigmshiftinsocialtheoryawayfromstructuralistthinking.Therearealsoanumber of theoretical fallacies and internal logical inconsistencies in the pessimistic views onmigrationanddevelopmentderivedfromcumulativecausationtheory.ThefirstobIectionisthedeterministicandself$affirmingnatureofthemodel,whichdoesnotgiveroomforheterogeneity inthespecific,locali2edmigrationimpacts.Korwhatprecisereasonswouldpositive spread effects -e.g., remittances. never match negative backwash effects under certainconditionsY)econd,likeother'circularfeedback(migrationmodelsOsuch asmigrationnetwork theoryOthe vicious cycle of 'pauperi2ation( in the periphery and migration and growth atthe coreseemstogoonadinfinitum.Thisissuspect,sinceitseemsunrealistictoassumethat therearenocounter$mechanismswhichlevel$offorchangethenatureofthissupposedlylinearprocessovertime.*owfarcan'under$development(goonwithoutdecreasingmigrationY+umulativecausationandrelatedmodelsimplicitlysupposealinear$negativeassociationbetweendevelopmentandmigration.*owever,ashasbeenarguedearlier,onecannotgenerallyassumealinear$positiverelationshipbetweenspatialdisparitiesinincomeandwelfareandtheoccurrenceofmigration./sempiricalevidencesupporting transitionalmigration theories has convincingly demonstrated, the relationship between 'development(Owhetherexpressedintermsofthedemographicand'vital(transition-Lelinsky1%1.orincomeand other opportunity disparitiesOand migration isP$ or inverted 4$curve like ratherthan linear and inversely proportional -Martin C Taylor 1#..Third,thereisaninherentlogicalcontradiction inthetwocentralargumentsthatmigrationpessimistsgenerallymake:3ntheonehand,theysay,migrationbreedsine,uality becausemigrantscomefrombetter$offgroupswithinsociety.3ntheotherhand,itisarguedthatfurtherimpoverishment oftheregionoforiginleadstomoremigration. Thisislogicallyinconsistent, asthefirstargumentcorrectly supposes thatacertain threshold of wealth needsto precede migration and the second argument supposes a negative$linear relationship betweenwealth and migration. Thus, the migration pessimists tend to make an imprecise, inconsistentanalysis ofthecausesoflabourmigration through theirselectiveignorance oftheinherentlyselectivecharacterofmigration.Therefore,they suggestthatincreasingdeprivationleadstoincreasedlabourmigrationperse.*owever,thisreflectsanerroneousunderstanding oftheroleofmigrationinthedevelopmentprocess,whichisfarmoretemporally andspatiallyheterogeneous than the migration pessimists suggest./fourth,empiricalreasonto,uestiontheone$sided negativeperceptions onmigrationanddevelopmentistheincreasingbodyofempiricalresearchthatappearedinthe1&"sand1"s indicating that, under certain circumstances, migration has in fact played a positive roleinthedevelopmentofregionsandcountriesoforigin.Insouth$5uropean countriessuchas)pain,Italy,andEreeceand5ast/siancountries suchasMalaysiaand)outh$=orea,remittanceshaveplayedasignificantpositiveroleintheirOsuccessfulOnationaleconomicdevelopmentand,inreactiontosustainedsocio$economic developmentintheircountriesoforigin, many international migrants have in fact invested significantly in private enterprises./fteralongperiodofsustaineddevelopmenthasout$migrationlevelled$offanddecreased,andtheseformerlyemigrationcountrieshavenowbeentransformedintonetimmigrationcountriesorareontheway todoingso.Thisagainexemplifiesthattransitionalmigrationtheory0drawingoncontributionsbyscholarssuchasLelinsky -1%1.,)keldon-1%.andMartin and Taylor -1#. 0 is more realistic and has more explanatory power than the largelycircularcumulative causation theory. /pparently, the self$reinforcing cyclical mechanisms ofasymmetrical, polari2ing development cannot be taken as axiomatic.)o,looking back,doesthisallmeanthatthemigrationoptimists were rightY3bviously, theanswerisno.Infact,neitherthestructuralistpessimistsnorthefunctionalist optimistswereright,asthevariationandcomplexity ofreal$lifemigration$developmentinteractionsistoohigh tobeabletofittheminto deterministic theoretical schemes predicting the developmentoutcomeofmigration.DapademetriouandMartin-11.alreadyarguedthatthereisnoautomatic mechanism by which international migration and remittances result in development.Kewmigrationresearcherswouldnowadays contestthisgeneralassertion-cf.1ussell12.,but it gives an uncomfortable feeling to leave it simply at that. The main challenge here is toelaborateanappropriatetheoreticalframeworkthatisnotdeterministicandthatisrefinedenough to deal with the heterogeneity and complexities of migration$developmentinteractions, butthatdoesnotrestrictitselftoempiricismand'allislocalandsingular(relativism.Kindings from different studies are clearly contradictory. In some cases, migration seems to have a positive effect on the different dimensions of social and economic development, in other cases it seems to have no effect or even negative effects -Ae *aas 2""%c.. This not Iustpertains to differences in paradigmatic orientationOleading to different interpretations of similar empirical dataOand research methodology, but also to real, existing differences. 