New Migration Destinations and Agglomeration Economies in the United States

Transcription

1 New Migraion Desinaions and Agglomeraion Economies in he Unied Saes José Francisco Pérez-De latorre FES Acalán, Naional Universiy of Meico (UNAM) Luis Quinana-Romero FES Acalán, Naional Universiy of Meico (UNAM) Marcos Valdivia-López CRIM, Naional Universiy of Meico (UNAM) Populaion Associaion of America Anual Meeing. Washingon D. C. Topic 6. Migraion, Neighborhoods and Urbanizaion Session 601. Immigran Desinaions: Issues and Impacs. Absrac: Laino immigrans, regardless of heir abiliies, have been hisorically seled in he so called radiional migraion desinaions such as he Sae of California and ohers. Neverheless, migraion desinaions ehibied an imporan geographical urn in he nineies since hey sared o feaure regional dispersion, siuaion known in he lieraure as he emergence of "new immigraion desinaions." This paper conribues o he undersanding of his discussion by analyzing he role of agglomeraion economies, as suggesed by he heoreical framework of he New Economic Geography, o eplain he formaion of new Laino migraion desinaions in he USA during he period Our resuls show ha he new migraion desinaions can be in par eplained by agglomeraion and congesion economies ha have impaced firms and migran worker mobiliy decisions. 0

2 Inroducion The immigraion paern in he Unied Saes has displayed an ineresing feaure during recen decades, which consiss in a change in he process of foreign-born populaion mobiliy ha subsiue radiional immigran regions for "new growh saes" (Passel y Suro 2005). From an economic perspecive, ypically he eplanaion of new migraion desinaions would rely on sudying wage differenials beween origin and desinaion areas. However, under hese radiional perspecives wages are in somewha given, in he sense ha he modeling sraegy does no ake fully ino accoun he economic forces ha deermine wage differenials and by eension labor mobiliy. Wih his respec, he heoreical perspecive of he New Economic Geography (Fujia-Thisse, 2002) offers an eplanaion of hese regional forces (ha arac or repulse firms and workers) hrough modeling agglomeraion and congesion economies. We consider ha he New Economic Geography (NEG) can be a reliable eplanaion for he new migraion desinaions in he USA. Besides, one imporan advanage of he NEG perspecive is ha makes he spaial (geographical) dimension, conained in he wage differenials across labor markes regions, eplici in he modeling of migraion process. In his paper, he new migraion desinaions in he USA a couny level are sudied as a geographical diversificaion of migraion flows. Our cenral hypohesis is ha new migraion desinaions mus ehibi spaial concenraion because agglomeraion and congesion economic forces are playing an imporan causal role in he developing of hese regions. This sudy focuses in he Hispanic migraion a couny level. In secion 1, we make a review of he heoreical economic models of migraion and also we provide sylized facs o idenify agglomeraion economies role in he deerminaion of migraion flows. In secion 2, we rely on Eploraory Spaial Daa Analysis (ESDA) o highligh he spaial feaures of he Hispanic migraion flows a couny level. In he hird secion, i is described a formal model proposed by Croze (2004) under NEG approach o eplain migraion flows hrough agglomeraion economies. Finally, in he secion 4, we propose a spaial economeric implemenaion of a simple version of he core-periphery model a la Croze (2004) o show he imporance of he marke access effec on he new migraion desinaions. We conclude wih a secion of final consideraions. 1

3 1. Migraion heory and Agglomeraion Economies Labor migraion, boh inernal and inernaional, responds in some een o economic differenials a he regional level, mainly o wage differenials. An economic sandard model of migraion assumes wo regional labor markes (norh and souh) in a region wih free mobiliy of he labor force and similar skills. If average wages in he norh are greaer han wages in he souh, workers move o he norh pursuing higher wages and higher level of uiliy. The unfolding sory of his is a single regional wage ha eliminaes wage differenials because labor mobiliy forces wages down in he norh. (Borjas 2000) The above resul "can be enough" o eplain inernal migraion, bu in he case of inernaional migraion oher facors as race, disance, legal provisions, ec. mus be considered. Inernaional migraion has been sudied from he perspecive of economic developmen, which posulaes ha migraion is caused by geographical differences in he regional labor markes ha eplain differences in he endowmens of capial and labor.. The wage gap beween developed and developing counries encourages lower wage workers counry migrae o counries wih higher wages;and in he long run, migraion causes a endency of wage equalizaion in boh counries. ; and he only prevailing wage differenial would be coss of inernaional movemen, which are mosly psychological. (Lewis 1954; Ranis y Fei 1970; Harris y Todaro 1970). Migraion decisions are eplained by Sjaasad (1962), who argue ha migraion is a ype of invesmen which increases human produciviy in an individual sense. A poenial migran evaluaes economically coss and reurns associaed o mobiliy. Migran s coss and reurns are classified ino moneary and nonmoneary under he assumpion ha wages are marke deermined and here are no barriers o labor mobiliy and oher inpus among indusries or regions. 1 In his sense, he decision o migrae is like an invesmen in human capial where ne presen value of he fuure income less coss (aribued o he movemen) are calculaed (Borjas 2000). This framework can be adaped including he probabiliies of finding a job and of being depored (in he case of undocumened migrans) 1 In addiion o hese coss and benefis ha are called privae coss, he individualis faced wih social coss and benefis, which in his analyical framework are inroduced as eernaliies such as a srucures, school sysems, among ohers. For furher reference of each of hese elemens see Sjaasad (1962). 2

