Sie sind auf Seite 1von 31

GenderedMigrantSocialCapital:Evidencefrom Thailand

SaraR. Curran,University Washington of FilizGarip, Princeton University University ChangY.Chung,Princeton Kanchana MahidolUniversity Tangchonlatip,

Abstract to data studiesof cumulative we causation, Employing longitudinal from Thailand replicate extendcurrentknowledge measuring by frequency trips,durationof time away,levelof of networkaggregation and sex composition migrantnetworks to (villageor household), of a menandwomenin Thailand. estimate modelofprospective Wefindthat among migration that level tripsand duration timeawayhavedistinct influences of uponmigration; household are that migrantnetworks more influentialthan village level migrantnetworks; female and male migrantnetworkshave differentinfluencesupon migration migrantnetworks socialcapitalinfluences andwomen's men outcomes; thatmigrant and, migration differently. Our elaboration evidenceas to how genderand family providessignificantquantitative imbuemigration variously dynamics. Introduction networks evolve,accumulate generatehigher can and Theideathatmigrant thanexpected out levels of migration of communitiesof originhas yielded a considerablenumberof studies and some policyattention(Massey1990a;Massey 1990b;Massey and empirical et 1987; Massey,Goldring, al. 1994; MasseyandZenteno1999).Thisidea, Garcia-Espana is cumulative socialcapital is and causation, frequently equatedwiththe conceptof migrant describedas a process by which migration those who are in origin propensities, among or communities(whetherthey are return migrants otherwise),grow with each additional in stream.Itgrows as information increasesand reciprocal develop ties migrant a migrant of and the betweenorigin destination. Further, theoryholdsthatthe importance otherfactors in predicting migration propensitiesbecomes less important a context of high levels of and Garcia-Espana 1987; Massey, (Massey 1990a; Massey 1990b; Massey migration et Goldring, al. 1994; Massey and Zenteno 1999). Muchof this researchfocuses upon and behaviors et Mexico-U.S. Stecklov, al.2002; (Davis, patterns Mexican migration migration
in and at Theauthorsaregrateful thesupport theProgram Urbanization Migration Princeton for of and to the Centerfor Migration and Universityfor a researchfellowship to Tangchonlatip to couldnot have summerresearch Garipand Chung.Thisresearch support Curran, Developmentfor withoutthedata collection teamfrom the Carolina beencompleted efforts the research of Population at Centerat the University North Carolinaand the Institute Populationand SocialResearch of for and the cooperation the villagersof Nang Rongdistrict,Buriramprovince, Mahidol University of Thailand. We are extremelygrateful to all of them for their work and contribution.German all statisticaladviceregarding of our statisticalevaluations,modeling Rodriguez providedexcellent MelanieAdamsprovidedinvaluableeditorialassistance and of approaches, interpretations effects. Directcorrespondence SaraCurran, to and technical of supportin thepreparation this manuscript. WA Thomson Schoolof International Seattle, 98195Studies, Hall, University Washington, Jackson of 3650. E-mail: scurran@u.washington.edu.
? The Universityof NorthCarolinaPress Social Forces, Volume84, Number 1, September 2005

226 * SocialForces Volume84,Number1 * September 2005

et and Massey 1999; Massey, Goldring, al. 1994; Massey and Zenteno 1999; Espinoza Massey and Espinoza 1997; Winters,de Janvry,et al. 2001). Because of the data requirements causation,extensionsto other necessaryto test the theoryof cumulative case settingsoutsideof the Mexico-U.S. havenot been undertaken. We examinelongitudinal fromThailand replicate data to studiesof cumulative causation thathavepreviously beentested inthe Mexican-U.S only setting.Thedatawe use is the only datacurrently available affordsan extensionoutsideof the Mexican-U.S. that setting.Given the profundity the migration of momentum described the Mexican-U.S. we expectto case, by findsimilar resultsinthe Thai such settingandtherefore arguethatdemonstrating effects in Thailandwould portend heightened concern and need for understanding migration in momentum othersettingsaround world.We extendourtheoretical the of understanding the concept by examininghow migrantnetworksconvey differentlevels of trust and information when migrantsmake frequenttrips or have extended lengths of stay in a how networks based in households differ fromvillage-based destination; migrant networks; of networks further elaborates and,how sex composition migrant socialcapital and migrant its effect upon futuremigration. we Further, demonstratehow migration processes are contribution the to differently experienced menandwomen.Our by studyoffersa significant literature the causes of migration, on and because,to date, no studies havesystematically evaluated of and simultaneously trips,duration stay,levelsof aggregatemigrant experience sex composition the propensity migration on of amongmen andwomen. Ourstudyis different fromprevious studiesevaluating effect of cumulative the causation, notonlybecauseof the settingandourelaboration migrant of socialcapital, also because but ouranalysisconcernsinternal Previous workhas migration, mostlyrural-urban migration. will causation be less meaningful internal for suggestedthatcumulative migration processes recentanalysessuggests thatgendermayexplain differences the in 1986),although (Taylor effects of migrant networks internal international for and and (Curran Rivero 2003). migration There is also growingevidence that cumulativecausationis particularly for important out communities smallcities, but notfor migration of and out explaining migration of rural causationis large urbanareas (Fusselland Massey 2004). We also find that cumulative for rural-urban in and context,netof otherfactors, that important explaining migration the Thai measuresthat capturegender relations incorporating significantly improvesour models. we information opportunities and for Specifically, findthatmenandwomenconveydifferent theirmigration behavior experiences,andthatmen andwomen are and migration through affectedbythese effects amongmigrant socialcapital. risksof migration The for differently women are sufficiently social capitalis most usefulwhen it is high,andthereforemigrant and socialcapital most influential is men,migrant highly trustworthy basedon strongties. For when it is based in moreexpansive networks. the networks is upongenderas a critical Focusing aspect describing contentof migrant becauseof whatis assumedaboutthe way cumulative causation understood is to important affect individual to and rate networks propensities migrate the overall of migration. Migrant and accumulatedmigrant or experiencesdecrease the costs of migration, increasethe demonstrativeeffects (diminishing psychic costs or familialresistance),or increase information about and access to labormarket If opportunities. men and women livethe then the magnitudeof the decrease in costs and the migration experiencedifferently, increaseindemonstrative and effects, orthe amountof information access to labormarkets will varydependingon the gender content of the accumulated experiences.Inturn,this variation yielddistinctive compositiondistinctions migration sex in may outcomes, which for giventhe momentum causation, implied cumulative by mayyieldprofound implications the socialorganization lifein placesof origin destination. of and Inorigin this the communities, couldinfluence waycareis provided theelderly children, to or

on in TheEffects Gender Migration Thailand* 227 of

in inheritance of orthe redistributionwealth preferences awayfromdaughters the case (shifting for of of Thailand, example) investments or (abandonment upland plots agricultural thatrequire If behavesignificantly with male differently regards significant labor). maleandfemalemigrants thena gender withplacesof origin remittances visits), or to maintaining contact (either through in rates difference migration andaccumulated experience alsoaffectthe amount migration may back of andlevelof resource flowsdirected migrants to their by villages origin. destinationcommunities,if they are defined significantly the sex by Alternatively, composition of migrantpatterns, may also be transformed.For example, if males then of predominate, the organization domesticworkmaybe affected.Insome cases, men take on householdwork (Hondagneu-Sotelo 1994) or may outsourcethe work to may if to and entrepreneurs, they can afford do so. Or,if females predominate establisha sector nicheforthemselves,they mayaffecta redefinition women'sandmen's andoccupational of in work.Forexample,inthe case of HongKong,Filipina predominance domesticworkhas meant that the Filipino male migrantsfollowingtheir female relativesare more likely channeled domesticservice,an occupation wouldotherwise havejoined,either into not they inothermigrant destinations inthe Philippines or we 1997;Tyner (Constable 1996).Although do not quantitatively evaluatethese particular in we possibilities this paper, offerthem as for to of of justification why it is important examinethe influence sex composition migrant behavior. networks uponmigrant We examine 10-yearretrospective data longitudinal from22 villagesin one districtin Thailand evaluate roleof genderandmigrant to the socialcapital influencing Northeastern for the probability beinga migrant anypointintime during 10 years,net of alternative of at the We as of explanations. do so bydistinguishing experience number tripsandduration migrant of stayandby distinguishing socialdistanceof the migrant the network household-based as or village-based. evaluatethese fourdistinctmeasuresof migrant We social capitalupon men'sandwomen'smigration the Then,we evaluate effectof sex composition propensities. of experienceand social distanceupon men'sand women's migration And, propensities. we finally, evaluatewhetherthe sex compositionof migrant experiencesand networks matters moreorless forwomen'sandmen'smigration of the district. ourdiscussion out In of some insightsfromourqualitative fieldwork aboutthe possibleeffects results,we provide of genderedmigration of patternsuponthe reorganization social life in origin villages.We thatgenderrelations affectthe migration becausegenderinfluences process,inpart, suggest and socialcapital, measuredbytripsand the information trustavailable as through migrant levels- household village. and and experience as observedat different

Background Toourknowledge thereareveryfew studiesquantitatively how evaluating genderrelations might impact cumulativemigrationexperiences, but numerous review articles and havepointed the importance considering to research of when ethnographic genderrelations and 1992;Grieco Boyd1990;Hondagneu-Sotelo 1994; processes(Chant migration studying and Pessar1999a;Pessar1999b;Tienda Booth1991).Beforediscussingstudiesof gender, and we describethe conceptof cumulative causation causation, briefly migration cumulative andits effect uponmigration behavior. in or to are understood the Socialnetworks, socialcapital, relation migration commonly as in links of and betweenresidents a community origin individuals inanother place,orwith living who of individuals previously residence migrated regardless theircurrent (Hugo1991;Massey andGarcia-Espana Massey1990). in Thesetiesto migrants incorporatedmosttheories are 1987; world duallabor international usedto explain (Portes 1978), systemstheory migration including

