Transnational Entrepreneurs: An Alternative Form of Immigrant Economic Adaptation

by Alejandro Portes, Luis Eduardo Guarnizo, William J. Haller
Citation
Title:
Transnational Entrepreneurs: An Alternative Form of Immigrant Economic Adaptation
Author:
Alejandro Portes, Luis Eduardo Guarnizo, William J. Haller
Year: 
2002
Publication: 
American Sociological Review
Volume: 
67
Issue: 
2
Start Page: 
278
End Page: 
298
Publisher: 
Language: 
English
URL: 
Select license: 
Select License
DOI: 
PMID: 
ISSN: 
Abstract:

ALEJANDROPORTES    WILLIAMJ. HALLER
Princeton University    Princeton University
LUIS EDUARDOGUARNIZO    
University of California, Davis    
The recent literature on immigrant transnationalism points to an alternative form of economic adaptation of foreign minorities in advanced societies that is based on the mobilization of their cross-country social networks. Case studies have noted the phenomenon Spotential significance for immigrant integration into receiving coun- tries and for the economic development in countries of origin. Despite their sugges- tive character, these studies consistently sample on the dependent variable (trans- nationalism), failing to establish the empirical existence of these activities beyond a few descriptive examples and their possible determinants. These issues are ad- dressed using a survey designed explicitly for this purpose and conducted among selected Latin immigrant groups in the United States. Although immigrant trans- nationalism has received little attention in the mainstream sociological literature so far, it has the potential of altering the character of the new ethnic communities spawned by contemporary immigration. The empirical existence of transnationalism is examined on the basis of discriminant functions of migrant characteristics, and the relative probabilities of engaging in these kinds of activities is established based on hypotheses drawn from the literature. Implications for the sociology of immigra- tion as well as for broader sociological theories of the economy are discussed.

THE STUDY of immigration in econom- migration and the adaptation of immigrants ics and sociology has focused, since its to receiving societies (Park 1928; Ravenstein classic origins in the nineteenth century, on 1885). Economic historians (e.g., Thomas two central problems: the determinants of 1973) have examined the economic forces that gave rise to the ebb and flow of labor migration across the North Atlantic, between

Direct all correspondence to Alejandro Portes,

Great Britain and the United States. That tra-

Department of Sociology, Princeton University,

dition lasts to this day, having produced or-

Princeton, NJ, 08540, aportes@princeton.edu. The data were collected as part of the Compara- thodox push-pull models and also a set of tive Immigrant Enterprise Project (CIEP), sup- alternative theories on determinants of labor ported by grants from the National Science Foun- outflows collectively labeled the "new" eco- dation (SBR-9796286), Ford Foundation (960- nomics of migration (Massey, Arango et al.

0527 and 1005-0122), and Andrew W. Mellon 1998; Stark 1984, 1991; Todaro 1969, 1976). Foundation. We thank Steven Vertovec, and

Like economists, sociologists have addressed

members of his Transnational Communities

the origins of migration, but they also have

Seminar at Oxford University (Summer 2000),

focused on the adaptation of immigrants to

Ivan Szelenyi, David Apter, and their colleagues

their new environments. Concepts such as

at the Comparative Sociology Seminar at Yale University (October 2001), and anonymous ASR

assimilation, acculturation, and more re- reviewers for their comments. Responsibility for cently, incorporation, have been extensively the contents is exclusively ours. used in the sociological literature on immi-

AMERICANSOCIOLOGICALREVIEW,2002, VOL. 67 (APR1Lt278-298)

gration to provide conceptual guidance for the analysis of this topic (Alba and Nee 1997; Portes and Rumbaut 1996).

In recent years, a new concept, "trans- nationalism," has introduced an alternative analytic stance in international migration studies. Instead of focusing on traditional concerns about origins of immigrants and their adaptation to receiving societies, this emerging perspective concentrates on the continuing relations between immigrants and their places of origin and how this back- and-forth traffic builds complex social fields that straddle national borders. In its contem- porary usage, and as applied to immigrant populations, the term was introduced by so- cial anthropologists who first noted the in- tense interaction between places of origin and destination and the impact that such ac- tivities had in communities at both ends of the migration stream (Basch, Schiller, and Blanc-Szanton 1994).

Sociologists and social geographers fol- lowed suit with a string of empirical studies documenting the density of economic, politi- cal, and cultural ties of a number of immi- grant groups (Guarnizo and Smith 1998; Kyle 1999, 2001; Levitt 2001; Smith 1998). While not directly contesting the concepts of assimilation and acculturation, these studies strongly suggest that there had emerged in the modern world an alternative form of ad- aptation of immigrants to receiving societies that was at variance from what these tradi- tional concepts suggested. Immigrants to the developed countries were increasingly lead- ing dual lives, participating in two polities, and finding new paths of economic mobility on the basis of cross-border social networks (Levitt 2001; Poros 2001).