4 in he desinaion and; also including he likelihood of finding employmen in he place of origin o obain he epeced income differenial. This las perspecive emphasizes ha who migraes makes a raional valuaion of coss and benefis individually. In conras, he new economics of migraion argues ha he migraion decision is made a he household level, i.e. a collecive decision made by a family or household in order o maimize income and minimize risk in heir welfare. Households have he abiliy o reduce risk hrough diversificaion in he allocaion of resources of heir home or family. Tha is, he family decides who members will be allocaed o economic aciviies in he place of origin and who are allocaed o oher places making eplici he differences beween he labor markes of origin and desinaion. (Sark y Levhari 1982; Kaz y Sark 1986; Lauby y Sark 1988). Remiances and mechanisms of risk reducion are wo ways ha families use o maimize absolue income and minimize risks of loss of income. Typically, families reduce he risk of loss of income hrough remiances from relaives who are in he desinaion. On he oher hand, in desinaion places (developed counries), he risks of household income are minimized by privae insurance markes or governmen programs because i is assumed ha in places of origin (developing counries) he insiuional mechanisms are imperfec, absen or inaccessible o poor families. Likewise, he new economics of labour migraion also argues ha households no only decide o send migrans o increase family income in absolue erms, bu also o increase he relaive income compared o oher households in heir home communiy. The aim is o reduce he so-called relaive deprivaion compared wih oher households. Therefore, relaive deprivaion highlighs he disribuion of income in he place of origin. If income of rich households increases while income of poor households remains consan, hen he relaive deprivaion of he laer increases. By increasing he relaive deprivaion, he poor household's uiliy is negaively affeced, which causes o epand incenives for he household members o migrae. (Sark y Taylor 1991) An increasingly imporan aspec in he sudy of migraion is he human capial embodied in he migran. The lieraure indicaes ha human capial deermines wha ype of migrans move in relaion o he o he wage disribuion in he origin. Borjas hypohesis (1987) abou self-selecion of migrans is based on he idea of income maimizaion (Roy 1951). The hypohesis is ha migrans do no move randomly from he counry of origin 3

5 because hey are evaluaing heir skills reurns in local (origin) labor markes.assuming ha skilled workers may move freely among regions and ha workers wih higher skills will move o high income counries,, i can be derived wo ypes of selecion ha characerize migraion flows (Borjas 2000):1) Posiive selecion occurs when migrans have above average skills (Migran flows from he sie of origin o he desinaion sie are posiively seleced when he desinaion offers a higher reurn rae o skills). and; 2) Negaive selecion occurs when migrans have below average skills (Migran are negaively seleced when he desinaion region offers higher reurns o skilled workers, so ha less skilled workers will leave he region of origin). Under his perspecive, a branch of he lieraure analyzes migraion flows (Chiquiar and Hanson 2005). In general, he deerminans of migraion flows ha have been idenified by he lieraure are: age, since migraion is common among young people; educaion, because migraion is common among workers wih higher educaion; disance, he smaller he disance, he greaer he propensiy o migrae (Borjas 2000). Oher deerminans are, rade, populaion growh a he origin, he raio of per capia income beween he origin and desinaion, he sock of immigrans as a proporion of he populaion a desinaion in previous period (Michell and Pain 2003), unemploymen, foreign direc invesmen (Mendoza 2006), poliical facors (asylum, immigraion and work resricions), living sandards, social neworks, (Roer and Vogler 1998), macroeconomic variables (Bruker and Siliversovs 2004), among ohers.likewise, he self-selecion approach is accompanied by a soring eplanaion of migrans in desinaion sies. Migrans from any sending counry should be classified according o he yields ha heir skills would produce in oher desinaions. (Grogger and Hanson 2008). Thus, workers wih lower skills move o regions wih lower reurn raes, while more skilled workers choose regions wih higher raes of reurns ;finally, workers wih average skills will move o regions wih average raes of reurn. (Borjas,Bronars and Trejo, 1992). One aspec o highligh is ha self-selecion and soring are independen evens. While seleciviy of migraion depends on he reurns o skills in he counry of origin (among oher facors), migran soring depends on he reurns o skills in he counry of desinaion. So far, we have eposed he main neoclassical economic eplanaions abou migraion decisions; addiionally, he economic approach has also analyzed he ype of economic impacs in he origin and desinaion places ha migraion cuases: 4

6 assimilaion and job search of migrans in desinaion (Borjas 1999, Borjas 2000, Lalonde and Topel 1990), welfare (Borjas 2000 and 2001), brain drain, drain gain (Kanbur and Rapopor 2005, Beine, Defoor and Docquier and 2005, and Docquier 2006) and economic convergence (Solow 1956, Ramsey 1928 and Braun 1993) The heoreical eplanaions of migraion flows discussed above are quie similar in assumpions, consequences and mechanisms. Bu wo feaures mus be highlighed. The firs is he wage differenial as a resul of differenials in labor and economic srucures beween sies. The second is he space issue, i.e., economic differenials eis because economic aciviy in a counry-region conained in a specific geographical area. These wo feaures are implici in any analysis of migraion processes and hey mus be aken ino accoun, seriously, in he eplanaion and predicion of labor mobiliy. In his sense, he migraion neoclassical approach ends o impose small spaial heerogeneiy in heir models (Borjas, 1999). Migran flows follow a paern ha is far from being homogenous. For eample, see map 1 ha displays, he regional disribuion of Hispanics respec o he oal of U.S. populaion a couny level in Map 1 shows ha in fac he larges concenraions of Hispanic populaion in 2008, is concenraed in counies belonging o seven saes, : California, Nevada, Arizona, New Meico, Colorado, Teas and Florida. Map 1. Share of Hispanic Populaion o oal populaion. Unied Saes Fuene: Elaboración propia con base en US Census Bureau (2009) Given he spaial paern displayed in map1, a quesion ha arises is if here is any spaial-regional economic force (beyond wage differenials) ha eplain precisely he 5

