Massey 2012. Origins, Destinatios and Paths of Mexican Migration

8/13/2019 Massey 2012. Origins, Destinatios and Paths of Mexican Migration

1/34

Pathways to El Norte: Origins,

Destinations, and Characteristics of Mexican Migrants to the United States 1

Fernando Riosmena University of Colorado

Douglas S. Massey Princeton University

The geography Mexican migration to the U.S. has experienceddeep transformations in both its origin composition and the desti-nations chosen by migrants. To date, however, we know littleabout how shifting migrant origins and destinations may be linkedto each another geographically and, ultimately, structurally as rela-tively similar brands of economic restructuring have been posited

to drive the shifts in origins and destinations. In this paper, wedescribe how old and new migrant networks have combined to fuelthe well-documented geographic expansion of Mexican migration. We use data from the 2006 Mexican National Survey of Popula-tion Dynamics, a nationally representative survey that for the rsttime collected information on U.S. state of destination for allhousehold members who had been to the U.S. during the 5 yearsprior to the survey. We nd that the growth in immigration tosouthern and eastern states is disproportionately fueled by undocu-

mented migration from non-traditional origin regions located inCentral and Southeastern Mexico and from rural areas in particu-lar. We argue that economic restructuring in the U.S. and Mexicohad profound consequences not only for the magnitude but alsofor the geography of Mexican migration, opening up new region-to-region ows.

1 We thank Luis Felipe Ramos Martnez and Jorge Valencia for guidance with the use of the ENADID data and Nancy D. Mann for her careful editing and suggestions. Weacknowledge administrative and computing support from the NICHD-funded University of Colorado Population Center (grant R21 HD51146).

2012 by the Center for Migration Studies of New York. All rights reserved.DOI: 10.1111/j.1747-7379.2012.00879.x

IMR Volume 46 Number 1 (Spring 2012):336 3


2/34

INTRODUCTION

The geography of MexicoU.S. migration has experienced deep transformationsin recent decades as migrant origins and destinations have diversied away from their traditional Mexican sources and U.S. gateways. The heartland foremigration to the U.S. has historically been Mexicos West-Central Region,principally localities in the states of Guanajuato, Jalisco, Michoaca n, SanLuis Potos, and Zacatecas. Since early in the 20th century, these states haveaccounted for a majority of all emigrants to the U.S. (Durand, Massey, andZenteno, 2001). Over the past three decades, however, new sending regionshave slowly emerged (Durand and Massey, 2003), particularly in areas southand east of Mexico city (e.g., Cohen, 2004; Smith, 2006; Rosas, 2008),reducing the importance of the Central-West to less than half of the ow inrecent years (Durand and Massey, 2003).

Within the U.S., the spatial distribution of Mexican immigrants hasalso diversied, but more recently and at a quicker pace. Before the1990s, around 85 percent of all Mexican migrants went to just threestates Texas, Illinois, and California with the latter receiving upwards

of 60 percent all by itself (Massey and Capoferro, 2008). By the late1990s, however, the share of new arrivals settling in California droppedfrom two-thirds to only one-third, with the bulk of the difference being absorbed by emerging and re-emerging states of destination in fourregions: in the Southeast, North Carolina, Georgia, and Florida; in theMidwest, Iowa, and Nebraska; in the West, Arizona, Colorado, andNevada; and the Northeast, the tristate area of New York, New Jersey,and Pennsylvania(Durand, Massey, and Capoferro, 2005; Massey andCapoferro, 2008). These shifts initially occurred as large numbers of immigrants living in traditional gateway states migrated internally to thesedestinations (Malone et al., 2003; Durand, Massey, and Capoferro, 2005;Lichter and Johnson, 2006, 2009; Donato et al., 2007). Thereafter, how-ever, they appear to have been fueled by even more substantial directimmigration from Mexico, as we show and discuss below.

Scholars argue that the changing spatial distribution of migrantorigins has come in response to liberalization-driven economic restructur-ing (Delgado-Wise and Covarrubias, 2007; Ferna ndez-Kelly and Massey,2007; Nevins, 2007; Massey, Kalter, and Pren, 2008) associated with vari-ous shocks to the Mexican political economy (e.g., Lustig, 1990; Hanson,2003; Polaski, 2004; Zepeda, Wise, and Gallagher, 2009). Likewise, the

4 International Migration Review


3/34

shift in the spatial distribution of destinations has been attributed to therestructuring of various industries and their relocation to non-metropoli-

tan places in the Southeast and Midwest (Herna ndez-Leon and Zuniga,2001; Grifth, 2005; Kandel and Parrado, 2005).

To date, however, we know little about how shifting migrant originsand destinations may be linked to each another geographically and, ulti-mately, structurally (for an exception we discuss below, see Massey, Rugh,and Pren, 2010). In particular, we do not know the extent to whicheconomic restructuring south of the border may have accentuated the shiftin destinations north of the border. In this paper, we describe how old

and new migrant networks have combined to fuel the well-documentedgeographic expansion of Mexican migration. We use data from the 2006Mexican National Survey of Population Dynamics, a nationally represen-tative survey that for the rst time collected information on the U.S. stateof destination of all household members who had been to the U.S. during the 5 years prior to the survey. We nd that the growth in immigrationto the southern and eastern U.S. is disproportionately fueled by undocu-mented migration from non-traditional origin regions located in Centraland Southeastern Mexico and from rural areas in particular. We arguethat economic restructuring in the U.S. and Mexico had profound conse-quences not only for the magnitude but also for the geography of Mexi-can migration, opening up new region-to-region ows.

PREVIOUS STUDIES

New Destinations, Industrial Restructuring, and Declining Gateways

The spatial distribution of the population of the U.S. has changed signi-cantly over the past two decades, as conrmed by the 2010 census andthe accompanying shifts in congressional reapportionment.2 AlthoughHispanics, in particular, and immigrants, in general, have been an integralpart of this geographic diversication (Singer, 2004; Goodwin-White,2007; Massey, 2008; Lichter and Johnson, 2009), shifts have been partic-ularly marked in the case of Mexican immigrants. In just a few years,Mexican migration previously directed almost entirely to the Southwestand the Chicago Metropolitan Area has become a national phenomenon

2 See . Last accessed December28, 2010.

Pathways to El Norte 5


4/34

(Malone et al., 2003; Durand, Massey, and Capoferro, 2005; Donatoet al., 2007). In contrast, spatial dispersion away from traditional gatewayshas been more limited for other immigrant groups.

The shift in the spatial distribution of migrant destinations has beenattributed to the restructuring of various industries, mainly food process-ing (Herna ndez-Leon and Zuniga, 2001; Grifth, 2005; Kandel andParrado, 2005). While some (e.g., poultry packing, farming) were already rmly established in the new areas, others (e.g., beef packing) increasingly moved into them during the 1980s and 1990s (Kandel and Parrado,2005). Both sets of companies hired a large number of immigrants,

mostly of Hispanic descent, recruiting workers both from elsewhere in theU.S. and directly from the country of origin (Krissman et al., 2000; John-sonWebb, 2002).

In addition to the aforementioned geographic shifts in labordemand, two other phenomena seemed to have motivated migrants tomove away from traditional gateways. First, industrial restructuring took place in the wake of a massive regularization of undocumented migrantsunder the Immigration Reform and Control Act (IRCA). Legal statusacquired in the early 1990s provided former undocumented migrants witha new freedom to move and pursue job opportunities elsewhere. Withnewly acquired papers, migrants were no longer tied to labor marketswhere they knew they could work without documents (Neuman andTienda, 1994; Durand, Massey, and Parrado, 1999; Durand, Massey, andCapoferro, 2005). Second, living conditions deteriorated in many tradi-tional gateways, prompting immigrants to look elsewhere to escape expen-sive rents, crowded housing, failing schools, violent neighborhoods, andrising anti-immigrant sentiment (e.g., Hernandez-Leon and Zuniga, 2000;Herna ndez-Leon and Zuniga, 2001; Fennelly, 2005). Light (2006) arguesthat local policies in cities like Los Angeles also created harsher living conditions for low-income immigrants.

Internal Redistribution versus Direct Immigration

Although internal migration was initially responsible for the changing spatial distribution of Mexican migrants in the U.S. (see also Malone

et al., 2003; Durand, Massey, and Capoferro, 2005; Lichter and Johnson,2006, 2009; Donato et al., 2007), it has more recently been fueled by direct immigration from Mexico. Using data from the Integrated PublicUse Micro Samples from the 1990 and 2000 census (IPUMS-USA, publicly



5/34

available through the University of Minnesota, Ruggles et al., 2010),Table 1 shows the place of previous residence for Mexican immigrants

aged 15+ who reported moving in the 5 years prior to the census date.In this table, we dene traditional, re-emerging, and new destinations

following Singers (2004) depiction of metro-area immigrant gateways,3

basing our classication on historical trends in census data reported by Du-rand, Massey, and Capoferro (2005) and Massey and Capoferro (2008). Although not shown in this table, in subsequent tables we subdivide thesethree categories regionally following the scheme used by Durand and Mas-sey (2003). Thus, Traditional Gateways include those states with a long his-

tory of Mexican migration and the largest inows of migrants on a continuing basis and are composed of two geographic subregions: the Bor-derlands (California, Arizona, New Mexico, and Texas) and the Great LakesRegion, which incorporates states in the greater Chicago area (Indiana, Illi-nois, and Wisconsin), along with Michigan and Ohio.