This emerging literature on immigrant transnationalism is characterized by an em- pirical base consisting exclusively of case studies. Qualitative case studies consistently sample on the dependent variable, that is, they document in detail the characteristics of immigrants involved in transnational activi- ties but say little about those who are not. As a result of this methodological orienta- tion, these studies often convey the impres- sion that transnationalism has become the prime form of immigrant adaptation-politi- cal and economic. This impression may or may not be accurate.

We investigate this question using the first survey designed specifically to address the phenomenon of cross-border economic, po- litical, and cultural activities among contem- porary immigrants. The analysis focuses on transnational economic enterprise because this is the form most commonly highlighted by the existing literature and the one with the greatest potential to affect the socioeco- nomic mobility of immigrant groups and their influence on the communities of origin. Based on this unique data set, we investigate immigrant transnationalism to answer the following questions:

(I) Does immigrant transnational entrepreneurship exist and is it empirically dis- tinct from more traditional forms of immi- grant economic adaptation? (2) If so, how common is it among contemporary immi- grant groups and what are its main manifes- tations? (3) What are the major factors asso- ciated with its emergence?

THE CONTESTED TERRAIN OF TRANSNATIONALISM

There exist at present two alternative and opposite positions concerning this concept. The first highlights the significance and po- tential influence of immigrant transnation- alism in the receiving and sending nations, while the second questions its importance or even its existence as a novel phenomenon. The best way to present the case for trans- nationalism as a phenomenon worth study- ing is to summarize several instances in the research literature. Instances of transnation- alism have been documented among a num- ber of immigrant groups both in the United States and Western Europe.

In their study of the large Salvadoran im- migrant populations of Los Angeles and Washington, D.C., Landolt and her associ- ates discovered a "vibrant entrepreneurial community embedded in a web of social re- lations" (Landolt, Autler, and Baires 1999: 296). The study identified four types of transnational enterprises: Circuit firms are involved in the transfer of goods and remit- tances across countries and range from an array of informal international couriers, known as viajeros, to large formal firms, such as El Gigante Express, which is head- quartered in Los Angeles. Cultural enter- prises rely on daily contacts with El Salva- dor and depend on the desire of immigrants to acquire and consume cultural goods from their country. Salvadoran newspapers are readily available in Los Angeles and Wash- ington, D.C., as are compact discs and vid- eos with the latest musical hits. Ethnic en- terprises are small retail firms catering to the immigrant community. These depend on a steady supply of imported goods from El Salvador, such as foodstuffs and clothing. Finally, return migrant microenterprises are firms established by returnees to El Salva- dor that rely on their contacts in the United States. They include restaurants, video stores, auto sales and repairs, laundromats, and office supplies. Summarizing their find- ings on this last type of enterprise, the au- thors conclude that:

Typically, the idea for a microenterprise originates with the migrant's experience in the United States, and the investment capi- tal comes from the migrant's personal sav- ings. Given the precarious and often low rentability of their business ventures, expan- sion and maintenance costs often force the entrepreneur to seek wage work in the United States on a regular basis. (Landolt et al. 1999:299)

A similar pattern is detected by Itzigsohn et al. (1999)in their study of the Dominican immigrant communities in the Washington Heights area of New York City. These re- searchers also uncovered a number of infor- mal transnational couriers operating between the United States and the Dominican Repub- lic; the proliferation of stores selling im- ported Dominican foodstuffs, music, and newsprint in New York and Providence; and the rapid growth of remittance agencies, known locally as financieras. Return-migrant firms in the capital city of Santo Domingo also include an array of businesses based on examples found in the United States. They include video stores, laundro- mats, car detailing, home delivery of fast food, and computer software stores.

The residential construction industry in the Dominican Republic has become trans- nationalized through its increasing depen- dence on immigrant demand. Construction and real estate firms regularly advertise in the immigrant press in New York City. En- tire residential neighborhoods in Santo Domingo have been built with the expatriate community in mind. Reflecting the growing importance of remittances and investments, the Dominican government has facilitated the election of a representative of the New York immigrant community to the national Congress and appointed an immigrant as its consul in New York City. In official par- lance, Dominican immigrants are no longer "absent" (ausentes) but only "temporarily abroad" (Itzigsohn et al. 1999).

A third example with a unique cultural twist is provided by Kyle's (1999, 2001) study of the Otavalan indigenous community in the highlands of Ecuador. Traditionally, the town of Otavalo had specialized in the production and marketing of clothing, devel- oping and adapting new production skills since the colonial period under Spain. Dur- ing the last three decades or so, Otavalans have taken to traveling abroad to market their colorful wares in major cities of Europe and North America. In contrast to other in- digenous producers who sell their crafts to intermediaries, Otavalans appropriate the full value of their production by bringing it themselves to First World consumers. Dur- ing the same period, semipermanent Otavalan enclaves began to emerge abroad. Their distinctive feature is that members do not make their living from wage labor but from the sale of goods broughtfrom Ecua- dor. The back-and-forth movement required by their trade has made Otavalans a common sight, not only at Quito airport but also in street fairs in New York, Paris, Amsterdam, and other First World cities (Kyle 2001).