7 spaial configuraion of hispanic migraion in USA. In he following secion will discuss a heory of economic geography ha provides a possible answer. 2. Agglomeraion economies and migraion. Agglomeraion of economic aciviy in space is an aspec of economic analysis a he regional level ha has become relevan in recen decades. Agglomeraion or concenraion is a opic sudied by economic geography, as a feaure of he spaial organizaion of economic aciviy. Fujia and Thisse (2002) argue ha he concep of agglomeraion economies refers siuaions, in where high income naions are clusered in indusrial ceners and per capia produciviy declines he greaer he disance from hese ceners. In addiion, regions wihin counries are clusered in high growh raes, sies clusered by indusry specializaion, indusrial business disrics wih high echnology are clusered producing linkages and rade among hem. Before discussing he heory of agglomeraion from economic geography, i is useful o disinguish beween he conceps of concenraion and agglomeraion. Boh concenraion and economic agglomeraion refer o how he economy or a par of i is disribued in space, for eample, if a specific par of economic aciviy can be found in a few locaions, eiher a ciy, or counry. The basic disincion beween hese wo conceps is ha while concenraion analyzes he spaial localizaion of secors or indusries, agglomeraion analyzes locaion in space of economic aciviy, such as he manufacuring secor as a whole. (Brackman, e. al. 2009). Specifically, we can posulae ha agglomeraion is he spaial disribuion of aggregae economic aciviy.the sudy of agglomeraion from his perspecive has been addressed byeconomic geography. Cronon (1991) in a way ha economic aciviy is deermined geographically by wo naures. The firs naure, indicaes ha agglomeraion of economic aciviy in space is due o he disribuion of naural resources as raw maerials, climae, landforms and naural means of ranspor (such as rivers). In his sense, firs naure deermines he agglomeraion of economic aciviy by he characerisics of he sies which are given eogenously; in conras, he second naure esablishes ha agglomeraion is endogenous o economic aciviy as a resul of human acions o seize he firs naure. (Combes, e. al. 2008). In he case of he firs naure eplanaion for economic agglomeraion, we can find in economic 6

8 heory eplanaions based on comparaive advanages and echnology ha have formulaed under he influence of Ricardo s classical rade model and, more recenly, Hekscher-Ohlin model of inernaional rade. These approaches focus o eplain paerns of specializaion, producion and rade among counries. In he case of second naure approach, economic agglomeraion is resul of eernal economies. Posiive eernaliies or economies of agglomeraion in economic heory are classified in wo ypes: urbanizaion economies and localizaion economies (Graham 2007), The firs refers o iner-indusrial eernal economies of scale, while localizaion economies o eernal economies of scale in specific indusries. (Brackman, e. al. 2009).Some auhors posulaes ha urbanizaion economies are deermined by sie ameniies and nice weaher, proimiy o oceans and lakes, beauiful places, qualiy of life, global-ype ciies, suiable employmen, ec.(parridge 2010). While he economic agglomeraion caused by localizaion economies can be aribued o hree marshallian ypes of eernal economies, proposed by Marshall (Combes, e. al. 2008): 1) Specialized inpus disribuion whose uni coss are lower when demand for hese inpus is high enough, 2) Thick labor local markes ha allow maching jobs and workers; and 3) Inense echange of ideas and spillover effecs ha increase produciviy and rigger growh. Mos of he lieraure depars from hese hree eernaliies or any of hem o build up formal eplanaions where economies of agglomeraion are considered (Duranon and Puga 2004, Oaviano and Thise 2004). In addiion o Marshallian eernaliies, he lieraure posulaes ha imperfec compeiion in markes also creaes economies of agglomeraion. For eample, firms ha maimize heir profis in a cone of imperfec compeiion se prices because hey depend on he spaial disribuion of firms and consumersthe srucures of imperfec compeiive markes are wo main ypes: monopolisic compeiion and oligopolisic compeiion. (Combes, e. al. 2008). Economies of agglomeraion eplained by eernaliies of localizaion economies ype are hose ha have been modeled mosly. The differences beween he models lie in he fundamenals in which hese forces are analyzed. The main fundamenals are: price mechanism, ren of land, increasing reurns, ransporaion coss, eernaliies, monopolisic compeiion, among ohers 2. 2 For a deailed review of hese fundamenals see Fujia and Thisse (2002). 7

9 Lieraure on agglomeraion economies shows a common elemen in where spaial configuraion of economic aciviy is resul of a process ha involves wo ypes of forces: agglomeraion forces (cenripeal forces) and dispersion forces (cenrifugal forces). In he empirical lieraure hese agglomeraion and dispersion forces can be analyzed hrough five empirical componens from he perspecive of localizaion economies (Brackman, e. al. 2009). 1) Home marke effec. Regions wih high demand for goods produced by indusries wih increasing reurns produce more and are ne eporers of hese goods. 2) A large poenial marke increases prices of local facors. A large marke increase demand for local inpus, and his increases he facor prices. 3) A srong marke poenial induces economic facors o move. Producion facors wih free movemen will be araced o hose markes in which companies pay relaively high prices o producion facors. 4) Shocks sensibiliy. Changes in economic environmen may rigger a change in equilibrium spaial disribuion of economic aciviy. This hypohesis posulaes he eisence of muliple equilibria in models of economic geography. 5) Reducion in rade coss lead o agglomeraion, a leas unil a criical level of ransporaion or rade coss is reached. In erms of our research is imporan o esablish how migraion can be eplained by agglomeraion economies. In his sense, domesic marke effec due o increasing reurns indusries would creae a large marke poenial ha increases local facors prices, aracing facors of producion such as labor. Specifically, he economic geography lieraure ha aacks he undersanding beween migraion and agglomeraion economies is he so-called New Economic Geography (NEG), mainly hrough he cener-periphery model. Fujia, Krugman and Venables (2000) consider he eisence of agglomeraion economies ha allow concenraion of economic aciviy in space because of increasing reurns o scale and a marke srucure of imperfec compeiion. In a seminal paper, Krugman (1991) developed a model of his ype assuming increasing reurns and imperfec compeiion as he basis of agglomeraion economies. A he sie where he cenripeal forces operae, agglomeraion of economic aciviy (cener) is riggered, producing a wide variey of goods ("love of variey") ha arac workers from he periphery. In paricular, here is a migraion economic lieraure ha focuses in he cone of he cener-periphery model of he NEG. (Holmes 1996, Thiessen and Van Oor 2001, Croze 2004). Likewise, oher models have been incorporaed ino his framework o analyze migraion heerogeneiy (Russek 2009, Morei 2010), skilled workers and human 8