228 * SocialForces Volume84,Number1 * September 2005

market socialcapital and and 1979), (Piore theory (Massey1990) new economicsmodels(Stark Levhari for has 1982;Stark 1991;Taylor 1986).Theconceptof networks explaining migration and most networks as support. strongtheoretical empirical Typically, studiesmeasure migrant froma commonsocialunit, either a simplecountsof otherpeoplewho havealready migrated or The to tie socialunit. Whether or village family. network is presumed be basedon a common in notprior of with of actually participatea network relationships the members thesocial migrants unitof originis usually quantified, not there is plentyof ethnographic evidenceto although in of do Therefore, suggestthatonthewholea vastmajority migrants participatethese networks. we use the termsmigrant network migrant and the socialcapital interchangeably, reflecting and current literature the presumption about meaning cumulative the of experiences. migration in Socialnetworks facilitate aboutrelative migration several ways.Theyconveyinformation on livelihood opportunities (Hugo1991),reducetravelcosts throughinformation safe and costs by lessening"assimilation shock" emotional cheap routesorsmugglers,anddiminish if immigrants in arrive an environment whereothersspeaktheirlanguage 1973)and (Choldin can whereliving deportation (Massey1990). amongotherforeigners easilyprevent Theyalso increase expectedbenefitsof migration through searchassistanceandreduceinitial job living assistance and shared livingspace. Migrant expenses and othercosts throughfinancial for heralded offering as trustedand reliable information prospective networks frequently are The networkfactor continues to be one of the most important contributing migrants. causedmigration, otherfactorscanalsocontribute' elementstowards cumulatively although evidence supportingthe importanceof migrantnetworksfor predicting Empirical a are and and migration for influencing dynamicrateof migration numerous mostlyfocus (Davis,Stecklov,et al. 2002; Espinozaand Massey 1999; upon Mexican-U.S. migration et Massey, Goldring, al. 1994; Massey and Espinoza1997; Massey and Zenteno 1999; evidencefromnumerous the de et Winters, Janvry, al. 2001).Qualitative settingsthroughout worldfurther supportthese findings.(See, for example:Arizpe1975;Arizpe1980;Arizpe 1995;Menjivar 2000;Tyner 1981;Arizpe 1985;Menjivar 1996). A few studies distinguish contentof migrant the networks dependingon the frequency of of tripsand the duration stay in places of destination (Masseyand Zenteno1999).The a of and number preceding tripsmaydemonstrate strongtie to a placeof origin the face-toface conveyance of information may enhance the degree of trust and accountability of associatedwithmigrant socialcapital thistype.Even of thoughthe number preceding trips may indicategreaterties to an originhouseholdor village,it may not convey highquality or information. the information enoughrelevant Instead, lengthof time migrants spend ina of of increasethe quality information the array resourcesavailable and to destination may The net of flow of both potentialmigrants, of the number tripsmade by migrants. return will of of amountandquality information dependon the nature the ties betweenthe migrant or ina placeof destination the migrant's and homevillage community. None of the quantitative between familystudies, however,systematically distinguish networks community-based and networks. There some suggestion is based migrant migrant in inthe qualitative studiesthattherearedifferences the kindsof information available from of members.Notonly might familymembersversusthose froma widerarray community information vary,but also the qualityof the social capitalwill be different.Trustmay socialcapitalquality betweenfamily community and but networks, community distinguish amounts a widerarray information Granovetter's and networks yieldgreater of weak may (e.g. ties theory (1983)).We distinguishmigrantsocial capital along these two levels of networks shorter socialdistances suggestingthathousehold-based aggregation, represent thanvillage-based networks. migrant Inmanydifferent contextsgenderhas been shownto differentiate the contentand both the character migrant linking of ties originand destinationhouseholdsand communities

* on in TheEffects Gender Migration Thailand 229 of and and 2003).Insome cases, femalemigrants (Curran Saguy2001; Curran Rivero maybe and morelikely maintain to places of origin(Curran Saguy2001)and in othercases to ties and To menmaybe morelikely do so (Grasmuck Pessar1991). date,we knowof no studies to that have systematically disaggregatedby sex the effects of trips and experienceupon We contextsgendermayserveas a proxy propensities. suggest thatin particular migration for measuring qualityand quantity migrant of social capitalor migrant the networks.Of ties importance be the way inwhichgenderdifferentiates betweenmigrants particular may and places of originand the trustand accountability associatedwith a particular formof In socialcapital. Thailand, we describeinthe nextsection,menandwomen have as migrant very differentsocial ties to theirvillagesand natalhouseholds.These relationships may affectthe contentof migrant or of networks the quality migrant socialcapital systematically andthereby influence behavior. differently migrant betweenhow migrant socialcapital Similarly, systematic analysesof the differences might affect men's and women's migration behavior have been limited.Numerous differentially studies havesuggested thatthe factorsaffectingmen'sandwomen'smigration differand If thatmenandwomenlivethe migration is processdifferently.women'smigration perceived as fraught morerisks, with thenseveraleffects areanticipated. family networks First, migrant than villagenetworksfor ensuringhigh qualityinformation and may be more important the associatedwithmigration. accumulated Second,womenneeda greater diminishing risks social capitalat the villageor community overcomebarriers migration. women's to to So, increasesafterthe migration processhas developedsignificantly migration (Boyd1989).Or, women'smigrant networks and different information maybe verydifferent offersignificantly bothmen'sandwomen'spropensities migrate to thereby differentially affecting (Hondagneuand Sotelo 1994;Curran Rivero about 2003;Tyner 1996).Despitefairly strongexpectations few studies haveexamined whethermigrant socialcapitalmatters differences, quantitative moreor less to men orwomenandallof these studieshavefocused uponthe Mexican-U.S. case. migration and finds that in a (2000) analyzeswomen's and men's first migration Kanaiaupuni constrained model,familynetworksmattermorefor women'sfirstmove thanfor men's. than Further, villagenetworks appearto havea strongereffect uponmen'smigration upon women's. And, finally,sex composition(a greater proportion women in a network) of positivelyaffects women's first move, but discouragesmen's. Withtime, however,the for effectof sex composition a disappears women,although significant positive discouraging effect remains men (Kanaiaupuni for and 2000).Cerrutti Massey(2001)findthat migration has but prevalencein the community littleeffect on husbands'or wives' migration, does influence eldestson'smigration notthe eldestdaughter's). an moves (and Family significantly for than aremoreimportant predicting Sons'and moves,ingeneral. migration arecommunity of moves raisethe probability bothmigrating, husbands significantly and are prior daughters' moreinfluenced theirchildren's movesthanarewives.Cerrutti Massey(2001) find and also by the of and thata mother's increases probability botha daughter son migrating, prior migration A effect upona daughter's buthas a greater migration. father'sprior migration significantly affectsa son's migration, nota daughter's. but and find Curran Rivero internal of (2003) thatthe prior migration womenfroma household of facilitatesthe migration both men and women, but international of migration female of the householdmembersonlyfacilitates migration otherwomen, not men. Men'sinternal has to destinations uponeithermen'sorwomen'smigration internal migration no influence buthas a significantly influence than uponmen'sinternational probabilities greater migration and 2003).Theresultsfromthis studysuggest thatthe effect uponwomen's(Curran Rivero of of gendercomposition prior household and uponinternational internal migrants migration barriers migration international for and can be explained bothbythe genderdifferentiated to

230 * SocialForces Volume84,Number1 * September 2005

internal as migration, well as the sex-segregatedlabormarket experienceandthe resulting to networks acrossthe two types of destinations. exposure sex-delimited migrant Inallof the abovecases, the empirical evidenceshows a strongly genderedstoryabout the migrationprocess, which is supported by numerous ethnographicaccounts. datacomes closest to ouranalysisof Thai of databy (2000)analysis Mexican Kanaiaupuni's and networks. extendthese analysesof the We takingintoaccountbothfamily community how cumulative at householdand process by analyzing migration migration the individual, we levelis differentially the for important menandwomen.Further, examine extentto village whichthe genderedcontentof family villagetripsandaccumulated and experience,net of as overallmigration prevalence,mightaffect any migration well as men's and women's we betweenthe number migrant and of And,finally, show howdistinguishing trips migration. months of migrant experiencerevealsmore insightsabout the gender dynamicsof the process. migration is similar data,we findsurprisingly Despitethe factthatouranalysis conducted usingThai to Our also on results some of the studiesof Mexican migration. findings offerfurther insights howdistinguishing betweendifferent of migrant socialcapital illuminates dynamics the types case provides important an to of the migration process.TheThai opportunity test a theoryin a different The data,setting and genderdynamics. settingwhere thereare verydifferent and refinement theoriesabouthowmigrant for networks analysis yieldgeneralizations further withregard how gendermightinfluence to influence outcomes,especially migration migrant and socialcapital different at levelsof observation (household village). Thailand: Gender, Migration and Distinctions Across Migrant Social Capital fromrural Thailand urbanBangkok to as migration may not seem as dramatic a Although Thaisduring mid-1980sit the move fromrural Mexicoto the UnitedStates,for manyrural was. (Thedataforthis studycoversthe period1984-1994.) the Although dangersof the trip are not as dramaticor deadly as those for Mexicanscrossing the borderillegally, the unsafeworkenvironments poorliving and experienceinthe placeof destination regarding who was our Thai told quarters similar. During fieldwork,rural villagers talesof failedmigrants in hadreturned homeandtoldof not beingpaidforweeks of work,of living 6-by-6 cubicles, of beingphysically and and Even mistreated, of losingalltheirmoneyto drinking gambling. for and moredevastating parentsarethe children who migrate areneverheardfromagain. our we of storiesof others.Beingableto During fieldwork learned one such case and heard trustsomeone anddistinguishing betweengood information aboutjobs fromacquaintances versusclose friendsandfamily in were critical determining whethera migrant wouldbe trip networks were particularly for worthwhile. Trustworthy important youngwomen and their whenconsidering for or families Preference kin-based immediate networks family migration. was much more important when decidingwhetherto allowwomen to migrate.Fearsof in were of utmostconcern.We observedduring exploitation, terms of eithersex or labor, severalepisodesof fieldwork youngwomenwouldoftentravel groupsto construction that in sites orto applyforfactory jobs.2 in Seasonalmigration notunheard inThailand, is of (the especially the Northeastern region site forthese data)as the monsoonrainsareoftenprecededby lengthy that droughts often the at livelihoods, least for a few months required movementof peopleto findalternative in took 1999;Phongpaichit 1990).However, (Nartsupha migration on addedsignificance Thai livelihoods fromthe mid1980sonward. was thenthatThailand's froman agricultureIt shift based export economy to a manufacture-based export economy took place (Bello, et and and 1980;Phongpaichit Baker 1996;Phongpaichit Cunningham, al. 1998;Phongpaichit