In contrast, Re-emerging Destinations include places in the North-west and Great Plains that received non-trivial numbers of immigrantsduring the earlier Enganche Era (19001929) or the Bracero Period(19421964), but which declined as destinations during the Undocu-mented Era (19651985). In this category, we also include regions thatexperienced signicant migration during the rst years of the Undocu-mented Era but stagnated for a time before reviving in the late 1980s and1990s. The Northwest Region includes the states of Washington, Oregon,Idaho, Nevada, and Utah. Although Durand and Massey (2003) distin-guished the Upper Plains (Montana, Wyoming, the Dakotas, and Minne-sota) from the Lower Plains (Colorado, Nebraska, Kansas, Oklahoma,Missouri, and Iowa), here we collapse these two groups of states into a single category, the Great Plains Region, for reasons of statistical power.Examples of states that received immigrants during earlier eras but

3 Singer classied metro gateways into various types according to the evolution of theirshare of foreign-born in every decennial census in the 20th century. Re-emerging gatewaysare those places with an above-average share of foreign-born individuals in 19001930,lagging below the national average until 1980, after which it increased rapidly (Singer,2004: 5). Given that we are dealing with much larger regions, our classication is lessstrict and more informal, but follows the same spirit. Re-emerging regions are those inwhich Mexican migration had been sizable before the 1980s (mostly during the Enganche or Bracero eras of 19001929 and 19421964), later to diminish in relevance relative toimmigration into traditional gateways, and which resurged during the 1980s and 1990s(Durand, Massey, and Capoferro, 2005).



6/34

declined during the Undocumented Era only to resurge in the 1990s areColorado, Kansas, Oklahoma, and Oregon (see Durand, Massey, andCapoferro, 2005), whereas the Yakima Valley in Washington received sig-nicant immigration early in the Undocumented Era that faltered only torevive again in the 1990s (see Darian, 2006).

We characterize states in the rest of the continental U.S. as New Desti-nations, which we divide into three regions located east of the Mississippi inwhich Mexican migration was by and in large non-existent or very small

before the 1980s (with the notable exceptions of Florida and New York). TheSoutheast Region is composed of Maryland, the District of Columbia, theVirginias, the Carolinas, Georgia, and Florida. The South includes Arkansas,Louisiana, Mississippi, Alabama, Tennessee, and Kentucky, and nally, theNortheast spans Maine, Vermont, New Hampshire, Massachusetts, RhodeIsland, Connecticut, New York, New Jersey, Pennsylvania, and Delaware.

Table 1 shows the changing distribution of stocks and recent streamsof Mexican immigrants among traditional, re-emerging, and new destina-tions, as well as the origins of the streams within or outside the U.S. Thepercentage of Mexican-born individuals (ages 15+) living in traditionalgateways diminished between 1990 and 2000, going from 93.4 to 84.3percent. This change occurred because of a dramatic increase in rates of

TABLE 1R EGIONAL STOCKS , FLOWS, AND ORIGIN OF INTER -REGIONAL MIGRANTS, FOREIGN-BORN MEXICANS AGES

15+

Distributionof stock,

census year

In-migrationrate in

5 years priorto census

Percent of thosearriving 5 yearsprior to census

from

Percent of inter-regional

migrantscoming from

AbroadOther

U.S. region Traditional Re-emerging New A. 1990 census

Traditional 93.4 3.8 92.9 7.1 N A 61.0 39.0Re-emerging 3.8 2.0 60.2 39.8 96.2 N A 3.8New 2.8 3.8 62.2 37.8 95.5 4.5 N A

B. 2000 censusTraditional 84.3 3.6 89.9 10.1 N A 49.9 50.1Re-emerging 7.5 8.9 63.4 36.6 93.4 N A 6.6New 8.2 10.5 77.3 22.7 91.4 8.6 N A Source: Weighed estimates based on 1990 and 2000 census data, obtained from IPUMS-USA 1990, 2000 (Ruggles

et al. 2010).Notes: Traditional gateways comprise the Borderlands and the Great Lakes regions.

Re-emerging gateways include the Great Plains and the Pacic Northwest regions.New destinations include the South, Southeast, and Northeast regions.See text for denition of regions



7/34

in-migration to re-emerging and new destination states and not becauseof a substantial decrease in the rate of in-migration to traditional gate-

ways. Whereas the in-migration rate (i.e., the annualized ratio of Mexi-can-born individuals moving into the region in the 5 years prior to thecensus divided by the number estimated to live there at the beginning of the period) decreased only slightly in traditional gateways (going from 3.8to 3.6%), it more than quadrupled in re-emerging gateways (going from2.0 to 8.9%) and almost tripled in new destinations, reaching very highlevels (rising from 3.8 to 10.5%).

A non-trivial number of migrants who arrived in re-emerging

and to a lesser extent new destinations during the late 1980s and late1990s came from locations within the U.S., mostly from traditional gate-ways. In Re-emerging Destinations, 39.8 and 36.6 percent of recent arriv-als came from within the U.S. during 198590 and 19952000,respectively. In contrast, 92.9 and 89.9 percent of arrivals to gateway states in the same periods came from abroad. In new destinations,meanwhile, the percentage of new arrivals coming from within the U.S.was 37.8 percent in 198590 and 22.7 percent in 19952000. Thedecline in this percentage does not mean that internal migration to new areas diminished in absolute terms; rather, direct immigration fromMexico grew at a faster rate and thus fueled most of the growth of Mexican-born communities east of the Mississippi.

The importance of direct immigration from Mexico is not surprising in light of its rapid growth during the late 1990s and most of the rstdecade of the 21st century (Bean et al., 2001; Passel and Cohn, 2008).Given the rising tide of immigration to new destinations from abroad, weargue that it is necessary to understand the changing distribution of geo-graphic origins in Mexico before speculating about the degree to whichthe shift to new destinations might stem from IRCAs legalization, deteri-orating conditions in gateways, or U.S. industrial restructuring versuschanging conditions in Mexico. Linking specic regional origins to spe-cic regional destinations will enable us to evaluate the relative contribu-tions of transformations north and south of the border in producing thenew geography of Mexican immigration.

New Origins, Mexican Restructuring, and Migrant Networks As mentioned at the outset, the spatial distribution of Mexican migrantsending regions has also shifted in recent times away from its traditional



8/34

heartland in the Central-West Region, which Durand and Massey (2003)call the Historical Region. This region, which encompasses the states of

Durango, Zacatecas, San Luis Potos, Aguascalientes, Guanajuato, Jalisco,Nayarit, Colima, and Michoaca n from the 1920s until recently, has com-prised at least 50 percent of the total outow to the U.S. The sharepeaked between 60 and 70 percent during the 1970s, however, and sincethen has declined steadily to reach levels just below 50 percent early inthe new century (Durand and Massey, 2003) and more recently tolevels just below 40 percent (see Table 2).

Until the mid-1990s, the second most important sending region was

the Border Region, which includes the states of Baja California, Baja

TABLE 2CHARACTERISTICS OF MIGRANTS TO THE US AND NON -MIGRANTS AGES 15 + SURVEYED IN MEXICO

Characteristic Non-migrants US Migrants Age 38.4 44.5Female 52.4% 40.1%Educational attainment

No education 8.4% 11.6%Incomplete primary 16.1 29.7Complete primary 17.4 20.8Some lower secondary 29.1 22.7Some upper secondary 28.9 15.1

Marital statusSingle 31.3% 7.7%Currently married 58.5 78.1Previously married 10.3 14.2

Position in householdHead 35.9% 81.3%Spouse 25.8 11.1Child 28.5 5.4Other 9.8 2.2

Urban level of locality of residenceMetro area (population > 100,000) 51.4% 30.0%Medium urban (15,000 -100,000) 13.8 14.2Small urban (2,500 -15,000) 13.1 15.5 Rural area (population < 2,500) 21.7 40.3

RegionHistorical 22.2% 38.2%Border 21.2 15.4Central 40.4 33.4Southeast 16.2 13.0 N 98,486 2,477

Notes: Estimates adjusted for complex sampling designBold indicates that mean is signicantly different from that of non-migrants ( p


9/34

California Sur, Sonora, Sinaloa, Chihuahua, Coahuila, Nuevo Leon, andTamaulipas. According to Durand and Massey (2003), between 20 and

28 percent of migrants historically came from these eight states, butafter the mid-1990s, this share fell below 20 percent and then dipped to11 percent by 2000 as the region came to house the most rapidly grow-ing sector of the Mexican economy (Hanson, 2003). Rather than shed-ding migrants, the Border Region now attracts a large number of internal migrants (Lozano-Ascencio, Roberts, and Bean, 1999) and,despite a long history of cross-border movement, residents of bordercommunities are now less likely to migrate to the U.S. than interior res-

idents, especially once step-migration to the border is taken into account(Lozano-Ascencio, Roberts, and Bean, 1999; Fussell, 2004; Fussell andMassey, 2004).

As the Border and Historical Regions saw their share of migrantsdecline in recent decades, two sending regions have grown in impor-tance: the Central Region, which includes the states of Queretaro, Mex-ico, Distrito Federal, Hidalgo, Tlaxcala, Morelos, Puebla, Guerrero, andOaxaca, and the Southeastern Region, which includes Veracruz,Tabasco, Chiapas, Campeche, Quintana Roo, and Yucata n. The CentralRegion was relatively unimportant until 1980, accounting for no morethan 10 percent of migrants to the U.S., but rose steadily thereafter toreach just over 30 percent by centurys end (Durand and Massey,2003). Mexicos Southeastern Region remained insignicant as a migra-tion source until recently, contributing fewer than 2 percent of migrantsthrough the early 1990s (Durand and Massey, 2003). By the end of the millennium, migrants from this region composed 7 percent of thetotal (Durand and Massey, 2003). More recently, this gure has goneup to 13 percent (see Table 2).