10 capial (Sanchis-Guarner, and Lopez-Bazo 2006), labor marke fricions (Epifani and Gancia 2003) aes (Hafner 2005), ec.. Finally, he relaionship beween migraion and agglomeraion economies also has been discussed in a lieraure ouside he framework of he NEG, mainly by economies of urbanizaion or ameniies (Glaeser, e. al. 2001, Parridge and Rickman 2003, Sorper and Sco 2009, Parigde 2010, Marchiori, e. al. 2010). Moreover, here are also oher approaches ha aemp o model agglomeraion economies in conjunion wih spillover effecs (Hirose 2005) and informaion asymmery (Berlian and Yu 2010). New Migraion Desinaion of Hispanics in Unied Saes: Sylized Facs Theory and empirical evidence eposed in he laer secion propose ha migraion flows have a spaial characerizaion, i.e. ha here are places in which regional economic heerogeneiy deermine migraion flows. These sies, in principle, can be characerized by increasing reurns and eernal economies ha rigger agglomeraion economies. And hese cenripeal forces arac workers. In his sense, we can find in hehe New Economic Geography propose formal frameworks ha model he relaionship beween agglomeraion economies and migraion (Croze 2004). Moreover, from our perspecive, before modeling hese processes, i is imporan o indenify firs wheher or no migraion conveys a spaial characerizaion. A useful echnique o address his issue is eploraory spaial daa analysis (ESDA). In his secion, we eplore spaial diversificaion of Laino migran flows o new desinaions places in he Unied Saes. In he Unied Saes of America has emerged a new paern of he flow of immigrans ha consiss in srong inernal migraion of foreign-born populaions from saes wih large selemens oward new new saes, which are called "new growh saes" (Passel and Suro 2005). In addiion, Duran, Massey and Capoferro (2005) show ha in he case of Meican migrans in he Unied Saes, i is indenified geographical diversificaion during differen periods being he las period characerized by new desinaions. These auhors poin ou ha here are 4 facors ha deermine hese new desinaions : 1) he dramaic increase in he coss and risks of crossing San Diego-Tijuana border, 2) deerioraion of he economy of California and is respecive ani-immigran policy, 3) sudden privileges of free movemen of undocumened migrans in he Unied Saes 9

11 esablished decades ago hrough he IRCA policy and 4) he emergence of a srong demand for labor hroughou he counry. As i is said above, ESDA 3 is a useful echnique ha helps o show he disribuion of economic phenomena in space, and in his case will help us o analyze he new migraion desinaions of Hispanics in he Unied Saes a couny level. In all cases ha folow, ESDA is performed hrough open source sofware called GeoDa 4. Panel char 1 shows disribuion by quariles of he Hispanic populaion growh raes in he Unied Saes counies from 1980 o Four regions are deeced in he maps (he inensiy of he color deecs he dynamics of he couny). When map (A) is compared wih maps (B), (C) and (D), iis clear ha he mos dynamic region from 1990 o 2000 shifs oward counies associaed o "new migraion saes" such as Minnesoa, Iowa, Arkansas, Kenucky, Tennessee, Souh Carolina, Norh Carolina, Virginia and Georgia. From 2000 o 2007,oher counies from Souh Dakoa, Norh Dakoa, Wes Virginia, Maryland, Uah and Monana mus be also included in he "new desinaions". The las map of he panel corresponds o he growh beween , and i suggess ha counies from Colorado, Wyoming and Idaho are now par of he "new desinaions cluser." Panel Char 1. Quarile analysis (A) ESDA summarizes spaial properies of he daa, deecs spaial paerns, helps o make assumpions abou he geography of he daa, idenifies cases and sub ses of unusual cases, ec.. Anselin (1988 and 2005) eplains he basic principles of his echnique. 4 GeoDa sofware was developed by Luc Anselin and can download i freely from he websie of Geoda Cener a he following address: hp://geodacener.asu.edu/. 10

12 (B) (B) (D)

13 Source: Auor`s Calculaions base on U.S Census Bureau (2009) Bo maps analysis provide addiional evidence o highligh he eisence of a group of counies wih growh raes ouliers, well above he ohers in he fourh quarile (red) and far below he firs quarile (blue). Panel Char 2 shows ha while from 1980 o 1990 ereme growing counies were locaed in he old (radiional) desinaions, period shows a clear change where a new area of 313 new desinaions wih high dynamics is formed. These new desinaions correspond o he saes of Arkansas, Souh Dakoa, Norh Carolina, Souh Carolina, Georgia, Kenucky, Tennessee, Mississippi and Virginia. Panel Char 2. Bo-Map analysis (A) (B)

14 Source: Auor`s Calculaions base on U.S Census Bureau (2009) Panel char 3 shows migraion dynamics evoluion for periods and The firs period displays a process of dispersion, and i is clear ha high dynamic counies reduced heir number, if compared wih he period,from 313 o only 208. Finally, period shows he emergence of a new dynamic area ha corresponds o counies in he saes of Colorado, Wyoming, Idaho and Uah, heir consolidaion as a new desinaions depends on wheher hey are able o mainain heir high dynamism o arac migrans. Panel Char 3. Bo-Map analysis (A) (B)

15 Source: Auor`s Calculaions base on U.S Census Bureau (2009) Finally, panel char 4 provides sandard deviaion analysis of he growh raes showing evidence ha hispanic migraion above he mean values were locaed a he old desinaions from 1980 o 1990, while in recen periods hese above growh raes were locaed in he new desinaions saes. Panel Char 4. Sandard Deviaion Analysis. (A) (B)

16 (C) (D) Source: Auor`s Calculaions base on U.S Census Bureau (2009) The formaion process of new migraory desinaions is no ransiory, because he paern coninued during he nineies, and i also mainained he spaial associaion. We can see his rough he prevalence of counies ha ehini wih geographic proimiy and hey are clusered in highly dynamic regions. To demonsrae he eisence of his paern of spaial dependence, i is used he Moran's inde as a global measuremen of spaial auocorrelaion. 5 N N w ij ij ( i )( j ) 5 I N 2 S0 Moran`s inde is a spaial auocorrelaion indicaor and is defined as ( i i ) 1 where i is a quaniaive variable in region i, is heir sample average, W is a mari wih weighs w ij, N is a sample size and S 0 w ij i j. 15