5mpirical research has highlighted that the spiralling down mechanisms of cumulative causation do not always hold true, but that the perfect neo$classical world does not exist in reality either. )tructural constraints such as highly une,ual access to employment, markets, education and power do matter in the daily struggle of most people in the developing world, and do severely limit their capability to overcome their situation of poverty and general underdevelopment. *ence, discarding the rigidity of classical structuralist and neo$Marxist approaches is not to say that structural constraints do not matter. Fhile neo$classical and developmentalist perspectives on migration and development tend to underestimate,structuralist perspectives tend to overestimate the importance of structural constraints. *ence, an improved theoretical perspective on migration and development has to be able to account for the role of structureOthe constraining or enabling general political, institutional, economicsocial, and cultural context in which migration takes placeOas well as agencyOthe limitedbut real capacity of individuals to overcome constraints and potentially reshape the structuralcontext.#. $luralist !iews on migration and de!elo"mentinteractionsThesecondhalfofthe1&"smarkedinseveralwaystheendofnotonly rigidhistorical$structuralism,butalsoofrigidtheoreticalthinkingingeneral.Inthisnewera,socialscientists,influencedby post$modernist thinking andEiddensB -1&8. structuration theory11,sought to harmoni2e actor$ and structure$oriented approaches. 1ecognition of the relevance ofbothstructure andagency seemsessentialforthemigration anddevelopment debate,asthisenablesustobetterdealwiththeheterogeneity ofmigrationimpacts.Insucha'pluralist(approach, the results of the structure$actor interactions allow for a greater variety of outcomethanwouldhavebeenallowedfromeitherthesingleaggregationofindividualdecisionmaking -)keldon 1%:1&. or from the unidirectional imperatives of structures.Thisgeneralparadigmshiftinsocialtheoryhasalsodeeplyaffectedthespecificmigrationanddevelopmentdebate.3verthe1&"sand1"s,themaincontribution tothinking onmigrationanddevelopmentcamefromtheneweconomicsoflabourmigration-;5 )tark C Taylor1>1>TaylorCFyatt1#.. )uchmarketsareoftenweakly developed ordifficulttoaccess for non$elite groups. In particular through international remittances, migration can be ahouseholdstrategytoovercomesuchmarketconstraints,andmaypotentiallyenablehouseholdstoinvestinproductiveactivitiesandtoimprovetheirlivelihoods-)tark1&"..Fhile remittances do not play a role in neo$classical migration theory, within ;5 -b. increasing household income>and -c. overcoming constraints on economic activities and investments in the region of origin,the new economics of labour migration also critici2ed the very methodological design of mostprior migration research. /ccording to Taylor et al. -1#a:1.,priorworkhasbeenundulypessimisticabouttheprospectsfordevelopmentasaresultofinternationalmigration,largelybecauseithasfailedtotakeintoaccountthecomplex,oftenindirect ways that migration and remittances influence the economic status of households and the communities that contain themThis pertains to criticism of the lack of analytical rigor, the prevalence of deductive reasoningoverempiricaltesting,aswellastheimportantmethodological deficienciesofmuchpriorempiricalwork.;5McAowellCde*aan1%>)coones 1&..Itisincreasingly recogni2ed that labour migration is often more than a short$term survival or crisiscoping strategy or a stereotypical 'flight from misery(. 1ather, it is often a deliberate decisiontoimprovelivelihoods,enableinvestments-Gebbington1:2"2%.,andhelptoreducefluctuations in the family income that has -often. been entirely dependent on climatic vagaries-Ae*aanetal2""":2&>McAowellCde*aan1%:1&.. Migrationcanthenbeseenasameans to ac,uire awider range of assets which insure against future shocks and stresses -Ae*aanetal2""":7"..Thereisnoapriorireasonwhy thisdiversification$through$migrationargument cannot also be extended to urban households.Thiscomessurprisinglyclosetothepremisesof;5-II. thedevelopment contextatthe localorregional level>and-III. thefactorsrelatedtothe migrantandherdirect social and economicenvironmentOinmostcasesthehousehold.Thesethreesetsofvariablesarereciprocallylinked through various direct functional relations and feedback mechanisms./igure 4. General conce%tual framewor for analysing migration6de$elo%ment interactionsMacro development contextbaeMigrationdc