Scholars have argued that these transformations were triggered by economic restructuring and shocks to the Mexican political economy beginning in the 1980s (Ferna ndez-Kelly and Massey, 2007; Nevins,2007; Massey, Kalter, and Pren, 2008). These transformations dispro-portionately affected the countryside, especially rural areas in the south-ern part of the country (Hanson, 2003; Polaski, 2004; Zepeda, Wise,and Gallagher, 2009). Mexicos political economy steadily liberalized

following its entry into the General Agreement on Tariffs and Tradein 1985, its adoption of the North American Free Trade Agreement(NAFTA) in 1993, and its accession to the World Trade Organizationin 1995. Economic liberalization increased job opportunities in export



10/34

manufacturing plants located overwhelmingly north of Mexico City,especially in the Border Region; but it also brought about a concomi-

tant loss of primary sector jobs, mostly in the south (Hanson, 2003;Polaski, 2004; Zepeda, Wise, and Gallagher, 2009), where open com-modity markets caused prices for coffee and corn to decline (Nevins,2007; Zepeda, Wise, and Gallagher, 2009), rendering many small-scaleproducers incapable of competing and exacerbating their dependenceon subsistence agriculture.

These transformations, along with a series of economic crises thataficted the country in 1982, 1988, 1994, and 2009, kept the purchasing

power of Mexicans at or below 1980 levels (cfr. Lustig, 1990), broadenedthe income gap between northern and southern regions of the country (Hanson, 2003; Borraz and Lopez-Cordova, 2007; Jensen and Rosas,2007; Chiquiar, 2008), and stimulated a series of adaptation strategies by the working poor (De la Rocha, 1994), including international migration(Nevins, 2007). If the shift away from immigrant gateways toward new destinations was fueled by U.S. industrial restructuring, the unleashing of internal mobility through IRCAs legalization, and deteriorating living conditions in traditional destinations, one would expect that migrantsfrom Central-Western Mexico would continue to dominate the streaminto new destinations, given their historical dominance of migration fromMexico and their well-established social networks (Massey, Goldring, andDurand, 1994; Massey and Espinosa, 1997; Palloni et al., 2001; Curranand Rivero-Fuentes, 2003; Flores, Herna ndez-Leon, and Massey, 2004;Flores, 2005; Lindstrom and Lopez-Ramrez, 2010; Massey and Riosme-na, 2010) (but see Fussell and Massey, 2004; Menjvar, 2000). If, how-ever, economic restructuring in Mexico was more relevant role inexplaining the shift in destinations, then we might expect migrants fromnon-traditional sending areas to be over-represented in the composition of streams to new destinations.

Thus far, the geographic diversication of immigrant origins and thechanging spatial distribution of migrant destinations have been separately studied using census and survey data from U.S. and Mexican sources,respectively. In Mexico, the Decennial Census and the Encuesta Nacional de la Dina mica Demogra ca (ENADID) traditionally asked which

household members had been to the U.S. during the preceding 5 years.Tabulating the number of people so identied by state offers a snapshotof the geographic origins of Mexican immigration, and classifying themby individual, family, and household characteristics yields a socioeconomic



11/34

prole of migrants to the U.S. (see Massey and Zenteno, 2000; Durand,Massey, and Zenteno, 2001; Marcelli and Cornelius, 2001). In the U.S.,

the Decennial Census and Current Population Survey historically haveasked place of birth, enabling social scientists to examine the distributionof Mexican-born persons by state and region in the U.S. and to study their socioeconomic characteristics (see Durand, Massey, and Capoferro,2005; Massey and Capoferro, 2008).

Until recently, however, there was no way to link information onplaces of origin with information on places of destination, at least using nationally representative surveys. From the U.S. census and current popu-

lation survey, for example, we knew where Mexican migrants lived in theU.S. but not where they came from in Mexico. Likewise, the Mexicancensus and ENADID told us where people with U.S. experience lived inMexico, but not where they came from in the U.S. It was thus impossibleto identify specic place-to-place streams not even their size, much lessthe characteristics of the immigrants involved and how they might differfrom place to place.

This situation has changed with the appearance of two new data sources.4 The rst, Mexicos Matr cula Consular Program, provides a reg-istration document to nationals who register with Mexican Consulatesabroad. Issued since 1871, the document was originally a paper certicatethat recorded the foreign address of the bearer, but in response to the ris-ing number of undocumented migrants living north of the border andthe growing repression directed against them, beginning in March 2002

4 There is of course a plethora of community studies, many of them in Zuniga andHerna ndez-Leon (2005) and Smith and Furuseth (2006) in which oftentimes the place orstate of origin (as a proxy for the age of the migrant network) in Mexico is explored. Among others, scholars have studied destinations in New York (mostly with Pueblans, seeSmith, 2005, 2006), Pennsylvania (Kenneth Square, see Shutika, 2005), North Carolina (Grifth, 2005; Parrado and Flippen, 2005), Northern Georgia (e.g., Dalton, seeHernandez-Leon and Zuniga, 2005; Herna ndez-Leon and Zuniga, 2001), Iowa (Grey and Woodrick, 2005), Nebraska (Gouveia, Carranza, and Cogua, 2005), Maryland (Dunn, Aragones, and Shivers, 2005), Kentucky (Rich and Miranda, 2005), and Louisiana (Donato, Stainback, and Bankston, 2005). However, the main goal of these studies has by and large been to understand the perils and mechanisms of community formation andinter-ethnic relations in these places while emphasizing the context of reception migrantsface and the role of restructuring in bringing certain jobs to these areas (also see Murphy,Blanchard, and Hill, 2001; Smith and Furuseth, 2006).



12/34

the ID card underwent major changes and was replaced with a high secu-rity, wallet-sized card that uses the same security standards as the Mexican

passport, thereby enabling its use for identication purposes in the U.S.(Instituto de los Mexicanos en el Exterior 2004).

By the end of 2006, some 5.7 million consular IDs had beenissued, and the Instituto de los Mexicanos en el Exterior prepared a sampleof nearly one million registrants. These data enable one to cross-classify state and region of origin in Mexico with state and city of destinationin the U.S. to identify specic place-to-place networks, which Massey,Rugh, and Pren (2010) have analyzed. They found that by 2006, migra-

tion to the U.S. had moved well beyond its historical origins in West-Central Mexico and had come to embrace sending communities in theCentral, Southeast, and Border Regions. Although Massey et al. foundtraditional gateways continued to dominate destinations in the U.S., a variety of newer gateways had emerged in the south and Midwest by 2006, and California had lost its overwhelming dominance as a favoreddestination.

The Matr cula data, however, have coverage and depth limitations. Although the data are available for a very large number of people, they only cover undocumented residents of the U.S. and are not necessarily even representative of that population because the database only includesthose who voluntarily came forward to register for the consular ID, whichpresumably includes people who are better established as U.S. residents.In addition, the Matr cula data only offer aggregate tabulations for a lim-ited set of characteristics, making it impossible to study characteristics andbehaviors at the individual level.

DATA

We use a second recent source of data, Mexicos 2006 round of theENADID, the National Survey of Population Dynamics. The 2006 ENADIDfor the rst time asked household members with U.S. experience wherethey had gone in the U.S., which also enables the identication of specicstreams. We build on the earlier work of Massey, Rugh, and Pren (2010)and use data from ENADID to characterize the size and composition of

specic migratory streams. Although the ENADID data offer a muchsmaller number of cases than the Matr cula Consular data, the survey isbased on a probability sample that yields a nationwide snapshot of allhouseholds with a member who had migrated to the U.S. during



13/34

the 5 years preceding the survey. The survey captures both legal andundocumented migrants and includes them even if they were still in the

U.S. as long as someone in the household remained behind to reporton them. Moreover, the data are available at the individual level, so thatmultivariate models of migrant decision-making can be estimated; though,explanatory variables are limited to the standard socio-demographic indi-cators included on the ENADID and exclude many factors specically rel-evant to migration decisions, such as access to social capital throughmigrant network connections. Further, the data do not include informa-tion on the locality or municipality where the interview was conducted

although the ruralurban classication of the locality of residence wasincluded in the database.The ENADID is based on a stratied, multi-stage probability sam-

ple of 41,926 dwellings. The sample was designed to obtain informationrepresentative of the nation, states, metropolitan areas, mid-sized urbanareas, and rural areas (INSP, 2008). The ENADID is a semi-continuoussurvey of repeated cross-sections, the rst of which was carried out in1992 by the Mexican Statistical Ofce (Instituto Nacional de Estad stica,Geograf a e Informa tica , or INEGI). Its purpose was to obtain informationon a variety of demographic processes, including mortality, fertility, andreproductive health, as well as internal and international migration. Subse-quent cross-sections of the survey were implemented in 1997 and again in2006 (for other migration studies using previous cross-sections of thesedata, see Canales, 1999; Durand and Massey, 2003; Durand, Massey, andZenteno, 2001; Massey and Zenteno, 2000; Zenteno and Douglas,1999).

Within each household surveyed by the ENADID in 2006, eld-workers enumerated all persons who normally resided there,

5

even if tem-porarily absent, and obtained basic socio-demographic information forthese individuals. In the section on international migration, the survey

5 According to the interviewers manual, a person normally residing in the household isthat who generally sleeps, prepares his her meals, and obtains shelter in the intervieweddwelling p. 8 and 49, available at last accessed January 13, 2012). Both the house-hold roster and the international migration section include a lter question to verify thealleged household member (and in the case of the latter, the recent U.S. migrant) lived inthe household prior to emigrating (see questions 3.3 and 4.5 on pages 9 and 19, availableat , last accessed January 13, 2012).