17 Moran`s inde in panel char 5 indicaes ha spaial dependence is posiive, which shows ha counies wih high migraion dynamics are surrounded by counies wih high migraion dynamics and vice versa (low dynamic counies are surrounded by counies wih low dynamics). Alhough he inde is relaively small, i is saisically significan according o permuaion ess. Comparison beween periods shows ha he spaial dependence increases from he firs period ( ) o he second ( ) which is when he new migraory desinaions emerge. This spaial fac sugges haeconomic spillovers could operae as aracors of migraion labor force. Panel Char 5. Spaial Dependence: Moran`s Inde. Source: Auor`s Calculaions base on U.S Census Bureau (2009) Analysis based on Moran`s inde has disadvanages because i is a global indicaor of spaial auocorrelaion. Therefore, i would be adequae o carry ou a local spaial correlaion analysis known as LISA 6 which allows indenify groups of spaial 6 For a deailed eplanaion of LISA see Anselin (1988 and 2005). An inroducory eplanaion is showed in Quinana and Mendoza (2008). 16

18 associaion a couny level counies, hese groups could sugges regions of new migraion.panel char 6 shows ha in period, Florida s counies and counies of Washingon and Oregon kep high growh dynamics and hey formed growh regions. While in he same period, counies of Wes Virginia, Kenucky, Alabama, Mississippi, Souh Carolina and Norh Carolina formed a migraion region wih low dynamic.however, facs change drasically from 1990 o 2000, on he Florida s norh region, where counies of he saes of Virginia, Alabama, Norh Carolina and Souh Carolina sared o display significan spaial local auocorrelaion of highly dynamic counies. In conras, counies of saes considered as old migraion desinaions generaed a a spaial associaion paern wih low migraion dynamics. The las map on panel char 6 indicaes ha old migraion desinaions kep heir low dynamic in , bu a new spaial cluser of counies wih high migraory dynamic emerge in where saes of Wyoming and Colorado are included. Also counies of Norh Carolina, Souh Carolina, Virginia and Maryland, show high migraory dynamics in boh periods. Panel Char 6. Local Spaial Dependence: LISA Inde. (A) (B)

19 (C) (D) Source: Auor`s Calculaions base on U.S Census Bureau (2009) 3. New migraion desinaions and agglomeraion economies: An empirical NEG approach. We posulae ha spaial paerns of migraion flows is caused by agglomeraion economies., NEG offers an analyic approach which is useful o model migrans localizaion and migraion choice based on increasing reurns and imperfec compeiion assumpions (Krugman 1991). In paricular, we rely on Croze s proposal (2004) proposal, ha is based based on NEG, o undersand migraion dynamics of Hispanics o he new desinaions.the model Producion and Consumpion. 18

20 Marke consiss of R regions endowed by wo facors, labor and free mobiliy labor. Each region produces hree goods; a radiional and homogeneous good z, non-radable services y and manufacured goods. Good z is assumed homogeneous and produced under perfec compeiion, which is ranspored wih high coss beween regions and labor is no mobile. Price of good z and worker s wage are similar in any place. Seing z as numerary, such ha p 1 in any region.manufacuring and services goods z are produced in monopolisic indusries and labor free mobiliy is used o produce horizonally differeniaed varieies. Each variey is produced under scale economies. Wihin each indusry (, y ), he required labor o produce a q quaniy is respecively; q y q, where y y y y while and y are fied inpus requiremens. If y are marginal inpus requiremens o producion, n i, y n yi, denoe he number of goods varieies, y, produced in region i on period, he secoral employmen of region i on period is: L y, ni, qi, and Li, nyi, yqyi, y i i, R i (1) Denoing L i, as oal of free mobiliy workers oal in region i on period, so y L i, Li, Li,. On he oher hand, consumers are idenical and maimize heir uiliy hrough consumpion of every good, wih preferences ype Cobb-Douglas in region i on period : U C C C i i, R (2) 1 i, yi, i, zi, Where, y 1 are consumpion share in manufacure, services and radiional goods, respecively. C zi, is he amoun consumed in region i on period. C i, is a compound good which shows manufacured goods variey and i is epressed as follows; C i, mn c i, 1 1 m i, i i, R (3) 19

21 Where denoes subsiuion elasiciy beween varieies, c m i of variey m in region i on period and economy n, ir, is consumed quaniy n, is he number of varieies available in he, ni. Consumers canno impor services goods varieies from oher region, herefore, available varieies number of good y in region i is he variey number produced wihin region n yi,, and C yi, is: C yi, mn c yi, 1 1 y y y y m yi, i i, R (4) As i is usual in his framework, each producer has he same price ha maimizes benefis, price is a consan mark-up over marginal cos. Leing w i, o be workers wage wih free mobiliy in region i on period, fob price (for over board ) of a produced variey in region i is: y and p yi, ywi, 1 p i, wi, 1 i i, R y (5) However, free enry in each secor leads o zero profis in equilibrium. Furhermore, using (1) and (5) equaions and equilibirum condiion in each regional labor marke, would obained enerprises number in each region: Li, ni, and X X n yi L y y i,, i i, R y (6) Finally, assuming ransporaion coss ranspor of iceberg ype in manufacured goods, shipmen beween regions and assuming ha a fracion good ij ij 1 mels in he ransporaion so ij 1 good`s unis have o be ranspored from region i o deliver a uniy in region j. Assuming ha his ranspor coss are a disance increasing funcion beween wo regions d ij. ij 20

22 ij Bd i, j i, R, 0 y B 0 (7) ij Marke Poenial Funcion Tradiional good price z is normalized o one, such ha real wage of mobiliy workers in region i is: i P w i, yi, Pi,, (8) Where P i, and P yi, are CES ype prices inde of indusrial and services goods aggregae, respecively, in region i : P i, rr mn r, ir p r, rr n r, Bd ir p r, (9) P yi, 1 1 y 1 y 1 1 y p yi, nyi, p yi, (10) mn i, I is clear in (9) ha price inde of manufacured goods, can be reaed as he inverse of marke poenial funcion: his shows a comparable sum of he marke sizes in every regions weighed by heir disances. Furhermore, an inerpreaion would esablish ha price inde is greaer in disan regions where consumers should impor a grea proporion of heir demand from remoe locaions. Similarly, wih a consan nominal wage, so ha workers real income is lower in regions ha offer a relaively small number of service varieies. This price inde effec wih a high densiy in services and a access marke low coss makes such sies more aracive places o live. This is he forward linkage of Hirschman ype ha conribues o he cumulaive process of spaial agglomeraion Migraion Choice Migraion Model follows Tabuchi y Thisse (2002). Considering a free mobiliy worker k, ha choices her localizaion from region j o any of he R regions, including 21