of on in TheEffects Gender Migration Thailand? 231

the to and Baker 1998;Warr 1993;Warr Nidhiprabha 1996).From mid-1980s the mid-1990s some expertsestimatethatThailand's economygrewon average10 percentperyear(Bello, and Thisgrowth fueledby production was et 1996). Cunningham, al. 1998;Warr Nidhiprabha in exportmanufacturing, whichwas a resultof the risingvalueof the Yen,risingwages in countries (NICs), quotasto the United nearby newlyindustrialized changes intextileimport direct fromJapan increasesinforeign investment, States,anddramatic primarily (Nidhiprabha and 1994;Phongpaichit Baker 1998).By 1985Thaimanufacturing exportshadoutpacedrice andotheragricultural and 1994; 1996).With exportsinvalue(Nidhiprabha Warr Nidhiprabha the growthin manufacturing exportcame an increaseddemandfor labor.Rural migrants muchof this labor, part many provided comingmostlyfromthe Northeastern of the country, of themyoung,andmanyof themwomen(Chamratrithirong, et Archavanitkul,al. 1995;Mills and to most 1997;Phongpaichit Baker 1996;). Survey According the 1992National Migration to area migrants the Bangkok metropolitan were intheirteens orearlytwentiesandat least halfof these migrants were women (Chamratrithirong, et Archavanitkul, al. 1995).These women were not movingto urban areasfor marriage. Theirs were economically motivated to of origin's their household moves,primarily helpsupport family economy(DeJong,Richter, etal. 1996). Thai ruralwomen have long played an importantrole in household economies and 1992).Theyworknext to theirhusbandsand (Singhanetra-Renard Prabhudhanitisarn in brothers the ricefieldsandareoftendescribed "holding pursestrings" regards as the with to financial Also manyhistoric ethnographic and studies 1992). planning (Yoddumnern-Attig describewomen'srelations theirhusbands egalitarian with as and (Knodel, Chamratrithirong, in streams is considerable, 1987).Women'sparticipation rural-urban Debavalya migrant as Archavanitkul al. 1995; et reaching highas 60 percentof all migrants (Chamratrithirong, Tantiwiramanond rates of 1995).These rates are only surpassedin Asia by the migration women fromthe Philippines Japan(Tantiwiramanond Itis important notethat and to 1995). these moves are rarely associational for moves) but primarily jobs for the women (family themselves(Chamratrithirong 1995). et al. of These positive characterizationswomen'sstatusareincreasingly questioned recent by in These researchersnote that women predominate the low wage, low skill scholarship. low sectorsof the economy, including wage servicejobs,prostitution, agricultural labor, wage andlow skillmanufacturing textiles,partsassemblyforelectronics, food processing and (like women consistently earn 1993;Tantiwiramanond 1995).Importantly, plants)(Sussangkarn to as one-third one-half muchas men insimilar and 1993; (Phananiramai Richter occupations Havanon of muchof the 1994;Tantiwiramanond Outside seasonalconstruction labor, 1995). labor market sex segregated. is Mentendto workinheavier taxi destination industries, driving andmotorcycle and services,automobile 1993).Hence, servicing, construction (Sussangkarn of the kinds information men for to migrant can provide womenmaynotbe as helpful women as the information migrant from womenandviceversa. of and ties Further, patterns marriage settlement adjudicate againststronghousehold with in sons andtendtowards Thailand (Blancstrongties withdaughters, especially Northeastern Szanton1990; De Jong, Richter, al. 1996;Yoddumnern-Attig Thaihouseholdsare et 1992). matrilocal. is, a husbandmoves to the wife'sfamily'shouse fortwo daysto fiveyears That the (usuallyuntilthe first childis a year old). Inthe ideal situation, husbandprovidesan normis bilateral, sons usually sourceof farmlabor. inheritance The but abdicate important to becausetheywillbe moving their to theirlandinheritance theirsisters or brothers-in-law wife's householdand receive,instead,some otherformof inheritance eithermoney,cattle or education et Patanothai, al. 1982).Youngest (morerecently) (Keyes1984; Limpinuntana, with regardsto householdresourcesbecause they advantaged daughtersare particularly the often inherit homesteadand a largerportionof the land.Access to these resources

232 * SocialForces Volume84,Number1 * September 2005

comes at a price. Thesedaughters theirhusbands expectedto careforthe daughter's and are 1992). parentsintheirold age (Keyes1984;Yoddumnern-Attig Onthe otherhand,sons are based networks, daughtersare socializedto and encouragedto developoutside, non-kin cultivatetheir kinshipties. For men, this cultural emphasis has lead to an interesting et of relations crosscutkinship (Limpinuntana, that ties Patanothai, phenomenon patron-client al. 1982; Hanksand Hanks1963).The non-kin based network men providesaccess to for political patronage, jobs,andresources. Thesegenderandfamily relations suggest different also ways inwhichthe socialcapital Different access to different of migrant networksmightinfluencemigration probabilities. to types of labormarkets maybe moreor less beneficial membersof the oppositesex. For advertised femaleexportmanufacturing jobs example,migrant tripsby women for clearly if networks maybe less influential men'smigration upon probabilities, especially recruitment we oftenused fieldwork foundthatfactory arehighly specific.During sex previous managers is and back This their workers recruiters intheirhomevillages. strategy notunusual as factory of can be productive, a forcethatis controlled social relations obligation by providing labor are for theirkinand and minimizing expenses since recruiters responsible training training friends. Wolf(1994)also describesthistype of management strategyin herstudyof women laborersin Indonesia factories.Besides the sex segregated exportmanufacturing work, in womenarealso employed domesticandothertypes of services,whichareeven moresex segregated. Working against this sex-segregated effect of migrantsocial capital,might be the and strengthof ties between destination villageof originandthe lengthof time a migrant Timein a place of destinationserves to broadenthe localties a spends in a destination. has and otherjobs. Withtime a migrant maybe migrant to information contactsregarding as able to help the next migrantfrom their village regardlessof their sex. This just ties information only be useful to potentialmigrantsif currentmigrantsmaintain to may or places of origin,eitherthroughletters,messages sent viaotherreturn migrants, return visits. Ifwomen are morelikely maintain to theirnatalvillagesbecause of cultural to ties access to theirfutureinheritance), the then reasons(maintaining expectationsor material theirexperiencein places of destination greater amountof information the the they longer may have and the more helpfulthey may be to potentialmen or women interestedin Further, given householdmembers'expectationsaboutwomen'scommitment migration. to otherfamily members' well being,the ties generatedbyfemalemigrants be imbued may with greatertrust, especiallyties based in the household.We share a few examples of sons and these sex-differentiated expectationsbetween men and women, particularly as we heardthem during fieldwork: our daughters, "Ifgirlsgo to workin Bangkok make5,000 baht3 and they will send 5,000 baht.The boys would not send us any money! Boys would not even send moneyto theirmothers.Theydo not make enoughforthemselves.We cannotdependon sons."(FocusGroup, Women40-49yearsold). "Ithink girlswillsend moremoneybecause boys use moneyfor and cigarettes, whiskey, baitiaow4. Boysuse a lotof money.Ifthere is anymoneyleftover, Girls havechancesto tiaow theygo outagain. butfewerthanboys.Daughters to think have moreabouttheir future at home because they depend on our inheritance." (FocusGroup, Women40-55yearsold).

TheEffects Gender Migration Thailand* 233 of on in

"Yes,we send money home very often. Whenthe end of the monthcomes, we save and send it all home. We do not haveany We We personalresponsibilities. haveto takecareof the family. all haveto help.Theyoungerones are stilllittle,and we do not want them to haveit tough.We wouldlikethem to study.Iwouldliketo with migrant improveour financialstatus at home." (Interview woman,22 yearsold). Men'sextended time in places of destinationmay be less useful to householdsand contact with householdsand villages of origin,because they are less likelyto maintain thereis a lowerexpectation they do so, andthey are perceived be that to villagesof origin, less vested in the householdand villageeconomies. Some men overcomethese social and commitment householdsof originsis through to expectations one way to demonstrate In our several frequentvisits home, as well as remittances. severalcases during fieldwork theirsons as "good" sons becausethey remitted theirearnings fromtheir parentsdescribed workandtheyvisitedfrequently. Pramualrathana makesa similar observation (1991) migrant in his studyof the elderlyin central when he drawsa correlation betweena son's Thailand, statusof "goodness" having and characteristics. these relationships are daughter-like Again, for at levelthanat the expectedto be moremeaningful the odds of migrating the household villagelevel. Migrantsocial capital, as described in the literature,is a combinationof both informationand trust, decreasing the risks and costs of migration.Thai villagers' discussionof migration the differential and of also patterns migration suggest thatmigrant social networksconvey information are evaluatedby potentialmigrantsand their and families in terms of theirtrustworthiness. Thus, based on our fieldworkand common sense, we suspect that household migrantsocial capitaland village social capitalare distinctfromeach otherin two ways. Theformeris morelikely be imbuedwith greater to trust and expectations about reciprocity. former is therefore more important The for risk on diminishing and the perceptionsof risk.The latter, the other hand,is a reflection of informationand more informationis availablewithin a village migrantnetwork. that information much more likelyto be transmitted return is via However, tripsand the Timespent awayby migrant personalconveyanceof news andjob opportunities. villagers is likely diminish conveyanceof that information. to the We thereforepropose three sets of hypotheses, based on previousresearchabout and as migration, genderand migration, genderinThailand, well as based on ourown field workexperiencesinThailand: 1. Migrant of different about tripsand duration stay offerdistinctly types of information withtrips socialtiesandtrustworthinessinformation duration of and migration, reinforcing of staywidening deepeningknowledgeabouta destination, weakening ties and but the to origin households communities. and 2. Thesourceof migrant socialcapital its significantly distinguishes relevanceandinfluence outcomes. uponmigration the a. Thenarrower socialdistance, greater perceptions trustworthiness the the of about but aboutmigrant migration opportunities, the less richand diversethe information possibilities. Hence, household-basedmigrant networks will proffer more but than networks. information, possiblyless information village-based trustworthy household-based networkswill be more influential Therefore, upon migration outcomesthanvillage-based for aversepotential networks, especially risk migrants.

2005 Volume84,Number1 * September 234 * SocialForces

within householdssignificantly and b. Sex segregatedlabormarkets genderrelations of and the expectations maleandfemalemigrant distinguish information trustworthy networks sourcesof influence as process. uponthe migration socialdistance and as socialcapital distinguished experience(trips duration), 3. Migrant by will and and (household-based village-based), sex composition affectmenandwomenin sets different ways,because men and women face different of constraints significantly to and risks withregards migration. Study Site and Data in in Thedataforthisstudycomesfromone district Northeastern Thailand, Rong, Buriram Nang land and arable forfuture Becauseof poverty, highfertility limited development, past province. centersinThailand, to sourceof migrants urban has the region becomean important primarily of region alonga major portion the Northeast NangRongis locatedinthe southern Bangkok. border. with the built States,linking Bangkok the Laotian highway during 1970sbythe United conduitof people and resourcesto and fromcentral Thishighway serves as an important within Thailand and location 1 a Thailand. Figure displays mapof the studysite andits relative in with districts the country was one of the largest In of the province Buriram. 1990,NangRong an of a population over200,000within areaof about240,000hectares. effortconducted the Carolina data are TheNangRongSurveys a longitudinal collection by and for and of Centerat the University NorthCarolina the Institute Population Population in We at SocialResearch Mahidol University Thailand.5 employthe firsttwo waves of datafor this analysis(the 1984 and 1994 surveys).The 1984 data collectionwas a census of 50 on data, householdassets and demographic villagesand includedinformation individual the not The characteristics. 1994datacollection onlyreplicated 1984survey, including village aboutformer1984 villagemembers,but also a census of all householdsand information as abouteducation, life included 10-year a work,andmigration, well as retrospective history births(askedonly of women),and events such as marriage, key social and demographic about information serviceorthe Buddhist entranceintomilitary Sangha(askedonlymen)6, and andtheircurrent residence, a specialsurveyof migrants. siblings as and from We employthe information the lifehistory follow-up, well survey the migrant in was The from as information the 1984survey. migrant follow-up component conducted 22 of as and of the original 1984villages counteda migrant someonewhowas a member a 1984 for and household hadsince lefta village morethantwo monthsto one of fourdestinations: and Korat Nakhon or the the provincial Ratchasima; Buriram; regional Bangkok capital, capital, Seaboard the Bangkok Chonburi, Area;and,Eastern (Chachoensao, provinces Metropolitan In of 2 of andRayong). project Figure shows the location the fourdestinations interest. related it has been documented how successful the surveys were at following manuscripts of et this and households individuals Kaneda, al.2002).For kind migrant (Rindfuss, follow-up, et is the success at finding Kaneda, al.2002). remarkably (Rindfuss, high migrants considered were successfully about44 percentof the migrants for Onaverage, the twenty-two villages, the interviewed some pointinthe six monthsfollowing 1994village at surveys. Inouranalysiswe builda datafile thatstartswith 1984 householdmembersthatare8and 25 years old fromthe twenty-twomigrant follow-upvillages7 are matchedwith data to fromthe 1994 surveys.We use the life historyinformation constructa person-agefile that thatbeginswiththose individuals are 13-25yearsold in 1984andthen add personsto our dataset, as they become 13 years old. We chose 13 years old as the lower bound because it marksthe end of primary schoolingand the beginningof exposureto the risk

The Effects of Gender on Migration in Thailand * 235 Figure 1. Map of Study Site

ProvincialMap of Thailand
0 150
Kilometers

300

Myanm.ar

Population size Laos ...