14/34

asked whether any household member moved to the U.S. between January 2001 and the date of the interview.6 Although this denition does

include people not living in the sampled households at the time of thesurvey (63% of recent migrants were still in the U.S. at the time, seeTable 2 below), this denition clearly does not include members of entirehouseholds that left for the U.S. without returning during the referenceperiod, a population that may be better captured with U.S. data sources.

To minimize potential biases in under-coverage, we focus our analy-ses on household members old enough to engage in independent migra-tion, given that the emigration of children tends to occur in tandem with

that of other household members, making them most likely to be under-estimated by the design of the ENADID. We set a lower bound forinclusion at age 15, thus yielding a sample of 100,963 individuals, 2,477of whom had U.S. migratory experience in the 5 years prior. These2,477 migrants belong to 2,071 households, yielding an average of 1.2migrants per household.7

6 The exact wording in the questionnaire was: De enero de 2001 a la fecha, alguna persona

de este hogar se fue a vivir a los Estados Unidos de Ame rica, aunque haya regresado? (see question 4.2 on page 19, available at , last accessed January 13, 2011). See also interviewers manual(p. 8 and 49, available at last accessed January 13, 2012). This wording suggests thattemporary visitors were not included in the migrant count.7 Note that the percentage of migrants in the sample (a design-adjusted 2.34% over slightly more than 5 years) is low compared to estimates derived from other data sources (cfr. Hilland Wong, 2005; Rubalcava et al., 2008), including the recently released 2009 ENADID,which included similar information to the 2006 ENADID for migrants leaving the inter-viewed household since 2004 and from which we tabulated origin-destination distribu-tions. While indeed the percentage of migrants in the sample is higher in the 2009ENADID (5.7% , partly due to the continued growth of documented and undocumentedmigration from Mexico during most of the decade, see Passel and Cohn, 2008), the regio-nal distributions of origins and destinations were quite consistent with the gures pre-sented here. For instance, the percentage of migrants coming from the Historical (39.5%),Border (16.4%), Central (29.1), and Southeastern (14.8) Regions are very close to (andnot statistically different from) the estimates we present on Table 2. Likewise, the distribu-tion of migrants going toward the Borderlands (54.1%), Great Lakes (7.25%), Northwest(3.3%), Great Plains (2.6%), Southeast (20.7%), South (3.1%), and Northeast (7.1%) areoverall consistent with the estimates presented in the rst Panel on Table 3; though, they also suggest a smooth continuation of the shift toward new destinations. The 2006 resultsthus provide an earlier snapshot of the shift toward new destinations and how they arerelated to the shift in migrant origins as well.



15/34

In addition to identifying recent international migrants, the ENADIDalso collected information about their last U.S. trip, such as the documenta-

tion used by the migrant and their state of destination in the U.S. This lastquestion was not asked in previous versions of the ENADID or in the 2000Census Migration Supplement, making it possible for the rst time to iden-tify specic inter-regional streams from a nationally representative data source. However, given the wording of the state of destination question (towhich U.S. state did the migrant go to during his her last trip?),8 survey respondents could have interpreted this as their main but not rst destina-tion, thus exaggerating the extent of direct immigration from Mexico

toward new destinations. However, we believe our main ndings are notoverly affected by this potential interpretation for two reasons. First, ourestimates in Table 1 above indicate that most migrants into new destina-tions moved directly from Mexico. As such, even if present, the extent of this bias should not be dominating our results. Second, if there was a sizableamount of masked step-migration to traditional gateways and on to new destinations measured by the survey as movement toward new destinationsalone, we would expect that people from the Historical Region would repre-sent migrant streams toward new destinations: on the contrary, we nd thatstreams toward new destinations have a distinct origin composition (e.g., seeFigure 1 below), which suggests the stream toward new destinations is notdominated by the better-established networks of the Historical Region(see also Massey, Rugh, and Pren, 2010: Table 4).

Some 83 percent of migrants (2,061 individuals) provided validinformation on state of destination. Given that this is a relatively smallnumber, we focus on region-to-region rather than place-to-place or state-to-state streams. Although this approach has the obvious drawback of ignoring intra-regional shifts in migrant origins and destinations, the anal-yses stemming from our regional classication provide albeit in broadstrokes a clear picture of the changing geography of Mexican migrationat a relevant scale of analysis.

ORIGINS, DESTINATIONS, AND CHARACTERISTICS OF MIGRANTS

Table 2 presents general socio-demographic characteristics for Mexicanmigrants aged 15+ who left for the U.S. between 2001 and 2005

8 See question 4.11 on p. 21 A que estado de los Estados Unidos se fue la u ltima vez? .



16/34

(N = 2,477) and compares them to non-migrants aged 15+ (N = 98,486). Although the age structure of migrants tends to be quite young, those in

the sample are older, on average, than the general Mexican population(mean age of 44.5 versus 38.4), because the latter still has a young struc-ture and the former includes those making return trips to the U.S. as wellas those going for the rst time. While women are still under-representedin the ow (40 versus 52% in the general population), they represent a larger share than studies have found in the past (cfr. Durand, Massey,and Zenteno, 2001: Table 2). Some female moves are independent of male migration (Kanaiaupuni, 2000), whereas others are tied to family

reunication (Donato, 1993; Cerrutti and Massey, 2001) oftentimes by way of legal migration (Riosmena, 2010). Migrants and non-migrants dif-fer in educational attainment, partly because migrants in the sample areolder. Relative to non-migrants, migrants have higher proportions of per-sons with no schooling (12 versus 8%), incomplete primary education (30versus 16%), and completed elementary education (21 versus 17%), andlower proportions with some lower secondary (23 versus 29%) and someupper secondary (15 versus 29%).

Given differences in age structures, it is not surprising that migrantsare considerably less likely to be single than non-migrants (8 versus 31%)and more likely to be either currently or previously married (78 versus59% and 14 versus 10% respectively). Age, gender, and marital status dif-ferences are reected in the household position of migrants. Householdheads are over-represented in the migrant ow by a large margin (81 ver-sus 36%).

As Durand, Massey, and Zenteno (2001) point out, the growing representation of urban origins among migrants partly stems from theurbanization Mexico has undergone since the 1950s (Garza, 2003), as vil-lages have become towns and towns have become cities. Although themajority of migrants no longer come from rural areas, these are still over-represented in the ow: while only 22% of the Mexican population livesin rural localities, 40% of migrants come from rural places. As shown inthe table, small urban areas (2,50015,000 inhabitants) are also slightly over-represented in the migrant ow (16 versus 13% for non-migrants),whereas medium-sized urban areas (15,000100,000 inhabitants) yield

around 14 percent for both migrants and non-migrants. The proportionof people coming from medium-sized cities has in fact remained strikingly stable since the 1970s, uctuating between 11 and 15 percent (Durand,Massey, and Zenteno, 2001). Although the share of migrants coming



17/34

from metropolitan areas has been on the rise (Durand, Massey, andZenteno, 2001; Herna ndez-Leon, 2008), these areas are still under-represented

in the ow. Whereas more than half of all Mexicans live in metropolitanareas, only 30 percent of U.S. migrants come from such places, resultsthat are consistent with the hypothesis that mechanisms associated withthe cumulative causation of migration (Massey, 1990) do not operate asefciently in large urban settings because of their higher economic dyna-mism and diversity, and greater anonymity and impersonality (Fussell andMassey, 2004).

The regional composition of migration reveals continuities in the

Historical Region and changes elsewhere. Although the HistoricalRegions proportion of the outow has now dropped below 50 percentand stood at just 38 percent in 20012005 (cfr. Durand and Massey,2003; Durand, Massey, and Zenteno, 2001), it is still the largest sourceregion for migration to the U.S. and the only one over-represented rela-tive to its share of the population (only 22% of Mexicans aged 15 andover live in this region). In contrast, 15 percent of recent migrants comefrom the Border Region, which is slightly under-represented relative to itsshare of the total population (21%). The same is true for the Central andSoutheastern Regions, which compose 33 and 13 percent of the migrantow, respectively, compared with 40 and 16 percent of the Mexican pop-ulation. These gures are generally consistent with those derived fromprevious studies (cfr. Durand, Massey, and Zenteno, 2001); though, they also suggest that migrants from less traditional sending regions are joining the ow at very high rates: the share originating in the Southeast, forexample, increased from 7 percent during 19952000 (Durand andMassey, 2003: Table 13) to the estimated 13 percent during 20012005.