23 region j. The migraion choice is a resul of qualiy of life comparison beween several sies. For empirical convenience, i is assumed ha migraion choice is made hrough maimizaion of he following funcion: V ln 1 1 k d bf k k k ji, ji, i i, i, ij ij i (11) Where i, is likelihood of find a job for a migran in region i on period and ij d 1 bf is a migraion cos which increases wih disance beween home and ij desinaion places. y b are sricly posiive coeficiens, and indicaes if regions i and j do no share a common border. ha capures Fij is a dummy variable ha k i is a sochasic componen k`s personal percepions of region i characerisics. To avoid endogeneiy problems in he empirical implemenaion, i is assumed ha migraion choices on period are deerminae by comparison beween regions on period 1. Furhermore, k ji individual k would decide o move o in region i if k k V ji, 1 V jr, 1, r i. Wih convenien assumpion made abou disribuion of deerminae by he logi funcion: k V ji ji, e 1 P M k k i, he likelihood of choice region i is V jr, 1, e (12) rr Epeced migraion flow from region j o region i is L j PM ji, oal flow of region j is PM region i is: L j jj,,. Similarly, emmigraion, 1. The migrans share of region j who choose go o i j migr ji, migr ji, rr e V e V k jr, 1 k ji, 1 e V k jj, 1 Using (5), (6), (8), (9), (10) equaions and definiion of V,, migran share of region j ha k ji choose region i, may wrie as; 22

24 ln ln i j migr ji, migr ji, ln y 1 y Li, 1 ln Lr, 1w r1d ij 1 w, 1, 1 ln 1 ~ i i dij bfij a j, 1 rr 1 1 (13) wih a~ j, 1 ln rr e k V jr, 1 e k V jj, 1 Equaion (13) capures he valuaion faced by poenial migrans who have o choose beween several possible locaions. Lef side of equaion (13) is he migrans share from a given region who have decided move o region i. On he righ side, hird erm represens epeced wage on region, which increases wih nominal wage on home region and a likelihood of find a job in his region. Fourh erm capures he impac of bilaeral disance of migraion flows and i is inerpreed as a displacemen average cos. Firs wo erms denoes access marke of i` s regions, which are; price inde o varieies of nonradable services goods and o manufacured goods in region i, respecively. Second erm in equaion (13) in he mos imporan erm, because corresponds o a marke poenial funcion and relaes labor migraion wih indusrial aciviies localizaion, and would sugges forward linkages highlighed by NEG. On he oher hand, main parameers of he NEG framework (subsiuion elasiciy and parameers of rade cos funcion) can be esimaed from price inde funcion. Furhermore, if empirical analysis confirms ha his price inde encourages migraion flows, i.e. migrans follows marke poenial, validaes he role of forward linkages as a par of endogenous agglomeraion process. 4. Access Marke Effec on he configuraion of new migraion desinaions. A simple model from spaial economerics. ESDA analysis is only eploraory, furhermore confirmaory analysis requires a formulaion based on spaial economeric models. When spaial dependence of daa is 23

25 confirmed, is necessary specify a spaial regression model ha ake ino accoun dependence. The general model proposed is: y W y X 1 W 2 (14) wih μ~n(0,ω) being diagonal elemens of Ω ii =h i (zα) wih h i >0 where y is a endogenous variable vecor, X is a eogenous variables mari and ε is he error erm ha incorporaes an auoregressive spaial dependence srucure, W 1 y W 2 are spaial weighs mari. 7 The eplanaion of migraory process based on NEG approach, as discussed in he las secion, and he eisence of spaial effecs, leads o assume he number of Hispanic migrans in he Unied Saes depends on some een of he economic aciviy concenraion and dispersion variables across counies and heir spaial characerisics In he las secion, we presened a formal model of NEG model ha incorporaes agglomeraion economies o eplain space disribuion of labor migraion. (Croze 2004). The model posulaes ha labor migraion is no only deerminae by real wage differenial, bu also by oher facors like mobiliy coss and risks associaed o migraion. The hypohesis raised here is ha a grea marke poenial induce labor facor flows. Spaial dependence of daa showed in ESDA, see secion 2, sugges possible ineracion effecs of sies aribues in he migraory process eplanaion. Eplanaion of migraory process based on NEG approach and hese spaial effecs, may assume ha migraion growh raes depends on economic aciviy concenraion and spaial characerisics of he regions. To consider boh elemens in an economeric implemenaion, we proposed he following spaial model a la Croze (2004): TMigr i WTMigr 1 i 1TIPRi i (15) 7 See abou Quinana and Mendoza (2008) for an inroducory eplanaion o spaial weighs mari. 24

26 i W2 i i where: TMig i is he migraion growh rae of hispanics in couny i on period, TIPRi is he growh rae of real personal income in couny i on period. The laes variable is considered a cenripeal force o ake ino accoun marke access effec, which is a marke size, as i is proposed by NEG.The specified equaion is esimaed o periods , y The resuls are shown in able 1 from which i follows ha in each period he growh rae of real personal income is a variable ha impacs posiive and significanly o migraion growh rae. Comparison beween periods showed ha access marke effec is reinforced considerably in period, where a percenage poin of growh in income ranslaes ino 1.7 poins of growh in he rae of migraion, whereas in previous period corresponding o domain of radiional migraion desinaions income differenial has an impac of only 0.8%. Resuls for he las period show a lower impac of marke effec wih a coefficien of only 0.51, siuaion ha could indicae he presence of dispersion effecs in he new migraory desinaions, which somehow we have observed wih he creaion of new migraion counies in ha period. Table 1 Esimaion for Differen Periods Period Consan TIPR Spaial Effec Type p-valor p-valor p-valor LAG LAG SARMA Source: Auors`s Calculaions based on U.S Census Bureau (2009) and Couny Business Paerns (2009) In all cases, he spaial effecs es indicaes spaial models as alernaives, dominaing in he wo firss periods a spaial lag model. Resuls of spaial models esimaion are shown in able 2. Marke access effec, i.e. he growh rae of real personal income, remain are similar o he esimaions wihou spaial effecs in able 1; however, spaial effecs are significan and eplain 0.13% of migraion growh in and periods., In he period where new desinaions had greaer growh, spaial 25