Chianp Mai.'

5,000,000and greater 100,000to 250,000 Provincial Provincial Boundary Boundary

50,000 to 100,000 Less than50,000

Gulf of Thailand

n Si Thamrnarat

2005 236 * SocialForces Volume84,Number1 * September Figure2. Mapof MigrantDestinationsin 1994for Migrantsfrom Nang Rong

Map of MigrantDestinations
0 'li---..... Ivlyanmar -,
'-i-. iI .. X

250 Kilometers

500

Provincial Capital Regional Capital

Bangkok Metropolitan Area Eastern Seaboard

U.S. Friendship Highway

i,

-'

...I
f v '.,A,9,,.,' .......... I

X.1" 7 (

e.

NakhonRanohasima
Bori Rarn 0
..>#-''

....
.

Area of detail Gulf of Thailand

Provinces in the Bangkok MetropolitanArea and EasternSeaboard

.t

~~~~~~~~~~~~~~~~~~~

n,a < a

Pi JmftH n.na

.I

(;,qf

rf 1 Th.

Tsering Creled'Lhv reangvoal i

* on in TheEffects Gender Migration Thailand 237 of adult.We conductedouranalyseson two sets of data,one of movingas an independent that for the entireset of 13 to 25 year olds and subsequent individuals age into the file overthe 10-yearperiod,and one for a subset of those - that are 13-15years old (to limit the historical of prior bias Here,we experienceamongthose olderindividuals). migration reportonlythe resultsfor the largerdataset, but the resultsfromthe subset of data are also available uponrequest.8 As has been done with the MMPdata to demonstratethe cumulativedynamicof we ratesfor each yearfor each prevalence migration migration, also estimatecommunity et prevalencerate measures the community(Massey, Goldring al. 1994). A migration of et up proportion peoplethat haveever migrated to a pointintime (Massey,Goldring, al. inthe NangRongvillagesgrows dramatically the 10-year over 1994).Migration prevalence of 3 the time period the study.Thetop imagein Figure displays overall prevalence migration for information collected.Inthe was rateforeach villageforallindividuals whom lifehistory across the 22 villages and that variation Thai case, there is considerablevariation is over maintained time.In1985,at the lowend, inone village only5 percenthadevermigrated inthe pastyearandat the higher 50 percenthadevermigrated the pastyear.By 1994 end in allvillageshadincreasedtheirprevalence rates,butthe wide rangebetween highand low villageswas stillapparent. The middleand bottomimages in Figure displayvillages'female and male migration 3 There similarly variation is wide acrossvillagesinbothimages, rates,respectively. prevalence femaleprevalence ratesaregenerally lowerthanmalemigration rates. prevalence although for in a particular The patterns maleandfemale migration do not always prevalence village mirror othereither. each how These descriptive demonstrate the cumulative data in does build dynamic migration case. These dataalso show thatthereis significant the Thaicase as it does inthe Mexican in over an variability migration prevalence timeandacrossthe sexes. Thisvariability provides to evaluatehow gender affects the dynamicbetween cumulative opportunity migration and experience individual propensities. migration Measures and Analytic Approach Ouranalyticapproachbuilds on a model developed in Massey and Zenteno (1999) to measure the dynamicsof mass migration. employ theirapproachfor a numberof We reasons. Ourdata are limitedin that we do not know the date of first migrationand some verylargeassumptionscannotpresumeto estimateit.We do have withoutmaking a population villagers8-25 years old in 1984 and observethem forwardin time, much of like Massey and Zenteno(1999)do with theircommunities.We buildon theirmodel by adding a set of baseline attributes of individualsthat might account for possible relatedto ourexplanatory factorsand the dependentvariable. unobservedheterogeneity whethera personin time t is livingoutside of NangRong We are interestedin predicting districtor not. Ourmodeltakes the followingformand takes into accountthe correlated errorstructureof multipleobservationsfrom individuals estimate a randomeffects (we logisticequation): = (1) Prob(Migit) f(ltripsit-1, Htrips Hmonths Ctripsit-1, it-1, Cmonthsit1, it-1, Imonthsit1i, Invariant ) Variantit, outsideof NangRongin yeart, Migitis where Prob(Migit) a personi'sprobability living is of 1 if personimovedout of NangRongat some pointduring yeart and 0 otherwise,Itripsit_l

238 * Social Forces Volume 84, Number 1 * September 2005

Figure3. Variationin MigrationPrevalenceRatesAcrossVillagesand Time in lowestandblack a codedto reflecta village's Eachshadeddot represents village ranking 1984(gray
and lowest values in 1994. highest). Linesare drawnfrom highest, mid-point,

Prevelance RatesbyVillageandYear Migration


90% 80%.
. -,* .

70%J
o 60%I

50%d I = 40% 4) ,
30%

-~

~ ~
3r

Ii
?

& I

U~~
3

20%-

'

10% I
0% I
9

2 198:

1984

1986

1988

1990

1992

1994

1996

Prevelance RatesbyVillageandYear Female Migration


100%4.

90% -

80%- > 70%60D0/0


-

50% 40% -

30% 20%-

r.
4 16 1 1 1 1

0%
01%

1982

1984

1986

1988

1990

1992

1994

1996

in TheEffects Gender Migration Thailand* 239 of on Figure3. (continued) in lowestandblack a Eachshadeddot represents villagecodedto reflecta village's ranking 1984(gray fromhighest, and are mid-point, lowestvaluesin 1994. highest).Lines drawn Prevelance RatesbyVillageandYear MaleMigration
100% 90%80%70% /

60%-./

;,

(s so%o 40% 30% ,?

x .

I
.
"

20% -

10%--

0% 1982

1984

1986

1988

1990

1992

1994

1996

is the number of trips made by personi up through year t-, Imonthsit-1 is the number of months experienced as a migrant by personi up through year t-1, Htripsit-1is the number of trips made by other people in personi's household up through year t-1, Hmonthsit-1is the number of months experienced as a migrant by other people in personi's household up through year t-1, Ctripsit-1is the number of trips made by other community members up through year t-, and Cmonthsit_1 is the months of experience accumulated by other community members through year t-1.'0 The community migrant trips and months of or migrantexperience do not include the experience of the observed individual the members of the observed individual'shousehold. We include as controls a vector of time-varying factors, including age, educational attainment, marital status, village electrification, and the migration prevalence rate. We also include as controls a vector of time invariantmeasures , including: sex, household land ownership in 1984 , whether the person was a temporary migrant in 1984,13 whether there were any temporary migrants in their household in 1984, the person's village's proportion of 13 to 30 year olds in 1984 that were temporary migrants, the female temporary migration rate, and whether the person lived in a somewhat remote village or a very remote village in 1984.'4 Inthis specification, the probabilityof livingoutside of Nang Rong depends not only on the age and sex of the individual,but also on a person's priormigratoryexperience (i.e., on his or her accumulated human capital) and on the degree to which he is surrounded by other villagers with migratory experience (the quantity of social capital). Our model is different from Massey and Zenteno's model in the following ways: first, we add a measure of household migrant trips and experience (because we suspect that the quality of informationavailable to potential migrants is different at the household level than at the

240 * SocialForces Volume84,Number1 * September 2005

of and factorsthatmight array timevarying invariant villagelevel);second, we add a larger due be relatedto the migration decision;third,we controlfor sampleattrition to the data over at the collectiondesign by measuring the village-level success in followingmigrants rate time; and fourth,we includea measureof the village'stemporary migration in 1984 (which measures temporarymigrationbetween 1983 and 1984 - see footnote 13) as and (1994). suggested by Massey,Goldring Durand in factors Table1 displaysthe meansfor allthe variables the model.Thetime invariant factorsdisplayedin the 1994 panel. In are shown in the 1984 paneland the time variant 1984 the men in the sample are slightlyyounger,probably reflectinglife course related migrationfor marriageby men. Older men will have migratedout of the village for of postnuptialresidence, yieldinga younger followinga tradition matrilineal marriage, in of sampleof men inthe village.A majority the sampleoriginates householdswithsome were not temporary land.A vast majority migrants (gone for 2 or moremonthsinthe year in preceding1984).Onthe otherhandalmosta thirdof the individuals the sample livedin at householdsthatdid havea temporary migrant some pointbetween 1983 and 1984. On in 46 averagethe individuals the samplecame fromvillageswhere approximately percent of the 13 to 30 year olds were temporarymigrantsin 1984 and about 35 percent of levels of migration women were migrants. therewere fairly So, highlevels of pre-existing In1984 most livedinveryremotevillagesandonlya third at the startof the 10-year period. In livedin a villagewithelectricity. 1984, most respondentshadonlya primary education, but some secondaryeducationwas not unusual. Althoughvery few had been temporarymigrantsbetween 1983 and 1984, by 1994 more than 60 percent had ever migratedout of the Nang Rong district.This rate was by higheramong men. Overhalfof women had married 1994, whereas only 36 percent to Withregards schooling,ten of men haddone so. By 1994,allthe villageshadelectricity. percentof women hadcompletedsecondaryschoolingin bothsamplesand 15 percentof men in the overall sample haddone so. whereasmen Onaveragewomen hadmade littlemorethanone tripout of the district, had made closer to 1.5 trips.On average,women had spent morethantwo years out of the districtoverthe past 10 years, and the men had spent almost three years outside of the district.On average,respondentsin oursample came from householdswhere other othermemberswere morelikely have to householdmembershadever migrated, although of by spent fewer monthsoutsideof NangRong(thismeasureis standardized the number to our people in the householdcontributing the measure).We standardized measureof to village trips and months by the numberof people in the village contributing the in measure.The averagenumberof tripsby individuals each individual's villageis about one anda third the averagenumber monthsoutsideof the villageis morethantwo and of shows thatthe contribution men of years.Thegenderedversionof each of these variables and women to each of these patternsis roughly equal. Our modelingapproachintroducesa set of baseline controls, the vectors of time social capital(not yet disaggregated factors, measures of migrant varyingand invariant by sex), andwe includea measureof yeareffects to controlfor changes in economicand We structural opportunities. evaluatethe modelon a pooled sampleand then separately for men andwomen (resultsare foundin Table2). We sex disaggregateour measuresof socialcapital. These last modelsevaluate importance the genderedcontent the of migrant of social capitalfor the pooled sample of men and women and for men and women we discuss the effects of the otherfactorsin ourmodels. (Table First, briefly 3). separately Thenwe move on to discuss the resultsevaluating effect of migrant the social capitalon in the probability migrating the Thaicontext. of