Figure 1 and the rst line on Table 3 show the distribution of migrants according to region of destination. Although traditional gatewaysstill dominate direct streams from Mexico (streams toward the Borderlandand Great Lakes Region compose two-thirds of the total ow), the shifttoward new destinations is indeed quite notable: 15.5, 3.9, and 7.3 per-cent of migrants go to the Southeast, South, and Northeast, respectively,making up more than a quarter of the ow, a gure already higher thanthe late 1990s estimates reported by Durand, Massey, and Capoferro

(2005) and Massey and Capoferro (2008).Table 3 also shows the distribution of the region and level of urban-ization of the place of origin of migrants according to their region of des-tination in the U.S. The changing regional composition of Mexican



18/34

origins is contributing more than its share to the eastward expansion of immigrant destinations documented in the second section of the paper. Although the re-emergence of destinations in the West and Midwestindeed seems to be driven largely by networks emanating from the Histor-ical Region, the eastern expansion of Mexican migrants in the U.S. is not.First, the Historical Region is always over-represented in streams to tradi-tional destinations: its residents compose 38 percent of the total ow, but

43 and 48 percent of migrants going into the Borderland and Great LakesRegions. In contrast, migrants from the Historical Region represent only 30, 27, and 13 percent of migrants going to the Southeast, South, andNortheast of the U.S. The striking under-representation of the HistoricalRegion in ows to the Northeast and to a lesser extent, South andSoutheast suggest that old networks from Central-Western Mexico donot operate in the same way to guide migrants to new destinations as they do to channel migrants to communities in traditional and re-emerging

gateways.The concentration of people from Mexican Border states in the U.S.Borderland Region is also evident from Table 3. Migrants from the Mexi-can Border Region are over-represented in ows to U.S. states just across

TABLE 3MIGRANT REGIONAL AND URBAN ORIGINS BY REGION OF DESTINATION IN THE US

Region of destination in the US

Total

Traditional Re-emerging New

Border-landGreatLakes North-west

GreatPlains South-east South North-east

Share in eachdestination

57.5% 8.5% 3.4% 3.9% 15.5% 3.9% 7.3% 100.0%

Characteristics of region place of origin by region of destinationRegion of origin in Mexico

Historical 43.2% 47.9% 44.1% 47.2% 29.2% 26.9% 12.9% 38.2%Border 21.2 9.4 7.0 25.5 3.1 15.3 4.9 15.4Central 26.7 34.4 36.8 18.8 41.8 22.7 72.3 33.4

Southeast 8.9 8.3 12.1 8.4 25.8 35.1 9.9 13.0Total 100.0 100.0 100.0 99.9 99.9 100.0 100.0 100.0Urban-rural classication of locality of origin in Mexico

Metro area 35.2% 18.7% 39.3% 34.2% 20.1% 10.8% 28.6% 30.0%Medium urban 15.2 20.5 17.1 7.1 8.8 21.1 16.7 14.2Small urban 16.3 15.2 21.7 13.2 16.0 18.1 16.6 15.5Rural area 33.4 45.7 21.9 45.5 55.2 50.0 38.2 40.3N 1,224 173 76 105 293 67 123 2,477

Notes: Estimates adjusted for complex sampling designBold indicates regional is signicantly different from overall mean ( p


19/34

T A B L E

4

C H A R A C T E R I S T I C S O F M I G R A N T S B Y R E G I O N O

F D E S T I N A T I O N I N T H E

U N I T E D

S T A T E S .

C h a r a c t e r

i s t i c

R e g i o n o

f d e s t i n a t

i o n i n t h e

U S

T o t a l

T r a d i t i o n a l

R e - e m e r g i n g

N e w

B o r d e r - l a n d

G r e a t

L a k e s

N o r t h - w e s t

G r e a t

P l a i n s

S o u t

h - e a s t

S o u t

h

N o r t h - e a s t

L e g a

l s t a t u s a t e m

i g r a t i o n

U n d o c u m e n t e d

6 7 . 1

%

7 7 . 3

%

7 7 . 9

%

5 9 . 5

%

8 6 . 2

%

7 9 . 9

%

8 1 . 0

%

7 3 . 8

%

T o u r

i s t v i s a

2 0 . 2

7 . 9

1 2 . 0

2 0 . 9

8 . 3

4 . 8

5 . 5

1 4 . 8

W o r

k v i s a

r e s i d e n c e

1 2 . 7

1 4 . 8

1 0 . 1

1 9 . 5

5 . 6

1 5 . 3

1 3 . 5

1 1 . 4

P c t . r e t u r n i n g

b y 2 0 0 6

4 2 . 1

%

3 4 . 8

%

3 2 . 5

%

4 3 . 0

%

2 9 . 5

%

3 8 . 0

%

4 1 . 7

%

3 7 . 0

%

A v g . t r i p d u r a t i o n ( m o s . )

6 . 9

6 . 1

5 . 4

8 . 5

4 . 9

8 . 1

8 . 6

5 . 8

A g e

4 4 . 1

4 4 . 6

4 4 . 2

4 6 . 2

4 3 . 6

4 1 . 1

4 6 . 5

4 4 . 5

F e m a l e

3 8 . 9

%

3 5 . 0

%

4 7 . 8

%

3 8 . 3

%

4 3 . 8

%

4 2 . 4

%

3 6 . 3

%

4 0 . 1

%

E d u c a t

i o n a l a t t a i n m e n t

N o e d u c a t i o n

9 . 3 %

9 . 2 %

7 . 7 %

9 . 0 %

1 1 . 9

%

8 . 2 %

1 1 . 5

%

1 1 . 6

%

I n c o m p l e t e p r i m a r y

2 7 . 2

3 5 . 2

2 1 . 6

3 0 . 3

3 9 . 1

3 6 . 4

2 5 . 1

2 9 . 7

C o m p l e t e p r i m a r y

2 2 . 5

1 8 . 6

1 8 . 3

2 1 . 6

2 0 . 9

2 5 . 4

1 8 . 8

2 0 . 8

S o m e

l o w e r s e c o n d a r y

2 4 . 4

1 6 . 3

3 4 . 6

2 7 . 8

1 2 . 0

2 2 . 2

3 1 . 6

2 2 . 7

S o m e u p p e r s e c o n d a r y

1 6 . 6

2 0 . 7

1 7 . 9

1 1 . 3

1 6 . 1

7 . 7

1 3 . 0

1 5 . 1

M a r

i t a l s t a t u s

S i n g l e

8 . 3 %

4 . 5 %

1 1 . 2

%

9 . 3 %

7 . 4 %

2 . 7 %

7 . 1 %

7 . 7 %

M a r r i e d

7 8 . 6

8 2 . 1

7 4 . 1

7 8 . 3

7 9 . 1

8 7 . 2

7 6 . 3

7 8 . 1

P r e v

i o u s l y m a r r i e d

1 3 . 2

1 3 . 3

1 4 . 7

1 2 . 4

1 3 . 5

1 0 . 1

1 6 . 6

1 4 . 2

P o s i t

i o n i n h o u s e

h o l d

H e a

d

8 3 . 8

%

8 3 . 5

%

8 0 . 0

%

7 8 . 8

%

7 8 . 2

%

8 1 . 5

%

8 6 . 8

%

8 1 . 3

%

S p o u s e

9 . 7

1 1 . 9

1 0 . 6

1 1 . 9

1 2 . 4

1 3 . 5

7 . 4

1 1 . 1

C h i l d

4 . 8

1 . 3

9 . 4

4 . 7

7 . 3

2 . 2

5 . 0

5 . 4

O t h e r

1 . 7

3 . 2

0 . 0

4 . 6

2 . 1

2 . 9

0 . 8

2 . 2

N

1 , 2 2 4

1 7 3

7 6

1 0 5

2 9 3

6 7

1 2 3

2 , 4

7 7

N o t e s : E s t i m a t e s a

d j u s t e d f o r c o m p l e x s a m p l i n g

d e s i g n .

B o l d i n d i c a t e s t

h a t m e a n i s s i g n i c a n t l y d i f f e r e n t f r o m o v e r a l l m e a n ( p < . 0 5 ) .

I t a l i c s i n d i c a t e s t

h a t m e a n i s s i g n i c a n t l y d i f f e r e n t f r o m o v e r a l l m e a n ( 0

. 0 5 < p < 0 . 1 0 )

.



20/34

the border (21 versus 15% overall) as well in as the smaller ows going to the Great Plains (26 versus 15% overall). Flows from the Border

Region to the American South are non-trivial at 15 percent, but basically mirror the share of migrants leaving the Border Region overall. In con-trast, the share of migrants from the Border Region is much lower inows going to traditional Great Lakes gateways (9%), and new destinationareas in the Southeast (3%) and Northeast (5%).

The pattern of emigration from Mexicos Central Region most likely reects its combination of immigrants from both older and newer immi-grant origins (e.g., areas around Mexico City and Guerrero versus Puebla

and Queretaro see De Genova, 2005: 34). For instance, the CentralRegion is well represented in traditional gateways in the Great LakesRegion (34 versus 33% overall) and the Northwest (37%), as well as new immigrant destinations in the Southeast (42%) and the Northeast, wherenearly three-fourths of recent migrants arriving directly from Mexicocome from the Central Region. New destinations in the South (26%) andSoutheast (35%) disproportionately receive migrants from SoutheasternMexico, again suggesting that newer sending areas are contributing morethan their fair share to the growth of new immigrant destinations (see alsoFigure 1).

The ruralurban origin of migrants seems to relate in systematicways to the prole of sending regions described above. Migrants going tothe South and Southeast (who tend to work in primary sector industries,see Dunn, Aragones, and Shivers, 2005; Grifth, 2005; Haverluk andTrautman, 2008; Kandel and Parrado, 2004, 2005, 2006; Parrado andKandel, 2008) are least likely to come from a Mexican metropolitan area and disproportionately likely to come from rural areas. In the South andSoutheast, 50 and 55 percent of migrants, respectively, come from ruralareas, while only 11 and 20 percent are of metropolitan origin. In con-trast, the proportion of migrants coming from metropolitan areas is high-est in the more established ows going to the U.S. Borderland Region( p < 0.10 relative to the national average), whereas the proportion coming from rural areas is lowest in this region. The relative number of migrantswith rural origins is also quite low in the re-emerging areas of the North-west, a fact that might reect the historical transformation of migration in

traditional gateways away from seasonal agricultural labor toward work inurban services (e.g., Cornelius, 1992; Durand, Massey, and Parrado,1999; Riosmena, 2004). Interestingly, although migrants going to theGreat Lakes Region have traditionally worked in the industrial and urban



21/34

sectors (De Genova, 2005), they do not come from places that are partic-ularly urbanized even relative to the general migrant ow, except for a

slight over-representation of migrants from urban areas with 15,000100,000 inhabitants, at 21 percent (versus 14%).