27 effec is bigger ( he coefficien had a value of 0.23%). Aragones (2006) argues ha in he new migraion desinaions, social neworks are no he main facor o arac migrans, however spaial effec deeced during period can be a a spillover effec ha influence migraion among counies. Table 2 Esimaion for Differen Periods Period Consan TIPR Spaial Effec Type p-valor p-valor p-valor LAG LAG SARMA Source: Auors`s Calculaions based on U.S Census Bureau (2009) and Couny Business Paerns (2009) 5. Final Consideraions. This paper provides a review of he lieraure on migraion from economic discipline as well some sylized facs abou spaial configuraion of Hispanic migraion in USA. We proposed ha hese spaial paerns can be eplained hrough agglomeraion economies. Agglomeraion economies sudied from economic geography sugges ha hey are casued by economies of localizaion and urbanizaion. We see in his paper ha he cener-periphery model of he New Economic Geography which assumes increasing reurns and imperfec markes provides an ineresing approach o model migraory flows in he cone of agglomeraion economies.as proposed by Croze (2004) In his paper, we esimae a simple model based on spaial economerics based which is a "spaial version" of Croze s model (2004) ha helps o undersand he deerminans of new migran desinaions of Hispanics in he Unied Saes. Our resuls show ha he period was subjec o srong geographical-spaial reconfiguraion of he flows paerns of migraion leading o he emergence of new desinaions. During he period , we find evidence of a process of dispersal oward new counies ha arac 26

28 immigraion. We conclude ha during he concenraion period , he impac of marke access (demand) on he aracion of migraion was relaively higher if compared wih he dispersal periods ( in he old places and wih he emergence of new migraory regions). Likewise, we deec ha spaial effecs of conagion in migraory processes are significan, which provides furher evidence ha spaial aspecs of he new counies (suc as geographical locaion, producive linkages, economic spillovers) are aconribuing facors o he growh of migraion. The resuls shown in his sudy are preliminary because are based on a simplified model, herefore his sudy should be eended o esimae he equaion wih all he deerminans ha arising from a complee NGE model. In addiion, we believe ha his sudy mus incoporae urbanizaion economies ype facors (such as ameniies) o es wheher he deerminans of agglomeraion economies are indeed relevan o eplain he spaial disribuion of he new Hispanic migran desinaions in he Unied Saes. Parigde (2010) argue ha in he case of he Unied Saes, agglomeraion economies aribued o ameniies, are more predicive han he proposals from he NGE. I remains for furher research o eplore hese issues. References Anselin, L. (1988) Spaial Economerics: Mehods and Models. Kluwer Academic Publishers, Dordrech, The Neherlands. Anselin, L. (2005) GeoDa 0.9 User s Guide. Spaial Analysis Laboraory (SAL). Deparmen of Agriculural and Consumer Economics, Universiy of Illinois, Urbana- Champaign, IL. Aragonés e.al. (2006) Migración y mercados de rabajo en el nuevo siglo. Un acercamieno eórico. Ponencia en la reunión de la Asociación Lainoamericana de Sociología. Quio. Barro, R. y Sala-i-Marin, X (2004) Economic Growh. Second Ediion. The MIT Press. USA. Beine, M., Defoor, C. y Docquier, F. (2005) Skilled Migraion, Human Capial Inequaliy and Convergence. Mimeo. Universiy of Luemburg, Universié Libre de Bruelles y Caolic Universiy of Louvein. 27

29 Berlian, M. y Yu, C. M. (2010) Locaional Signaling and Agglomeraion. Mimeo. Borjas, G. (1987) Self-Selecion and he Earnings of Immigrans. The American Economic Review. Vol. 77, No. 4., G., Bronars, S. y Trejo, S. (1992) Self-selecion and inernal migraion in he Unied Saes. NBER Working Papers. No , G. (1999) The Economic Analysis of Immigraion. En Orley C. Ashenfeler y David Card, eds., Handbook of Labor Economics. Amserdam: Norh-Holland., G. (2000) Economics of Migraion. Inernaional Encyclopedia of he Social and Behavioral Sciencies. Secion No. 3.4, Aricle No. 38., G. (2001) Does Immigraion Grease he Wheels of he Labor Marke? Brooking Papers on Economic Aciviy, Vol. 2001, No. 1. Brackman, S., Garresen, H. y van Marrewijk, Ch. (2009) The New Inroducion o Geographical Economics. Cambirdge Universiy Press. Unied Kingdom. Braun, J. (1993) Essays on Economic Growh and Migraion. Ph. D. disseraion, Harvard Universiy. Brüker, H. y Siliversovs, B. (2004) The Macro-Deerminans of Inernaional Migraion in Europe: Evidence from Germany. Paper presened a he Conference Immigraion Issues in EU-Turkish Relaions Bogazici Universiy. Ocober Combes, P. P., Mayer, T. y Thisse, J. F. (2008) Economic Geography. The Inegraion of Regions and Naions. Princeon Universiy Press. USA. Chiquiar, D. y Hanson, G. (2005) Inernaional Migraion, Self-Selecion, and he disribuion of Wages: Evidence from Meico and he Unied Saes. Journal of Poliical Economy, Vol. 113, No. 2. Cronon, W. (1991) Naure s Meropolis. Chicago and he Grea Wes. W. W. Noron. USA. Croze, M. (2004), Do migrans follow marke poenials? An esimaion of a new economic geographic model, Journal of economic Geography, 4, pp Docquier, F. (2006) Brain Drain and Inequaliy across Naions. IZA Discussion Paper. No Durand, J., Massey, D. y Capoferro, Ch. (2005) The New Geography of Meican Immigraion. En New Desinaions. Meican Migraion in he Unied Saes. Ed. Zuñiga, V. y Hernández-León, R. Russel Sage Foundaion. New York. 28