TheEffects Gender Migration Thailand* 241 of on in

1: Table Descriptive Statistics AllVariables 1984and1994 for in Individual LevelMeansin 1984

Women

Men

Age in Married 1984 1-10rai Land Owned Land 11-25rai Land rai >25 in Temporary Migrant 1984(0/1) HH Temporary Any Migrant (0/1) %13-30yroldTemporary in Migrants Village in in %Female Temporary Migrants Village Somewhat Remote (0/1) VeryRemote (0/1)

Electricity (0/1) NoEducation (0/1) Education Primary (0/1) Some Secondary (0/1) Schooling School Secondary Completed (0/1)
Out EverMigrated of NangRongby 1994

# Migration Among Indiv. Trips # Migrant Months Indiv. Among # Migrant forHH HH Trips per Member # Migrant Months HH HH for per Members # Migrant for PerPerson Trips Vill. Months Vill. Person for Per # Migrant # Migrant byWomen, Indiv. Trips Months Women, # Migrant Indiv. by in Per Women HH Person # Migrant for Trips in Per # Migrant Months Women HH Person for in Per # Migrant for Women Vill. Person Trips in Per Months Women Vill. Person for # Migrant # Migrant byMen. Indiv. Trips # Migrant Months Men, Indiv. by in Per # Migrant forMen HH Person Trips Months Men HH Person for in Per # Migrant in Per # Migrant forMen Vill. Person Trips Months Men Vill. Person for in Per # Migrant Age
Married 1994(0/1) by Rate Follow-up Village Migrant has Electricity Village (0/1) No Education (0/1) Education Primary (0/1) SomeSecondary Schooling (0/1) Secondary School(0/1) Comnleted - ,
.------. --

18.05 17.82 .22 .13 .13 .11 .19 .18 .27 .28 .08 .12 .31 .32 46.35 46.47 35.44 35.40 .19 .18 .65 .63 .32 .30 .00 .01 .68 .60 .25 .33 .03 .05 Individual Meansin 1994 .59 .72 1.14 1.55 27.59 35.03 .65 .66 16.14 15.92 1.34 1.34 31.33 31.13 1.14 27.59 .29 .30 7.67 7.76 .61 .61 14.82 14.88 1.55 mrn(i .36 .35 8.47 8.15 .72 .72 16.51 16.25 25.06 24.65 .55 .36 43.68 43.97 1.00 1.00 .02 .02 .71 .63 .18 .21 .14 .09

242 * SocialForces Volume84,Number1 * September 2005

Results 2 Table displays resultsof the modelsevaluated the pooledsampleinthe firstthree the on columnsand then the resultsfor the models evaluatedon separatesamples for men and women. Looking the resultsforthe control at variables evaluated the pooledsamplewe on show thatas a persongets a yearoldertheirodds of beinga migrant timet) increasebut (in thisrateof increase diminishes oldera persongets. For the whenan individual 15 is example, from are yearsoldthe predicted probability thismodelshow thatthe odds of beinga migrant 0.09. Whenthe same individual to 20 yearsold, the odds of beinga migrant .34. are ages is haveonly However, the time the individual 25 yearsold, the odds of beinga migrant by increasedto .35. Men have 2.4 times the odds of being migrantsthan women. Some also the relative primary to secondary schooling increases odds of beinga migrant schooling, but completingsecondary school does so dramatically almost three times). Not (by decreasesthe odds of beinga migrant surprisingly, marriage quitedramatically, almost65 by percent. Land is associatedwithbeinga migrant, the ownership not significantly although effect becomes significant Models2 and3. Individuals householdsthatare nearlandless in from havehigher odds of beingmigrants, relative landlesshouseholds. to (1-10rai) If individuals temporary are migrantsin 1984 then they have higherodds of being a in if householdhada temporary migrant anysubsequentyear.Similarly,theirorigin migrant in 1984theirodds of beinga migrant 50 percentgreater. are levelratesof temporary Village have in migration no effect on the odds of subsequentmigration model1, butinsubsequent models it has significant small effect uponthe odds of being a migrant. gender but The of rate composition the migration prevalence has no effectuponthe odds of beinga migrant. in or increases Living eithera somewhatremotevillage a veryremotevillage significantly the odds of beinga migrant 2.3 times.We suspectthatthis is an effect of not beingnear by locallabor market whichareavailable the district in town.Electrification no has opportunities effect on the odds of being a migrant. success of the 1994 field workin The significant fromeach villagealso has no effect on the odds of beinga migrant a at followingmigrants pointintime. Therearea few differences similarities and acrossthe control variables the men and for women.Theage effect is similar bothmen andwomen.Schooling, for is however, different for men and women. Havingsome secondaryschooling relativeto primary schooling increasesthe odds of men'smigration, onlycompleting but to secondary schoolingrelative is deterrent primary schoolingincreasesthe odds of women'smigration. Marriage a greater of migration women thanfor men, but it is a deterrent bothsexes. Land for for ownership, for the especiallynearlandlessnessapproachessignificance men, increasing odds of out but for migration, is not significant women. Prior temporary migration experiencebetween 1983and 1984dramatically increaseswomen'sodds of moving also increasesthe odds and of men's migration, muchless dramatically. presence of temporary but The in migrants a household only increases the odds of women's migration. the other hand, village On ratesincreasethe odds of men'smigration, notwomen's.Thesex but temporary migration of composition those rateshas no effecton eithermen'sorwomen'soddsof beinga migrant. Migrant Social Capital and Its Effects on Migration 2 Table also shows the resultsof the modelsthataccountforthe effects of migrant social net in capital of the factorsincluded the baselinemodelforthe pooledsampleandthe split to sample by sex. This model is a similar the model suggested by Massey and Zenteno

TheEffects Gender Migration Thailand* 243 of on in Table2: LogisticEstimationof the Odds of Being a Migrant(LivingOutsideof Nang Rong)

Pooled z OR

Women OR z

OR

Men z *** ** *** ** **


t

Cumulative Migration Experience # Migration Among Indiv. 1.48 11.15*** 1.80 9.74*** 1.27 5.39 Trips # Migrant Months 1.01 7.21*** 1.01 5.38 *** 1.00 2.95 Indiv. Among # Migrant forHHPerHH 1.40 4.51*** Trips Member 1.27 2.33 1.57 3.98 # Migrant Months HH HH for Per 1.01 3.00 ** Member 1.00 1.40 1.01 2.96 # Migrant forVill.PerPerson .57 -1.58 1.66 1.00 .17 -3.42 Trips # Migrant Months Vill. Person for Per .95 -4.71*** .94 -3.96*** .97 -2.38 Controls 2.71 22.09*** 2.25 12.64*** 3.39 18.81 Age .97 -22.94*** .98 -13.35*** .97 -19.16 Age*Age Men 1.59 4.89*** Some2ndary School 1.26 2.191: 1.22 1.34 1.46 2.69 School 2.28 8.71*** 1.88 4.79*** 2.66 7.07 Completed 2ndary Married .38 -12.23*** .23 -13.96*** .75 -2.31 Own1-10Raiof Land 1.57 3.30** 1.54 2.33 t 1.94 3.15 Own10-25Raiof Land 1.20 1.53 1.02 .14 1.45 2.17 : OwnMore 25 Rai than .90 -.86 .86 -.87 .95 -.26 4.47 7.47*** 6.36 4.87*** 2.34 3.10 Temporary Migrant inHousehold 1.40 2.95 ** 1.53 2.78 * 1.34 1.69 Temporary Migrants %13-30YearOldsthatareMigrants 1.01 3.22 ** 1.01 2.05 * 1.01 3.05 %13-30YearOldWomen 1.00 .85 1.01 3.10 ** .98 -2.16 Migrants Remote 1.40 2.38 : 1.63 2.31 t 1.20 .93 Village 1.84 3.34 ** 2.37 3.12 ** 1.57 1.69 VeryRemote Village Follow Rate 1.01 1.23 .99 -.79 1.03 2.50 Up Migrant has Electricity 1.09 1.10 .96 -.28 1.18 1.47 Village Prevalence Rate 1.04 6.75*** 1.04 4.00 *** 1.06 5.78 Migration Year Effects 1984 1.24 .60 1.51 .79 1.05 .11 1985 .61 -1.59 .59 -1.19 .62 -1.06 1986 .73 -1.07 .73 -.76 .74 -.72 1987 .67 -1.53 .73 -.86 .63 -1.22 1988 .75 -1.17 .82 -.59 .73 -.91 1999 .83 -.87 1.02 .09 .70 -1.15 1990 .78 -1.36 .74 -1.16 .85 -.63 1991 .88 -.76 1.10 .45 .74 -1.34 1992 1.03 .29 1.06 .35 1.03 .19 1993 1.42 3.38** 1.51 2.80 * 1.37 2.16 s.e. (u) 2.01 1.95 2.14 Rho .80 .79 .82 Wald 3135.40 ** 1694.80*** 1435.11 ** Chi-square
tp<.05 *p<.01 **p<.005 ***p<.001

*** *** * *** t **

** **
t

***

244 * SocialForces 2005 Volume84,Number1 * September

we for First, discussthe results the pooledsample.Wefindthateachadditional by (1999). trip Andwitheach additional an individual increasesthe odds of beinga migrant 48 percent. by more likely migrate to monthof experienceat the individual a personis significantly level odds increaseby about 1.5 percent). and monthsof experienceby other Trips again(the householdmembersalso have similareffects on the odds of migration. Tripsby other haveno significant effect uponthe odds of beinga migrant monthsof migrant and villagers has effecton the likelihood migration. of amongothervillagers actually a negative experience Theodds of beinga migrant reducedbyabout4 percentifallvillagemembersincreased are as theirmonthsof migrant experienceby one month.Experience measuredby tripsand and effects uponmigration duration to havedistinct sometimesdifferent outcomes. appears The effect of migrant social capitalis significant distinctfor household-based and and as cumulative networks, well. Household migrant experiencealso village-based migration increasesthe odds of being a migrant. However, experience villagecumulative migration but for The decreasesthe odds of migrating, onlysignificantly monthsof experience. greater then the socialdistanceat the village-level the less usefulthe migrant network increasing for the odds of migration. Beforespeculating we on to further, turn the analysis the splitsamples formenandwomen. Individual for migration experienceis important both men and women. The odds of are women beingmigrants 80 percenthigher witheach additional previous Theodds of trip. men being migrants increaseby 28 percentwith each additional prior trip.Trips other by household membersarealso important bothmenandwomen. Eachadditional taken for trip membersincreases odds of a manbeinga migrant 57 percent, the but byotherhousehold by the Months household of onlyincreases oddsfora womanby27 percent. experience migrant haveno significant effectuponthe odds of beinga migrant women,butdo havea positive for effectuponthe odds of menbeingmigrants. haveno significant effect by Trips othervillagers women'smigration direction the effect is positive), a strongnegativeeffect of but (the upon For of uponthe odds of beinga male migrant. both men andwomen the number migrant monthsexperienced othervillagers decreasesthe odds of migration. by significantly social capitalmeasuredby Again,the analysisof the split sample shows that migrant different and and levels(household village) and has experiences(trips duration) at different for men andwomen. Household different is significantly consequences migrant experience for the for important predicting odds of men being migrantsand not at all significant the Finally, predicting odds of women beingmigrants. tripshavepractically villagemigrant monthsof experience oppositeeffects for menandwomenandvillagemigrant significantly reducethe odds of beingmigrants bothmenandwomen. for the Theselastpuzzling effectsof village andhousehold results, regarding contradictory trips experiencesupon men'sand women's behavior, migrant suggest to us thatthere may be further information yet distilled not aboutthe influence migrant of socialcapital, particularly aboutthe gendered contentof migrant networks. beforediscussing preceding the results So, further turn the effectsof the gendered to we contentof migrant socialcapital. Engendering Migrant Social Capital Whenwe disaggregate migrant our socialcapital measuresintothose capturing men'sand women's several standout.Theseresults foundinTable are 3. experiences, interesting findings Table3 presents the resultsfor a full model, which disaggregatesthe measures of cumulative and migration experience sex and includesbaselinecontrols yeareffects.We by on presentresultsforthe modelevaluated a pooledsampleandon separate samplesof men and women. As with the results in Table2, the effects of an individual's priormigrant