Table 4 shows individual-level characteristics of 20012005 migrantsaccording to their region of destination in the U.S., including documenta-tion and return behavior as well as socio-demographic traits (the right-most panel shows the total for all migrants regardless of region of destination). The proportion undocumented is somewhat higher among migrants going to new destinations in the South and Northeast (80% and

above) compared with the Borderland and Great Plains, where the pro-portions are 60 and 67 percent, respectively. The low Borderlands gurereects higher proportions using tourist visas (20%), perhaps because of the over-representation of migrants from the Mexican Border Region,who have greater access to crossing cards and other visas. The low gurein the Great Plains is attributable to higher proportions of migrants withwork residence permits (20%), perhaps owing to better-established owsof contract labor. Although return rates also vary across destinationregions, it is only in the extremes (and the two largest ows) that thesepatterns are signicantly different from the average for all migrants. Thelikelihood of return is lowest among migrants going to the Southeast(30%) and highest among migrants going to the Borderland Region(42%), where trip durations are nonetheless slightly longer than average(6.9 months versus 5.8 months for all returnees).

The demographic prole of migrants by destination does not vary greatly, certainly not enough to be distinguished from the overall mean inmost instances (and this variation is also much lower than that acrossregion of origin, results not shown). Only for migrants going to the Southis the mean age of migrants slightly lower than average, at 41 years versus44.5 ( p < 0.10). Although the proportion of women is generally lower intraditional destinations (3539%) than in new destinations in the South(4344%), these differences are not statistically signicant. Not surpris-ingly, the educational attainment of migrants reects their urbanruralorigins. The percentage of people with at least some lower secondary edu-cation (the sum of the two highest schooling categories shown in Table 4,

which can be interpreted as being around or above the educational aver-age for the Mexican adult population), is higher in the Northeast(31.6 + 13.0 = 44.6%), the Borderland Region (41%), and the GreatLakes (37%) compared with gures of 28 and 30 percent in the Southeast



22/34

and South, respectively. In all other regions, schooling levels were not sig-nicantly different from the average. Finally, the distribution of maritalstatus is similar across all destinations except the South, where the per-centage of migrants who are currently married is higher than in all otherregions (87% versus an average of 78%).

HOW DO THE DETERMINANTS OF MIGRATION TO

TRADITIONAL AND NEW ORIGINS DIFFER? Table 5 shows odds ratios from multinomial logistic regression modelsestimated to predict migration to traditional and Re-emerging Destina-

TABLE 5ODDS RATIOS FROM MULTINOMIAL LOGISTIC REGRESSION PREDICTING THE PROBABILITY OF MIGRATING TO

TRADITIONAL RENEWED AND NEWER GATEWAYS ( VS. NOT EMIGRATING ).

Predictor To traditional, re-emerging gateways To new destinations

Age 1.02+ 1.05**

Female 2.49*** 2.87***

Education (REF = Some upper secondary)No education 1.19 1.07Incomplete primary 1.55*** 1.56**

Complete primary 1.44*** 1.21Some lower secondary 1.26** 1.06Some upper secondary

Marital status (REF = Currently married)Currently married

Single Previously married 0.46*** 0.50***

Position in the household (REF = Head)Head Spouse 0.07*** 0.06***

Child 0.08*** 0.14 ***

Other 0.06*** 0.08***

Urban-rural origins (REF = Rural area)Metro area 0.46*** 0.22 ***

Medium urban 0.64*** 0.46 ***

Small urban 0.85* 0.65 **

Rural area

Region of origin (REF = Historical)Historical Border 0.40*** 0.22 ***

Central 0.41*** 1.26 *

Southeast 0.20*** 0.61 **

Log-likelihood ) 9,348.4N 96,531

Notes: + p < 0.10; * p


23/34

tions combined versus newer destinations in the Southeast, South, andNortheast, in both cases relative to non-migrants. Signicant differences

between coefcients predicting emigration to traditional and re-emerging versus new destinations are denoted in bold ( p < 0.05) and italics(0.05 < p < 0.10). The socio-demographic determinants of migrating to a new destination are not signicantly different from the determinants of going to traditional or re-emerging destinations, with one exception. Although children of a household head are considerably less likely tomigrate to either type of destination, they are 1.64 times more likely (1.64 = 0.14 0.08, p < 0.1) to move to a new than to traditional or re-

emerging destinations.Unlike the socio-demographic characteristics, rural and regional ori-gins are signicant in predicting migration toward traditional versus new destinations and generally conrm the results of the descriptive statisticspresented above. The ruralurban gradient is sharper in ows directedtoward new destinations than in those going to traditional and re-emerg-ing gateways, even after controlling for region of origin in Mexico. Rela-tive to rural areas, the odds of migration to traditional or re-emerging gateway are 15, 36, and 54 percent lower in smaller urban, larger urban,and metropolitan areas. These gaps are even greater for new destinations,where the odds of emigration are 35, 54, and 78 percent lower in smallerurban, larger urban, and metropolitan areas relative to rural areas. Again,this is not surprising given that the demand for migrant labor inthe American South and Southeast is heavily concentrated in primary sector industries such as food processing. As a result, the expansion of Mexican migration eastward (at least for ows stemming directly fromMexico) retains the relatively rural character Mexican migration has hadhistorically.

However, again, it does not retain the region-of-origin character of previous epochs. The regional pecking order in emigration rates variesconspicuously by region of destination. Residents of the Historical Regionare most likely to migrate to traditional and Re-emerging Destinations,followed by those from the Border, Central, and Southeast Regions, whoare 60, 61, and 80 percent less likely to migrate to traditional orRe-emerging Destinations than residents of the Historical Region, once

we control for other relevant characteristics including the ruralurban clas-sication of the locality of origin. In contrast, people from the CentralRegion are most likely to migrate to new destinations. Indeed, they are1.26 times more likely to do so than residents of the Historical Region.



24/34

In contrast people living in the Border and Southeaster Regions are 78and 39 percent less likely to migrate to new destinations than people liv-

ing in the Historical heartland of Mexican migration to the U.S. Although residents of the Historical and Border Regions are more

likely to go to traditional or re-emerging gateways than to new destina-tions in the East, Southeast, and South, the opposite is true for residentsof the Central and Southeast Regions. Residents of the Border Region,for example, have 44 percent (0.22 0.40, p < 0.01) lower odds of migrat-ing to new destinations than to regions with traditional re-emerging gateways. In contrast, people from Central and Southeast Mexico have

both 3.1 times (i.e., 1.26

0.41 and 0.61

0.20) higher odds of migrating to new destinations than to traditional or re-emerging gateways. It is thusclear that the eastward expansion of Mexican settlement previously docu-mented in the literature has a different origin composition net of otherconfounders, at least among migrants who come directly from Mexico.

DISCUSSION AND CONCLUSION

Our results reveal the divergent regional origins of Mexican migrants tonew destinations in the American South, Southeast, and especially theNortheast. A majority of migrants going to new destinations in the Eastcome not from the Historical Region but from Central and SoutheasternMexico. Migrants from the Central Region dominate ows into theNortheast (72% with an additional 10% coming from the Southeast) andthe Southeast (42%, plus 26% from the Central Region), whereasmigrants from the Southeast dominate migration to the South (35%, inaddition to 23% from Central Mexico, see Table 4). Although migrantsfrom the Historical Region remain a non-trivial portion of the ow going to the Southeast (29%) and South (27%, see Table 3), they are strikingly absent from ows into the Northeast (in our data, mostly comprised of the tristate area of New York, New Jersey, and Pennsylvania), in whichthey represent a meager 13 percent.

It is possible, of course, that the origin composition of current pop-ulation stocks in the new destinations could be less dominated by Centraland Southeastern Mexico because of past migration ows and internal

redistribution of migrants away from traditional gateways; though, thatpossibility remains something of an empirical puzzle given the lack of rel-evant data. Nonetheless, this seems unlikely given how migrant networkstypically operate (Massey, Goldring, and Durand, 1994; Fussell and Mas-



25/34

sey, 2004; Lindstrom and Lopez-Ramrez, 2010) and the rapid growth in

emigration from Mexicos Central and Southeast Regions (cfr. Table 2,Durand and Massey, 2003; Durand, Massey, and Zenteno, 2001), espe-cially given the rapid rise of migration to the U.S. South, Southeast, andNortheast (cfr. Table 4, Durand and Massey, 2003; Durand, Massey, andCapoferro, 2005; Leach and Bean, 2008; Massey and Capoferro, 2008).

More than half of all migrants to the U.S. South and Southeast, inparticular, originate in Mexican rural areas, whereas only a few (1120%)come from metropolitan areas. This pattern not only is consistent with

the distribution of migrant origins favoring ows from the Central andSoutheastern Regions (40 and 67% rural, see Table 2), but remains soeven after we control for the regional composition of ows into new desti-nations (Table 5). In any event, the ruralurban composition of ows

TABLE 6ODDS R ATIOS FROM L OGISTIC R EGRESSIONS PREDICTING THE PROBABILITY OF MIGRATING TO THE U.S.