30 Duranon, G. y Puga, D. (2004) Micro-Foundaions of Urban Agglomeraion Economies. En Handbook of Regional and Urban Economics. Volumen 4. Capíulo 48. pp Elsevier. Epifani, P. y Gancia, G. A. (2003) Trade, Migraion and Unemploymen. Mimeo. Fujia, M., Krugman, P. y Venables, A. (2000) Economía espacial. Las ciudades, las regiones y el comercio inernacional, ed. Ariel. Barcelona Fujia, M. y Thisse, J. F. (2002) Economics of Agglomeraion. Ciies, Indusria Locaion, and Regional Growh. Cambrigde Universiy Press. Unied Kingdom. Glaeser, E., Kolko, J. y Saiz, A. (2001) Consumer Ciy. Journal of Economic Geography 1. Graham, D. (2007) Agglomeraion Economies and Transpor Invesmen. Join Transpor Research Cenre (OECD-ITF). Dicussion Paper No Grogger, J y Hanson, G. (2008) Income Maimizaion and he Selecion and Soring of Inernaional Migrans. NBER Working Paper Febrero de 2008 Hafner, K. A. (2005) Agglomeraion, Migraion and Ta Compeiion. BERG Working Paper Series on Governmen and Growh. Working Paper No. 52. Hanson, G. (2005) Emigraion, Labor Supply, and Earnings in Meico. Universiy of California, San Diego and Naional Bureau of Economic Research. April Hanson, G y McInosh, C. (2007) The Grea Meican Emigraion. Universiy of California, San Diego and Naional Bureau of Economic Research. December Harris, J. R. y Todaro, M. (1970) Migraion, Unemploymen, and Developmen: a wosecor analysis. American Economic Review 60: Hirose, K. (2005) Migraion and Agglomeraion wih Knowlegde Spillovers. Mimeo. Holmes, T. J. (1996) Sep-by-sep Migraion o Efficien Agglomeraions. Federal Reserve Bank of Minneapolis. Research Deparmen Saff Repor 221. Kanbur, R. y Rapopor, H. (2005) Migraion Seleciviy and he Evoluion of Spaial Inequaliy. Journal of Economic Geography 5. Kaz, E. y Sark, O. (1986) Labor migraion and risk aversion in less developed counries. Journal of Labor Economics 4. Krugman, P. (1991) Increasing Reurns and Economic Geography. The Journal of Poliical Economy. Vol. 99. No

31 LaLonde, R. y Topel, R. (1990) The Assimilaion of Immigrans in he U.S. Labor Markes. NBER Working Papers Series. Working Paper No Diciembre de Lauby, J. y Sark O. (1988) Individual migraion as a family sraegy: Young women in he Phillippines. Populaion Sudies 42. Lewis, W. A. (1954) Economic Developmen wih Unlimied Supplies of Labor. The Mancheser School of Economic and Social Sudies 22: Marchiori, L., Maysad, J. F. y Shumacher, I. (2010) Anoher Inconvenien Truh: Climae Change and Migraion in sub-saharan Africa. Mimeo Massey, D. e. Al. (1993) Theories of Inernaional Migraion: A Review and a Appraisal. Populaion and Developmen Review. Vol. 19, No. 3: sep Mendoza, J. E. (2006) Regional macroeconomic deerminans of Meican Migraion. MPRA Paper No Michell, J. y Pain, N. (2003) The deerminans of Inernaional Migraion ino de UK: A Panel Based Modelling Approach. Naional Insiue of Economic and Social Research (mayo 2003) Morei, E. (2010) Local Labor Markes. NBER Working Paper Oaviano, G. y Thisse, J. F. (2004) Micro-Foundaions of Urban Agglomeraion Economies. En Handbook of Regional and Urban Economics. Volumen 4. Capíulo 58. pp Elsevier. Passel, J. S. y Suro R. (2005) Rise, peak, and decline: Trends in U.S. immigraion , Pew Hispanic Cener, Repor, sepiembre 27 de Paridge, M. y Rickman, D. (2003) The Waing and Waning of US Regional Economies: The chicken-egg of Jobs versus People. Journal of Urban Economics 53. (2010) The Duelling Models: NEG vs Ameniy Migraion in eplaining US Engines of Growh. Papers in Regional Science, Volumen 89 Número 3, agoso de Quinana Romero, Luis y Miguel Ángel Mendoza (2008) Economería Básica Modelos y Aplicaciones a la Economía Meicana, ed. Plaza y Valdés. Méico. Ramsey, F. (1928) A Mahemaical Theory of Saving. Economic Journal. No. 38. Ranis, G. y Fei, J.C.H. (1961) A Theory of Economic Developmen. American Economic Review 51:

32 Rosenhal, S. S. y Srange, W. C. (2004) Evidence on he Naure and Sources of Agglomeraion Economies. En Handbook of Regional and Urban Economics. Volumen 4. Capíulo 49. pp Elsevier. Roe, R. y Vogler, M. (1998) Deerminans of Inernaional Migraion: Empirical Evidence for Migraion from Developing Counries o Germany. IZA Discussion Paper No. 12. Junio Roy, A. D. (1951) Some Thoughs on he Disribuion of Earnings. Oford Economic Papers (3). Russek, S. (2009) Diferenial labor mobiliy, agglomeraion, and skill-biased migraion policies. BGPE discussion paper No. 72. Sanchis-Guarner, R. y López-Bazo, E. (2006) Are Skilled Workers More Araced o Economic Agglomeraion? AQR Research Group IREA, Universia de Barcelona. Preeliminary Draf. Sjaasad, L. A. (1962) The Coss and Reurns of Human Migraion. The Journal of Poliical Economy, Vol. 70, No. 5. Par 2: Invesmen in Human Beings (Oc. 1962) Solow, R. (1956) A Conribuion o he Theory of Ecomomic Growh. Quarerly Journal of Economics. No. 70. Sark, O. y Levhari, D. (1982) On Migraion and risk un LCDs. Economic Developmen and Culural Change 31. Sorper, M. y Sco, A. (2009) Rehinking Human Capial, Creaiviy and Human Growh. Journal of Economic Geography 9. y Taylor, E. (1991) Migraion Incenives, Migraion Types: The Role of Relaive Deprivaion. The Economic Journal. Vol. 101, No (sep 1991) Tabuchi, T. y Thisse, J. F. (2002) Tase Heerogeneiy, Labour Mobiliy and Economic Geography. Journal of Developmen Economcs, 69. Thissen, M, y Frank V. O. (2001) Labour migraion in Europe and he New Economic Geograph. Neherlands Insiue for Spaial Research (RPB). 31