TheEffects Gender Migration Thailand? 245 of on in Table3: EvaluatingGenderRelationsand MigrantNetworks(LogisticEstimationof the Odds of Beinga Migrant(LivingOutsideof Nang Rong)

OR Cumulative Migration Experience # Individual Trips # Individual Months Experience FemaleCumulative Migration Experience # Migrant for Trips Women HHPerMember in # Migrant Months for in Women HHPerMember # Migrant for Trips in Women Vill. Person Per # Migrant Months for in Women Vill. Person Per MaleCumulative Migration Experience # Migrant for Trips in Men HHPerMember # Migrant Months for MeninHHPerMember # Migrant for Trips in Men Vill. Person Per # Migrant Months for in Men Vill. Person Per Year Effects 1984 1985 #1986 1987 1988 1989 1990 1991 1992 1993 BaselineControls Also Included s.e. (u) Rho Wald Chi-square
tp<.05 *p<.01 **p<.005 ***p<.001

All Z

Women OR Z

Men OR Z

1.4810.96*** 1.80 9.66*** 1.28 5.35** 1.01 7.09*** 1.01 5.26*** 1.00 2.88**

1.28 2.31 : * 1.02 4.51 ** .41-2.13 : .99 -.15 :

1.21 1.23 1.01 2.28 t 1.90 1.08 .94 -2.84

1.28 1.52 1.03 4.18** .07 -4.35*** 1.06 2.90** 1.98 4.17** .99 -.23 3.01 1.33 .85 -5.87** 1.29 .49 .80 -.47 .96 -.09 .83 .96 .90 .95 .93 -.48 -.12 -.31 -.18 -.28

1.50 4.03*** 1.31 1.99 1.00 .13 1.92 1.12 .90-5.33*** 1.32 .73 .66-1.25 .81 -.71 .74-1.09 .84 .92 .86 .97 -.70 -.37 -.74 -.14 1.00 -.07 1.21 .23 .95 -1.71 1.38 .61 .54 -1.34 .68 -.91 .68 -1.01 .76 -.75 .97 -.09 .80 -.79 1.05 .25

1.03 .16 1.12 .85 1.48 3.73*** 1.48 2.62 * 2.03 .80 3139.08*** 1.96 .79 1699.10 ***

1.25 1.19 1.54 2.88** 2.15 .82 1438.14 ***

2005 Volume84,Number1 * September 246 * SocialForces

Whatis more of uponthe probability beinga migrant. experienceshavea stronginfluence is experienceare migration revealing thatthe effects of householdand villagecumulative when gender compositionis taken into account.We begin with a different significantly discussionof the effects forthe pooledsampleof menandwomenandwe thendiscussthe separateresultsforwomenandmen. The results for the pooled sample in Table3 show how priortrips and months of influence odds the and level experienceat the individual continueto positively significantly increases the odds of being a migrantby 48 An of being a migrant. additional trip prior monthof experienceincreasesthe odds of being a migrant by percentand an additional 1.5 percent.Tripsby both othermales and females in the householdhave a positiveand significanteffect upon the odds of being a migrant.An additionaltrip by other male householdmembersincreasesthe odds of migration 51 percentand an additional trip by by by otherfemale householdmembersincreasesthe odds of migration 29 percent.An monthof migrant additional experienceby otherfemale householdmembersincreases the odds of beinga migrant 2 percent,butmonthsof migrant experienceby othermale by trips by female migrantmembers has no significanteffect. At the villagelevel, migrant of reducethe odds of being a migrantby 59 percent,but duration villagerssignificantly On has no effect on the odds of beinga migrant. the otherhand,migrant trips experience effect uponthe odds of beinga migrant, by malevillagershas a positivebut insignificant reduces the of but duration migrant experienceamong other male villagerssignificantly monthreduces the odds of being a migrantby 9 odds of being a migrant additional (an socialcapitalhas a positiveimpact,whethermaleorfemale, migrant percent).Household socialcapital andvillagemigrant appearsto havemixedor negativeeffects uponthe odds of beinga migrant. the experiencesincrease migration Comparing resultsfor men and women, individual the odds of being a migrantfor both samples. At the household level the effects are to similar those forthe pooledsample.Migrant tripsby otherfemale householdmembers increase the odds of being a migrantfor both men and women, but not significantly. increases of Duration experienceamong otherfemale householdmemberssignificantly the odds of being a migrantfor both men and women. Formen the odds of being a monthof experienceby otherfemale are migrant 3.4 percenthigherwith each additional householdmembers,whereas it is only 1.6 percent higherfor women. Migrant tripsby for increasethe odds of beinga migrant men othermalehouseholdmemberssignificantly the the andwomen, increasing odds of beinga malemigrant 99 percentand increasing by odds of beinga female migrant 31 percent.Monthsof migrant experienceamongother by for malehouseholdmembershaveno significant impactuponthe odds of beinga migrant of the eithermen orwomen, although direction the effect is negative. A comparisonof villagelevel effects across the samples of men and women reveals startlingdifferences. Sex decomposition of migrantsocial capitalat the village level reveals some contraryeffects for men and women. Migrant trips by female villagers and increase the odds of being a migrantamong women (althoughnot significantly) An decrease the odds of being a migrant trip among men (significantly). additional by all femalevillagers reducesthe odds of beinga migrant 93 percentamong men. Duration by effects for men and women. An of experienceamong female villagersalso has contrary in additional monthof experienceamongallfemale migrants a villagereducesthe odds of amongwomen by 5 percent,but it increasesthe odds of beinga migrant beinga migrant haveno tripsby othermalevillagers amongmen by 7 percent.Onthe otherhand,migrant the for effect uponthe odds of beinga migrant eithermen orwomen (although significant directionof the effect is positive).Monthsof migrant experienceby othermale villagers for has a negative effect upon the odds of migration both men and women, but only

* TheEffects Gender Migration Thailand 247 of on in for monthof migrant experienceamongall malevillagers significantly men. Anadditional woulddecrease the odds of a manbeinga migrant 14 percent. by Discussion of Results: Gendered Nuances of Migrant Social Capital Inthis studywe extenda model of migration tested in onlyone otherlocation previously to We that (Mexico-U.S. migration) the case of Thailand. argued the processof out migration and from rural areas was similarly profound fraughtwith risksin the Thaicase and that socialcapital shouldalso havea meaningful influence We behavior. uponmigration migrant in also expectedthatthe verydifferent contextof genderrelations Thailand, relatively a high status of women withsome freedomto migrate, dramatic economicgrowthcreating jobs outsideof the localeconomyfor bothmen andwomen,andgenderedsocialties withnatal socialcapital withdifferent householdswouldcreateverydifferent effects types of migrant formen'sandwomen'sprobabilities beingmigrants. of a In We beganouranalysis replicatingmodeltested by MasseyandZenteno (1999). that by betweentripsandduration stayand of andZentenoarguefordistinguishing model,Massey of household we followtheirlead.Our extensions the modelincluded disaggregating migrant social capital,evaluating social capitalfromvillagemigrant separatemodels on men and socialcapital sex composition. motivation Our women,andfurther by disaggregating migrant which suggests conceptual for this approachwas based on a review of the literature of socialcapital, the of aboutthe character migrant distinctions including importance strong and the womenface when migrating the importance weakties for of ties forovercoming risks and contextwe proposed sourcesof information. Basedon the literature the Thai diversifying betweenfrequenttripsversus extended these two concepts by distinguishing measuring of and into migrant experiences bytaking accountthe sex composition each formof migrant our of data. We socialcapital. findthatthisdistinction improves understanding the Thai Table4 providesa summaryof our results organizedaccordingto our three sets of of of dimensions migrant experience hypotheses:1. tripsandduration stay measuredistinct 2. will information affectthe social capital; the source of migrant as conveyedvia migrant social capital,specifically if the source comes froma greatersocial contentof migrant (a.) less and it but distance(village-based)offersmoreinformation, is potentially trustworthy (b.) is information comes frommen orwomenthe relevant if the sourceof information migrant and3. women and huedby genderrelations sex segregatedmigrant experiences; distinctly in different the socialcapital distinctly andmen experience effects of migrant ways. Ingeneral,tripsand duration stay appearto measuredifferent of aspects of migration, effects in most models. Greater frequencyof tripshas independent displaying significant our the supporting hypothesesthattripsreinforce primarily positiveeffects uponmigration, socialcapital. Duration stayshows a moremixedset of results, of of trustworthiness migrant and its varyingacross social distance and sex composition.We discuss this variation shortly. implications socialcapital. information sourcealso imbuesmigrant Thesocialdistanceof the migrant networks Closerdistancesas measuredby household-based encourage generally migrant as sourcesof migrant Moresocially distant information, measured village-based by migration. via influence networks villagers' exceptinthe case of femalemigrant mostlydetersmigration, socialcapital The that monthsof experience migrant uponmalemigration. finding household of and effect uponthe likelihood beinga migrant thatvillage has a consistentpositive migrant of even be a hindrance socialcapital may suggests thatthe quality information be more may better or more helpfulthan that from village-basedmigrant trustworthy, significantly studiesof and This networks. is notsurprising has also beenfoundinstudiesof Mexican-U.S.

248 * SocialForces Volume84,Number1? September 2005 Table4: Summaryof Resultsfrom Table3 (not statisticallysignificantlog-odds coefficient)