BY R EGION OF ORIGIN IN MEXICO (SE NOT SHOWN )

PredictorRegion of origin in Mexico

Historical Border Central Southeast National Age 1.03* 0.97 1.05** 1.06* 1.03** Age - squared 0.9997** 1.0002 0.9995** 0.9992** 0.9996***Female 2.35 1.29 2.51*** 4.74*** 2.39***Education (REF = Some upper secondary)

No education 1.19 0.99 1.85** 1.81* 1.43**Incomplete primary 1.65*** 0.87 2.08*** 1.98** 1.65***Complete primary 1.62*** 0.81 1.65** 1.46 1.43***Some lower secondary 1.37 1.06 1.52 0.99 1.30Some upper secondary

Marital status (REF = Currently married)Currently marriedSingle 0.52*** 0.48** 0.49** 0.19*** 0.46***Previously married 0.49*** 0.75 0.57** 0.28*** 0.51***

Position in the household (REF = Head)Spouse 0.08*** 0.11*** 0.08*** 0.04*** 0.08***Child 0.10*** 0.04*** 0.15*** 0.16 0.11***Other 0.11*** 0.05*** 0.05*** 0.06*** 0.08***

Urban-rural origins (REF = Rural area)Metro area 0.44*** 0.70** 0.30*** 0.21*** 0.41***Medium urban 0.54*** 1.16 0.54*** 0.42*** 0.58***Small urban 0.83 1.04 0.64 *** 0.57 ** 0.75***

Region of origin (REF = Historical)Border 0.37***Central 0.58***Southeast 0.29***Log-likelihood ) 4165.0 ) 1525.8 ) 2859.3 ) 1064.8 ) 9699.1

N 28,216 21,839 29,158 17,729 96,942Notes: + p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001.

Bold indicates coefcient is signicantly different from Historical region ( p < 0.05).Italics indicates coefcient is signicantly different from Historical region ( p < 0.10).



26/34

into new destinations in the U.S. South and Southeast is not surprising given the demand for immigrant labor created by meat processing andother agro-industries in non-metro areas in the U.S. South and Midwest(Kandel and Parrado, 2004, 2005, 2006; Dunn, Aragones, and Shivers,2005; Grifth, 2005; Haverluk and Trautman, 2008; Parrado and Kan-del, 2008). As a result, the more rural character of Mexican migrationremains a common feature of ows to these regions, though not to theNortheast, where migrants enter diverse occupations in the urban sectorin cities such as New York (e.g., Smith, 2006) more often than they enterrural occupations in small towns in the tristate area (e.g., Shutika, 2005).

Note that these trends do not imply that Central and SoutheasternMexicans go exclusively to new destinations; as with other migrantstreams, the largest ows from the Central and Southeastern Regions goto the U.S. Borderlands Region (47 and 41%, respectively); though, theseshares are indeed smaller than among those leaving from the Historical(64%) and Border (78%) Regions.9 The Border Region in fact stands out

Borderland 57.5%Historical: 43.2%Border: 21.2%Central: 26.7%Southeast: 8.9%Total: 100%

Great Lakes 8.5%Historical: 47.9%Border: 9.4%Central: 34.4%Southeast: 8.3%Total: 100%

Great Plains 3.9%

Historical: 47.2%Border: 25.5%Central: 18.8%Southeast: 8.4%Total: 100%

Northwest 3.4%

Historical: 44.1%Border: 7.0%Central: 36.8%Southeast: 12.1%Total: 100%

Southeast 15.5%Historical: 29.2%Border: 3.1%Central: 41.8%Southeast: 25.8%Total: 100%

South 3.9%Historical: 26.9%Border: 15.3%Central: 22.7%Southeast: 35.1%Total: 100%

Northeast 7.3%Historical: 12.9%Border: 4.9%Central: 72.3%Southeast: 9.9%Total: 100%

Figure I. Distribution of 20012005 Migrants by Region of Destination in the U.S.and Composition of Region of Origin in Mexico by Region of Destinationin the U.S.

9 However, after controls are introduced, migrants from the Central and Southeast Regionsdo in fact have a larger propensity to emigrate to new than to traditional and re-emerging destinations (see Table 5).



27/34

in various ways from other origin regions. First, it is the only region inwhich a majority of migrants come from metropolitan areas (see Table 2).

Although this predominance almost dees the notion that cumulative cau-sation does not operate in large cities (Fussell and Massey, 2004), noteven in the Border Region do metropolitan areas contribute their fairshare to the migrant ow (results not shown but available from theauthors).

Migrants from the Border Region also have at educationmigra-tion gradients and above-average return rates, and they concentrate heav-ily in the adjacent Borderland Region. The prole of Border migrants

may stem more from the unique history of cross-border movement thanfrom the structural conditions that promote emigration from elsewhere inMexico, particularly the South. Indeed, large disparities in job creationpartially fueled by NAFTA have transformed the Border Region into themost rapidly growing portion of the Mexican in both economic (Hanson,2003) and demographic (Lozano-Ascencio, Roberts, and Bean, 1999)terms. In contrast, the liberalization of the Mexican economy and thegradual opening of agricultural markets under NAFTA have generatedlarge-scale displacement from the countryside, especially in SouthernMexico (Delgado-Wise and Covarrubias, 2007; Ferna ndez-Kelly andMassey, 2007; Nevins, 2007; Massey, Kalter, and Pren, 2008). Thus, therecent growth of Mexican migration and its geographic diversication onboth sides of the border reects a conuence of supply-side and demand-side factors operating in the same direction with a common thread: eco-nomic restructuring.

REFERENCES

Bean, F. D. et al.2001 Circular, Invisible, and Ambiguous Migrants: Components of Difference in Esti-

mates of the Number of Unauthorized Mexican Migrants in the United States.Demography 38(3):411422.

Borraz, F., and J. E. Lopez-Cordova 2007 Has Globalization Deepened Income Inequality in Mexico? Global Economy Jour-

nal 7(1):155.Canales, A.1999 Periodicidad, estacionalidad, duracion y retorno. Los distintos tiempos en la migra-

cion Mexico-Estados Unidos. Papeles De Poblacion 22:1141.Cerrutti, M., and D. S. Massey 2001 On the Auspices of Female Migration From Mexico to the United States. Demog-

raphy 38(2):187200.



28/34

Chiquiar, D.2008 Globalization, Regional Wage Differentials and the Stolper-Samuelson Theorem:

Evidence from Mexico. Journal of International Economics 74(1):7093.Cohen, J. H.2004 The Culture of Migration in Southern Mexico . Austin, TX: University of Texas Press.Cornelius, W. A.1992 From Sojourners to Settlers: The Changing Prole of Mexican Migration to the

United States. In U.S.-Mexico Relations: Labor Market Interdependence Eds J. A.Bustamante, C. W. Reynolds, and R. A. Hinojosa Ojeda. Stanford, CA: StanfordUniversity Press. Pp. 155195.

Curran, S. R., and E. Rivero-Fuentes2003 Engendering Migrant Networks: The Case of Mexican Migration. Demography

40(2):289307.Darian, L. T.2006 The Emerging Hispanic Homeland of the Pacic Northwest: A Case Study of Yakima

Valley, Washington . Bozeman, MT: Earth Sciences, Montana State University.De Genova, N.2005 Working the Boundaries: Race, Space, and Illegality in Mexican Chicago . Durham,

NC: Duke University Press.De la Rocha, M. G.1994 The Resources of Poverty: Women and Survival in a Mexican city . Cambridge, MA:

Blackwell.

Delgado-Wise, R., and H. M. Covarrubias2007 The Reshaping of Mexican Labor Exports under NAFTA: Paradoxes and Chal-lenges. International Migration Review 41(3):656679.

Donato, K. M.1993 Current Trends and Patterns of Female Migration: Evidence from Mexico. Inter-

national Migration Review 27(4):748771., M. Stainback, and C. L. Bankston III2005 The Economic Incorporation of Mexican Immigrants in Southern Louisiana: A

Tale of Two Cities. In New Destinations: Mexican Immigration in the United States Eds V. Zuniga, and R. Herna ndez-Leon. New York, NY: Russell Sage Foundation.

Pp. 76100., C. M. Tolbert II, A. Nucci and Y. Kawano2007 Recent Immigrant Settlement in the Nonmetropolitan United States: Evidence

from Internal Census Data. Rural Sociology 72(4):537559.Dunn, T. J., A. M. Aragones, and G. Shivers2005 Recent Mexican Migration in the Rural Delmarva Peninsula: Human Rights versus

Citizenship Rights in a Local Context. In New Destinations: Mexican Immigration in the United States Eds V. Zuniga, and R. Herna ndez-Leon. New York, NY: Rus-sell Sage Foundation. Pp. 155183.

Durand, J., and D. S. Massey

2003 Clandestinos: Migracio n Me xico-Estados Unidos en los Albores del Siglo XXI . Zacatecas:Universidad Autonoma de Zacatecas.



29/34

, , and C. Capoferro2005 The New Geography of Mexican Immigration. In New Destinations: Mexican

Immigration in the United States Eds V. Zuniga, and R. Herna ndez-Leon. New York, NY: Russell Sage Foundation. Pp. 120.

, , and E. A. Parrado1999 The New Era of Mexican Migration to the United States. Journal of American

History 86(2):518536., , and R. M. Zenteno2001 Mexican Immigration to the United States: Continuities and Changes. Latin

American Research Review 36(1):107127.Fennelly, K.2005 Latinos, Asians, Africans in the Northstar State: New Immigrant Communities in

Minnesota. Beyond the Gateway: Immigrants in a Changing America . Lanham, MD:Lexington Books.

Ferna ndez-Kelly, P., and D. S. Massey 2007 Borders for Whom? The Role of NAFTA in Mexico-U.S. Migration. Annals of

the American Academy of Political and Social Science 610:98118.Flores, N. Y.2005 The Interrelation between Social Context, Social Structure, and Social Capital in Inter-

national Migration Flows from Mexico to the United States . Philadelphia, PA: Ph.D.Dissertation in Sociology, University of Pennsylvania.

, R. Herna ndez-Leon, and D. S. Massey

2004 Social Capital and Emigration from Rural and Urban Communities. In Crossing the Border: Research from the Mexican Migration Project Eds J. Durand, and D. S.Massey. New York, NY: Russell Sage Foundation. Pp. 184200.

Fussell, E.2004 Sources of Mexicos Migration Stream: Rural, Urban, and Border Migrants to the

United States. Social Forces 82(3):937967., and D. S. Massey 2004 The Limits to Cumulative Causation: International Migration from Mexican

Urban Areas. Demography 41(1):151171.Garza, G.

2003 La Urbanizacio n en Me xico en el Siglo XX . Mexico City: El Colegio de Mexico.Goodwin-White, J.2007 Dispersion or Concentration for the 1.5 Generation? Destination Choices of

the Children of Immigrants in the US. Population Space and Place 13(5):313331.

Gouveia, L., M. A. Carranza, and J. Cogua 2005 The Great Plains Migration: Mexicanos and Latinos in Nebraska. In New Desti-

nations: Mexican Immigration in the United States Eds V. Zuniga, and R. Herna n-dez-Leon. New York, NY: Russell Sage Foundation. Pp. 2349.

Grey, M. A., and A. C. Woodrick

2005 Latinos have Revitalized our Community: Mexican Migration and Responses inMarshalltown, Iowa. In New Destinations: Mexican Immigration in the United States Eds V. Zuniga, and R. Herna ndez-Leon. New York, NY: Russell Sage Foundation.Pp. 133154.



30/34

Grifth, D. C.2005 Rural Industry and Mexican Immigration and Settlement in North Carolina. In

New Destinations: Mexican Immigration in the United States Eds V. Zuniga, and R.Herna ndez-Leon. New York, NY: Russell Sage Foundation. Pp. 5075.

Hanson, G. H.2003 What Has Happened to Wages in Mexico since NAFTA? NBER Working Paper

No. w9563 . Cambridge, MA: National Bureau of Economic Research. . Accessed on January 22, 2012.

Haverluk, T. W., and L. D. Trautman2008 The Changing Geography of U.S. Hispanics from 19902006: A Shift to the

South and Midwest. Journal of Geography 107(3):87101.Herna ndez-Leon, R.2008 Metropolitan Migrants: the Migration of Urban Mexicans to the United States . Berke-

ley: University of California Press.Hernandez-Leon, R., and V. Zuniga 2000 Making Carpet by the Mile: The Emergence of a Mexican Immigrant Commu-

nity in an Industrial Region of the U.S. Historic South. Social Science Quarterly 81(1):4966.

, and V. Zu niga 2005 Appalachia Meets Aztlan : Mexican Immigration and Intergroup Relations in

Dalton, Georgia. In New Destinations: Mexican Immigration in the United States Eds V. Zuniga, and R. Herna ndez-Leon. New York, NY: Russell Sage Foundation.Pp. 244273.

Herna ndez-Leon, R., and V. Zuniga 2001 A New Destination for an Old Migration: Origins, Trajectories, and Labor Market

Incorporation of Latinos in Dalton, Georgia. In Latino Workers in the Contempo- rary South Eds A. D. Murphy, C. Blanchard, and J. A. Hill. Athens: University of Georgia Press. Pp. 126135.

Hill, K., and R. Wong 2005 Mexico-US Migration: Views from Both Sides of the Border. Population and

Development Review 31(1):118.INSP2008 Disen o Muestral [ENADID 2006] . Cuernavaca, Mexico: Instituto Nacional de Salud

Publica. Accessed on January 22, 2012.

Jensen, N. M., and G. Rosas2007 Foreign Direct Investment and Income Inequality in Mexico, 19902000. Inter-

national Organization 61(03):467487. JohnsonWebb, K. D.2002 Employer Recruitment and Hispanic Labor Migration: North Carolina Urban

Areas at the End of the Millennium. The Professional Geographer 54(3):406421.

Kanaiaupuni, S. M.2000 Reframing the Migration Question: An Analysis of Men, Women, and Gender in

Mexico. Social Forces 78(4):13111347.



31/34

Kandel, W., and E. A. Parrado2004 Hispanics in the American South and the Transformation of the Poultry

Industry. In Hispanic Spaces, Latino Places: A Geography of Regional and Cultural Diversity , Ed. D. Arreola. Austin: University of Texas Press. Pp. 273292.

, and 2005 Restructuring of the US Meat Processing Industry and New Hispanic Migrant

Destinations. Population and Development Review 31(3):447471.Kandel, W. A., and E. A. Parrado2006 Rural Hispanic Population Growth. In Population Change and Rural Society Eds

W. A. Kandel, and D. L. Brown. Dordrecht: Springer. Pp. 155175.Krissman, F., N. Foner, R. Rumbaut and S. Gold2000 Immigrant Labor Recruitment: US Agribusiness and Undocumented Migration

from Mexico. In Immigration Research for a New Century . Ed. N. Foner, R. Rumbaut,and S. Gold. New York: Russell Sage. Pp. 277300.

Leach, M. A., and F. D. Bean2008 The Structure and Dynamics of Mexican Migration to New Destinations in the

United States. In New Faces in New Places: The Changing Geography of American immigration Eds. D. S. Massey. New York, NY: Russell Sage Foundation.Pp. 99123.

Lichter, D. T., and K. M. Johnson2006 Emerging Rural Settlement Patterns and the Geographic Redistribution of Amer-

icas New Immigrants. Rural Sociology 71(1):109131.

, and 2009 Immigrant Gateways and Hispanic Migration to New Destinations. International Migration Review 43(3):496518.

Light, I. H.2006 Deecting Immigration: Networks, Markets, and Regulation in Los Angeles . New York,

NY: Russell Sage Foundation.Lindstrom, D. P., and A. Lopez-Ramrez2010 Pioneers and Followers: Migrant Selectivity and the Development of U.S. Migra-

tion Streams in Latin America. Annals of the American Academy of Political and Social Science 630(1):5377.

Lozano-Ascencio, F., B. R. Roberts, and F. D. Bean1999 The Interconnections of Internal and International Migration: The Case of theUnited States and Mexico. In Migration and Transnational Social Spaces Ed L.Pries. Aldershot: Ashgate. Pp. 138161.

Lustig, N.1990 Economic Crisis, Adjustment and Living Standards in Mexico, 198285. World

Development 18(10):13251342.Malone, N., K. F. Baluja, J. M. Costanzo and C. J. Davis2003. The Foreign-Born Population in the United States: 2000. In Census 2000 Brief ,

ed. U.S.C. Bureau. Washington, D.C.: U.S. Department of Commerce, Economicsand Statistics Administration. Pp. 19.

Marcelli, E. A., and W. A. Cornelius2001 The Changing Prole of Mexican Migrants to the United States: New Evidence

from California and Mexico. Latin American Research Review 36(3):105131.



32/34

Massey, D. S.1990 Social Structure, Household Strategies, and the Cumulative Causation of Migra-

tion. Population Index 56(1):326.2008 New Faces in New Places: The Changing Geography of American Immigration . New

York, NY: Russell Sage Foundation. Pp. viii, 370 p, and C. Capoferro2008 The Geographic Diversication of American Immigration. In New Faces in New

Places: The Changing Geography of American Immigration Ed D. S. Massey. New York, NY: Russel Sage Foundation. Pp. 2550.

, and K. E. Espinosa 1997 Whats Driving Mexico-U.S. Migration? A Theoretical, Empirical, and Policy

Analysis. American Journal of Sociology 102(4):939999., L. Goldring, and J. Durand1994 Continuities in Transnational Migration: An Analysis of Nineteen Mexican Com-

munities. American Journal of Sociology 99(6):14921533., F. Kalter, and K. A. Pren2008 II. Migration. Ko lner Zeitschrift fu r Soziologie und Sozialpsychologie 48:134161., and F. Riosmena 2010 Undocumented Entry and Exit from Latin America in an Era of Rising Enforce-

ment. Annals of the American Academy of Political and Social Science 630:294321., J. S. Rugh, and K. A. Pren

2010 The Geography of Undocumented Mexican Migration. Mexican Studies Estudios Mexicanos 26(1):129152., and R. Zenteno2000 A Validation of the Ethnosurvey: The Case of Mexico-U.S. Migration. Interna-

tional Migration Review 34(3):766793.Menjvar, C.2000 Fragmented Ties: Salvadoran Immigrant Networks in America . Berkeley: University of

California Press.Murphy, A. D., C. Blanchard, and J. A. Hill2001 Latino Workers in the Contemporary South . Athens: University of Georgia Press.Neuman, K. E., and M. Tienda 1994 The Settlement and Secondary Migration Patterns of Legalized Immigrants:

Insights from Administrative Records. In Immigration and Ethnicity: The Integra- tion of Americas Newest Arrivals Eds B. Edmonston, and J. S. Passel. Washington,D.C.: Urban Institute Press. Pp. 187226.

Nevins, J.2007 Dying for a Cup of Coffee? Migrant Deaths in the US-Mexico Border Region in

a Neoliberal Age. Geopolitics 12(2):228247.Palloni, A. et al.2001 Social Capital and International Migration: A Test using Information on Family

Networks. American Journal of Sociology 106(5):12621298.Parrado, E. A., and C. A. Flippen2005 Migration and Gender among Mexican Women. American Sociological Review

70(4):606632.



33/34

, and W. Kandel2008 New Hispanic Migrant Destinations: A Tale of Two Industries. In New Faces in

New Places: The Changing Geography of American Immigration Ed D. S. Massey.New York, NY: Russell Sage Foundation. Pp. 99123.

Passel, J. S., and D. V. Cohn2008 Trends in Unauthorized Immigration: Undocumented Inow now Trails Legal Inow .

Washington, DC: Pew

Documents

Massey 2012. Origins, Destinatios and Paths of Mexican Migration