Women 1: HYPOTHESIS as SocialCapital Trips Migrant # Migrant for Trips in Per Women Vill. Person # Migrant for Trips in Per Men Vill. Person # Migrant for Trips in Per Men HH Member # Migrant for Trips in Per Women HH Member as SocialCapital Months of Migrant Experience Months for # Migrant in Per Women HH Member # Migrant Months for in Per Women Vill. Person # Migrant Months for in Per Men HH Member # Migrant Months for in Per Men Vill. Person HYPOTHESIS socialdistance: 2aHouseholdLevel Social Migrant Capital # Migrant Months f in Per Men HH Member # Migrant for Trips in Per Men HH Member # Migrant Months for in Per Women HH Member # Migrant for Trips in Per Women HH Member Level SocialCapital Village Migrant # Migrant for Trips in Per Men Vill. Person # Migrant forfor Trips in Per Women Vill. Person for # Migrant Months for in Per Women Vill. Person # Migrant Months for for in Per Men Vill. Person HYPOTHESIS sex composition: 2bFemale SocialCapital Migrant # Migrant forfor Trips in Per Women HH Member # Migrant forfor Trips in Per Women Vill. Person Months for for # Migrant in Per Women HH Member for # Migrant Months for in Per Women Vill. Person Male SocialCapital Migrant # Migrant forfor Trips in Per Men HH Member # Migrant forfor Trips in Per Men Vill. Person # Migrant Months for for in Per Men HH Member # Migrant Months for for in PerPerson Men Vill.
*

Men (+) + (+) + + (-)

3 HYPOTHESIS *

(+) (+) + (+) + (0) (-)

(0) + + (+) (+) (+)

(-) + + (+) (+ +

(-)

(+) (+) +

(+) + +

+ (+) (0) (-)

+ (+) (-)

Significant Differences Between Men and Women

TheEffects Gender Migration Thailand* 249 of on in

and and 2000; (Cerrutti Massey2001; Kanaiaupuni Espinoza Massey1999;Massey migration and Espinoza ties household1997).Stronger andgreater degrees of trustmaycharacterize orfamily-based socialcapital the migrant significantly decreasing costs of migration. Whenthe source of migrant comes frommen the effects on migration information are different than if the source of information comes fromwomen. Estimating sex distinctly socialcapital demonstrates women offersignificant that information migrant disaggregated aboutmigration the via especially householdpossibilities longerthey stay in a destination, based migrant networks. Whereas longermen areawaythe less influence the they haveon On whenthe migrant network is tripsby men,especially migration. the otherhand,frequent reinforces and information socialties andconveysvaluable encouraging household-based, aboutmigration possibilities. and Men'sand women's migration propensitiesare somewhat dissimilar the factors socialcapital measuredby as those outcomesaredifferent, particularly migrant influencing networks. These differencesmostlyconformto ourexpectations abouthow village-based aboutthe migration For women'sexperiences different information convey experience. men, We frequentmigranttrips by women villagersactuallydeter migration. predictedthis if sex possibility, we assumed that the migration experiencein a destinationis relatively and durations less, trips segregatedas we knowit is in most cases. Frequent reflectshorter On outsideof a femalelabor market. information aboutmigration broad-based opportunities othervillagewomen the other hand,as we expected to find, frequentmigrant trips by influences women'smigration. positively women fromthe village,migration As duration stays extendfor migrant of propensities but reduced women.Weexpectedthe former for for increase menbutaresignificantly finding, also reactions We that aresurprised the latter. speculate bothfindings might reflect villagers' by of to the loss of women froma villagecommunity, especiallyas duration stay mayweaken and socialties to villageand householdcommunities raiseconcernsaboutexpectedsocial is as This relations. speculation notunfounded the following quotesfromfieldwork suggest: are "Wedon'twantdaughters go. There onlya few of them. If to they go, who willfetch water?Whowillcookthe food for me, their motherinterrupts "But or Another her, mother, theirfather?" usually can save more money than sons. Men spend daughters migrant womenknowhowto save money,men liketo havefun. moreeasily, Like theirfathers!" (Focusgroupinterview amongmothers) from "Weneed factory jobs in ourvillageto keepourdaughters with mother) (interview 60-year-old leaving." in that socialcapital also important is our Inconclusion, studyhas demonstrated migrant Withthis extension to the Thaicontext, even when consideringrural-urban migration. if studiesinotherlocales,especially those studies the Thailand, case can be madeforsimilar from both men and multilevel affordthe collectionof longitudinal, migrantinformation in women. Because of the verydifferent patternsof migration the Thaicase andthe data in variation the gendered patternsof that there is significant available us, specifically to and migration that data were collectedfrom both men and women, we are able to offer refinementson theories about migrantsocial capital.We confirmwhat ethnographic studiesandexpanduponthe few prior foundintheirqualitative havepreviously researchers studiesof genderand migration. resultsfor withverydifferent of socialcapital imbuethe quality migrant relations Gender of The outcomes. patterning these effectscanbe quiteprofound men'sandwomen'smigration

250 * SocialForces Volume84,Number1 * September 2005

andhaveimplications understanding rateandpacemigration of rural for the out areas. Although we do notexplore these effects,we suspectthatthispatterning influence will social gendered in and of as influence impact the organization villages households origin, wellas differentially of migration village household and economies.Oneexample upon mightbe thathighratesof womenmayyield different flowsof remittances to villages origin back of than by migration very if menpredominate a migrant in stream. the quotesfromourinterviews As in indicate, the Thai contextwomenareconsidered morereliable remitters aremen(Curran Saguy2001). than and Onthe otherhand,highratesof femalemigration alsoweakentheirties to natal may villages andshiftpatterns careprovision fromdaughters sons, especially highratesof female if of to are deterrent maleout migration of and (Curran Saguy2001).Theseare migration a significant offered tentative as that moreexploration. speculations require Ourstudyaddsto the smallbutgrowingcollection quantitative of research whichbuilds on whatethnographers been suggestingfora longtime:ignoring gendercontentof have the social capitalmay mask important migrant insightsaboutthe dynamicsof the migration social category, process. Ouranalysispresumesthat because genderis such a profound of sociallife,thatit inevitably influences quality migrant the of social manyaspects organizing capital,as well. We speculate that this is because men and women live the migrant In becausemenandwomenface different barriers moving, to experience differently. part very maintain differentrelationships with householdsand villages of origin,and experience in Future researchshouldexplore completelydifferent opportunities places of destination. these particular dimensions moreexplicitly, intoaccount acrossdestinations, taking variability workexperiences, of and barriers migration. to patterns remittances visitshome,andvarying Notes 1. Cumulatively causedmigration also be initiated maintained and mechanisms may through otherthansocial networksor in combination withsocial networks. These mayinclude in in or inequality communityincome distributions, inequality landholdings, cultural et. an change.Massey,Ariango al (1994)provide excellentreviewon how these factors also affecta cumulative migration dynamic. 2. Thefirstauthor conducted fieldwork NangRongandthe Bangkok in Area Metropolitan for extendedstays between1992and2000 (identifying citation removed). 3. Thisis the Thai to the of currency 25 bahtwas equivalent $1 during period thisstudy. 4. Atermliterally "to but termsmeansto havefunandparty. meaning takea trip," incolloquial 5. The data and information aboutthe surveysare available http://www.cpc.unc.edu/ at projects/nangrong/. 6. Entrance the Buddhist into as but life Sangha a monkis a tradition, a temporary eventthat can be as shortas threedays or as longas one year.Typically, entrance the Sangha into occursafterschoolandsome workandbeforemarriage, around 25. age 7. Thesample selection resulting migration bias from affects composition our the of significantly lifehistory fileifwe wereto examine from 50 of theoriginal data data all Out villages. migrants from 22 villages the survey the in havesignificantly different of education patterns migration, andwork than in the experiences thosewhowereliving the villages during 1994survey.

* TheEffects Gender Migration Thailand 251 of on in 8. The directionof the bias is unclear.If ruralout migration takes a life course path (decreasing with age) then this may diminishthe possible effects of cumulative unobservedpriormigration migration. mayalso be relatedto some of the Significant righthand side variablesincreasingthe possibilityof spuriousresults. We do have some reason to suspect that these effects are relatively minimal. The baseline data capturesthe verystartof the periodof rapideconomic growthand the beginningof Thailand. Our large movementsof young people out of rural villages in northeastern modelincludesmeasuresof migration experiencein 1983thatshouldhelpto diminish the possiblebiases. Theyoungersampleuponwhichwe replicate models, reflects our the youngestcohortsinthe data.Given overwhelming in migrant the rise in prevalence the villages,these migrantsmay be much less influencedby the increasein quantity and qualityof social capital, given high levels migrantexperience saturation.We expect that our results from the analysis of the overallsample will be significantly there is no dampenedwhen evaluatingthe effects on the youngercohorts. Finally, the differenceinthe substantive resultsfromthe smallersubset, although yeareffects are significantly more important because they capturethe effects of aging for this the smallerage cohort.Given demandsof ourmodel,the analysisof the larger datafile yields morerobustresults. 9. Exacttime in a place of destinationwas not measuredby the survey.The surveydid allowfor multiple return tripswithinone year(upto six trips- including trips)and the To survey links trips with work, education and destination information. measure monthsof migrant experiencewe took the numberof tripstakenwithinone yearand divided it into 12 months. Forone trip in one year we calculatedthe amount of experienceas six months, for two tripswe counted it as four months for each trip. Only10 percentof the sampleever made morethanone roundtripin one year. of 10.Thisequationand datafile are a replication Massey and Zenteno'smodeland data (1999)except thatwe add measuresof householdmigrant tripsand experienceand a controls. householdand community vectorof individual, 11.All of these, except for the measure of sex, were data collected duringthe 1984 survey. 12.Thisis measuredwithfourcategoriesof landownership:landless,nearlandless 1-10 A rai,somewhat landed(11-25rai)and landed(morethan25 rai). raiis 0.46 acres. 13.A temporarymigrantis a person identifiedin the household rosteras a permanent to householdmemberwho migratedout of the house for two or more months prior the surveybut plansto return. locatedifthereareone to two obstaclesto 14.A villageis consideredsomewhatremotely town. A villageis consideredveryremotelylocatedif thereare to traveling the district to three or moreobstacles to traveling the districttown. The obstacles we includein town thatis a cart of ourmeasurearethe presenceof a portion the routeto the district to and narrow), lack of publictransportation the district the path (unpaved,rutted, town, travel to the districttown takes an hour or more (as reportedby a village that headmanor key informant), duringthe year there are four or more months of difficulttravel to leave the village (this is also a measure of road conditions and to town. and to susceptibility flooding) it is 20 or morekilometers the district

252 * SocialForces Volume84,Number1 * September 2005

References
En El De Marias. de Lourdes. 1975. De Secretaria Educaci6n Arizpe, Indigenous LaCiudad Mexico. Caso Las Publica. .1980.LaMigracion Relevos LaReproduccion Por SocialDel Campesinado. Mexico: Centro de y El Estudios de Sociologicos, Colegio Mexico. . 1981."Relay and of Household." PeopleMove,editor Why Migration theSurvival the Peasant Jorge Balan. UNESCO. . 1985.CampesinadoMigracion. Secretaria Educaci6n Publica. de Mexico Mexico D.F, y and Bello,Walden,Shea Cunningham Li K. Poh. 1998. A Siamese Tragedy. Development in Moder Thailand. Books,Ltd. Zed andDisintegration M. and Resource Allocation Thai Blanc-Szanton, C. 1990."Gender Inter-Generational Among andSino-Thai Households." 79-102inStructures Strategies. and and vol. Women,Work Family, 3, editors Pp. Leela DubeandRajni Palriwala. Publications. Sage 1989."Family Personal and Networks International in RecentDevelopments and Boyd,Monica. Migration: NewAgendas." Review23(3):638-70. International Migration and 2001."On Auspices Female the of From to Cerrutti, Marcela, Douglas Massey. Migration Mexico the United States." 38(2): Demography 187-200. Wathinee Archavanitkul, Richter, Guest, Varachai, Chamratrithirong, Aphichat, Kritaya Kerry Philip Thongthai and 1995.National of Boonchalaksi, Nittaya Piriyathamwong PaneeVong-ek. Migration Survey Thailand. Institute Population SocialResearch, for and Mahidol Thailand. University, Bangkok, 1992.Gender Migration Developing in and Countries. Belhaven Press. Chant, Sylvia. M. in Networks the Migration Process." International Review Choldin, Harvey 1973."Kinship Migration 10:163-75. Nicole. in 1997.Maid Order HongKong. to Stories Filipina of Workers. Cornell Constable, University. 2003."Engendering The Networks: Caseof Mexican Curran, andEstelaRivero. Sara, Migrant Migration." 40(2). Demography C. Sara 2001. "Migration Cultural and A and Curran, andAbigail Saguy. Change: RoleforGender Social Journal International of Networks." Women's Studies 2(3):54-77. and Winters. 2002."Domestic International and From Davis, Benjamin, Stecklov Paul Guy Migration Rural Mexico: the Structure Composition." and Studies Disaggregating Effectsof Network Population 56:291-309. DeJong,Gordon, Richter Pimonpan and Isarabhakdi. "Gender, 1996. and to Values, Intentions Move Kerry inRural Thailand." International Review30(3):748-70. Migration and and and Kristin, Douglas of Espinoza, Massey.1999."Undocumented Migration the Quantity Quality Social and Social ed. Aldershot: Capital." Migration Transnational Spaces,Ludger Pries. Ashgate. and to Causation: Fussell,Elizabeth, DouglasS. Massey.2004. "TheLimits Cumulative International from Urban Areas." Migration Mexican Demography (forthcoming). Mark. of Ties:Network Revisited." 1: Granovetter, 1983."The Strength Weak Theory Sociological Theory. 201-233. and R. 1991.BetweenTwo Islands: Grasmuck, Dominican Sherri, Patricia Pessar. International Migration. Pp. 1-161. of Press. University California

TheEffects Gender Migration Thailand* 253 of on in Elizabeth andMonica M. 1990."Women Migration: and Grieco, Gender International Into Boyd. Incorporating of State Science,Florida University. Migration Theory." Working PaperSeries. College Social Lucien Jr.,andJane R. Hanks. 1963."Thailand: Hanks, M., Betweenthe Sexes."Pp.424-51in Equality in Women theNewAsia.TheChanging Social Rolesof Menand Women South South-East in and E. Barbara Ward. Paris Amsterdam: and Asia,editor UNESCO. Pierrette. 1994.Gendered Transitions. Mexican of Hondagneu-Sotelo, Experiences Immigration. University of California Press. Graeme. 1991."Village-Community Village Norms Ethnic Social and and A Networks: Review Ties, of Hugo, the Evidence From Third the World." 186-225 Migration in Decision editorsGordon Pp. Making, and W. Press. DeJong Robert Gardner. Pergamon Shawn the Question: Analysis Men, An of Women Gender and Kanaiaupuni, M.2000."Reframing Migration inMexico." Social Forces 78(4):1311-48. Charles. 1984."Mother Mistress Never Monk: or but a Buddhist Notions Female of Gender Rural in Keyes, Thailand." American 11(2):223-41. Ethnologist and A. Shoemaker. 1984. "Graphical Landwehr, J.M., D. Pregibon Methodsfor Assessing Logistic Models." Journal American of StatisticalAssociation61-71. 79: Regression Terd Gordon Manu Seetisarn Kanok and Limpinuntana, Aran Viriya, Patanothai, Charaoenwatana, Conway, Rerkasem. 1982. Agroecosystem An of Thailand. Khon Kaen Analysis Northeastern University. S. Social Economic and of AAPSS 510:60-72. Massey, ANNALS, Douglas 1990."The Origins Immigration." 1990a."Social Household and Causation Migration." of Structure, Strategies, the Cumulative Index 56(1):3-26. Population Graeme Ali AdelaPellegrino J. Edward and S., Massey,Douglas Joaquin Arango, Hugo, Kouaouci, Taylor. 1994."An Evaluation International of The American Case." and Migration Theory: North Population 1987. Social S., DevelopmentReview 20:699-752.Massey, Douglas andFelipe Garcia-Espana. "The Process International of Science237:733-38. Migration." E. 1997. "What's S., Mexico-U.S. A Massey, Douglas andKristin Espinoza. Driving Migration?Theoretical, and American Journal Sociology of Empirical Policy 102(4):939-99. Analysis." and 1994."ContinuitiesTransnational in S., An Massey,Douglas LuisGoldring JorgeDurand. Migration: of Mexican Communities." American Journal Sociology of 99:1492-533. Analysis Nineteen of of Massey,DouglasS., and ReneZenteno.1999."The Dynamics Mass Migration." Proceedings the National of Academy Sciences:5328-35. Cecilia. 1995."Kinship Networks LessonsFrom Qualitative a Menjivar, AmongImmigrants: Comparative International Journal Comparative of 36(3-4):219-32. Approach." Sociology 1997."Immigrant Networks: Salvadoreans Mexicans Comparative and in Vietnamese, Kinship Journal Comparative of Studies. Family Perspective." . 2000.Fragmented Salvadoran in Networks America. Ties. of Press. Immigrant University California B. in Women the Global Labor Contested Mills, Force.Consuming Selves. Mary 1997.Thai Desires, Rutgers University. 1999.The Village in Thai Press. Nartsupha, Chatthip. Economy thePast.Silkworm 1994. Determinants Private of Investment and Nidhiprabha, Bhanupong. Expenditures DirectForeign in Investment Thailand. Thailand Research Institute. Development

254 * SocialForces 2005 Volume84,Number1 * September Patricia. 1999a. Studies: Caseof NewImmigrantsthe United The in States." Pessar, "Engendering Migration American Behavioral Scientist 42(4):577-600. in A .1999b."The of Gender, and Role Process: Review Households, SocialNetworks the Migration andAppraisal." 53-70inHandbook International of The editors Experience, Pp. Migration. American Charles and Russell Hirschman, Kasinitz Josh DeWind. Philip Sage Foundation. Pasuk. 1980."The and The of Pacific OpenEconomy ItsFriends: "Development"Thailand." Phongpaichit, Affairs 53(3):440-460. Pasuk Baker. 1996.Thailand's Boom!Allen Unwin. and Phongpaichit, andChris
1998. Thailand's Boom and Bust. SilkwormPress.

Life: inthe Seri. and Mooban Press. Phongphit, 1990.Thai Village Culture Transition Northeast. in Michael. 1979.Birds Passage: of Labor Industrial Societies. Piore, Migrant Cambridge University Press. 1978. and Politics Portes, 8(1):1-48. Alejandro. "Migration Underdevelopment." andSociety 1991."Consensual An of Children's Pramualrathana, Anthony. Neglect: Interpretive Analysis Adult Support to the OldinRural Thailand." presented theAnnual at of Association Paper Meeting the Population of America, D.C. 21-23. Washington, March ToshikoKaneda, and Rindfuss,Ronald, Chattopahyay ChanyaSethaput.2002. "PanelStudies and Carolina of Carolina. Center, Migration." Population University North Suthad and Patamasiriwat. 'Agriculture." 81-117inThe 1993. Siamwalla, Ammar, Pp. Setboonsarng Direk in Thai editor Warr. Press. Economy Transition, Peter Cambridge University Anchalee Nitaya and Prabhudhanitisarn. "Changing 1992. Socio-Economic Rolesof Singhanetra-Renard, Thai WomenandTheir and in Countries, Pp. Migration." 154-73inGender Migration Developing editor Chant. Belhaven Press. Sylvia Oded.1991.The of Basil Stark, Migration Labor. Blackwell. Levhari. 1982."OnMigration Riskin LDCs." and Economic and Stark, Oded,and Mario Development Cultural 31:191-96. Change 1993. Markets." 355-400 TheThai in in editor Pp. Sussangkarn, Chalongphob. "Labour Economy Transition, Peter Warr. G. Press. Cambridge University Darunee. 1995."Gender Development Thailand." presented theSoutheast and in Tantiwiramanond, to Paper Asian Studies of December. Seminar, Seattle, University Washington, Washington, 1986."Differential Edward. Information Risk." and in Research Human Networks, Taylor, Migration, Capital andDevelopment: Human and 4:147-71. Migration, Capital Development and Booth.1991."Gender, and Tienda, International Marta, Karen Migration SocialChange." Sociology 6(1):51-72. JamesA. 1996."The of International Migration." Labor Professional Tyner, Gendering Philippine Geographer 48(4):405-16. editor. 1993.TheThai in Press. Warr, Peter, Economy Transition. Cambridge University and 1996. Peter, Bhanupong Macroeconomic Warr, Miracle: Stable Nidhiprabha. Thailand's Adjustmentand Sustained Growth. World The Bank. de and Sadoulet. 2001."Family Community and Winters, Alain Janvry Elisabeth Paul, in Networks MexicoUSMigration." Journal Human of Resources 59-84. 36(1):1

in TheEffects Gender Migration Thailand* 255 of on and Bencha. 1992."Thai Structure Organization: RolesandDutiesin Family Changing Yoddumnern-Attig, of in Rolesand Statuses Women Thailand, editors Historical Perspective." 8-24 in Changing Pp. Amara and Bencha Soonthorndhada, Chanya Sethaput Anthony Kerry Yodumnern-Attig, Richter, for and Mahidol Pramualrathana. Institute Population SocialResearch, University.

Das könnte Ihnen auch gefallen