This dense transnational commerce has had a profound impact on the town. Revers- ing the traditional racial hierarchy of the Andes, in Otavalo it is the indigenous entre- preneurs rather than local whites or mesti- zos who inhabit the better houses and have the best cars. In the streets of Otavalo, it is not uncommon to find white women attired in the regal traditional indigenous dress-the wives of transnational entrepreneurs who have brought the attire back from their jour- neys in Germany, England, or the Nether- lands (Kyle 1999).

A final example concerns the informal trans-Atlantic trade linking the islands of Cape Verde to their former colonial country, Portugal, as well as to other Portuguese- speaking nations. This trade is conducted by women travelers. According to a research team from the University of Lisbon, these women have earned a name in the local dia- lect, rebidantes, which is roughly translated as "those who are able to overcome ob- stacles and create new life opportunities." The obstacles to be overcome are the costs of travel and the difficulties of buying goods in foreign countries; the opportunities relate to economic mobility and the acquisition of a "respectable" position for their families and communities (Marques, Santos, and Araujo 2001).

This informal trans-Atlantic commerce is made possible by the presence of settled Cape Verdean communities abroad and the long-distance networks among kin and friends that they make possible. "Bounded solidarity" (Portes and Sensenbrenner 1993) within these networks enables women of modest means to access wholesalers and other suppliers in distant countries; trust en- forced by kin in various national locations turns these women into reliable international couriers carrying quantities of money and goods across the Atlantic. Air transportation and long-distance communication have fa- cilitated the growth of this transnational trade that now employs thousands of women in the Cape Verde islands (Marques et al. 2001).

Despite the novelty and interest raised by these studies, which differ markedly from traditional immigrant assimilation stories, skepticism continues about the significance of these forms of grassroots transnational enterprise. The skeptical position toward the transnational perspective makes two key ar- guments: First, that these activities, although colorful, are exceptional. By and large, the bulk of contemporary immigrants continues to follow the traditional routes, either set- tling and assimilating in the receiving soci- ety or returning for good to their home coun- tries (Lopez 2001). The second argument simply asserts that there is nothing new in what is referred to as transnationalism. Eu- ropean immigrants coming to the Americas at the turn of the twentieth century engaged in the same patterns of back-and-forth travel and trans-Atlantic economic ventures link- ing places of origin and destination (Foner 1997).

There is some merit in the second argu- ment insofar as it establishes a connection between past and present migrant flows, but the existence of cross-border immigrant ini- tiatives in earlier periods does not imply a fatal flaw of the transnational perspective. First, technological conditions a century ago made it impossible for cross-border enter- prises to reach the range and intensity of the initiatives described in the contemporary lit- erature. Air transportation and long-distance telephone, facsimile, and internet communi- cations underlie the remarkable flexibility attained by Otavalan trade networks, the proliferation of Salvadoran viajeros traveling regularly between California and remote Central American towns, and the viability of rebidante networks across the Atlantic. Even scholars who have advanced this objection recognize that technological conditions at present support the rise of immigrant trans- national activities of a scope and density in- commensurate with those documented in the past (Foner 1997).

Second, the existence of precedents to contemporary immigrant transnationalism does not invalidate its potential theoretical importance. Indeed, the retrospective redis- covery of similar activities among immi- grants at the turn of the twentieth century highlights the value of the concept in point- ing to previously overlooked connections. The tendency to dismiss an idea once histori- cal precedents have been found was ana- lyzed in a classic essay by Merton (1968) who referred to it as the fallacy of adumbra- tion. Merton prefaced his analysis with Whitehead's trenchant remark that "every- thing of importance has been said before by someone who did not discover it." In Merton's (1968) own words:

What is more common is that an idea is for- mulated definitely enough and emphatically enough that it cannot be ignored by contem- poraries, and it then becomes easy to find anticipations of it. (P. 16)

In our case, once the transnational per- spective had been formulated "emphatically enough," it was a straightforward exercise to find precedents in the historical literature. Thus, it became possible to establish retro- actively links between previously uncon- nected phenomena such as, for example, the cross-Atlantic activities of Polish and Rus- sian CmigrCs to North America and the dense trans-Pacific trade conducted by Chinese merchants (Glick Schiller 1999). There is nothing wrong with such rediscoveries, but they do not invalidate the concept.

Of greater weight is the objection that transnational enterprise is exceptional and does not justify much attention. From this perspective, the mainstream sociological ap- proach to immigration, which focuses on the concepts of assimilation (Alba and Nee 1997) or, more recently, modes of incorpo- ration (Portes and Rumbaut 1996), appear quite capable of framing studies in this field. For critics such as Waldinger (1998) and Lopez (2001), the transnational activities documented so far represent little more than tangential phenomena of little numerical or substantive import.

We investigate the validity of these com- peting arguments using data collected spe- cifically for this purpose. We seek to estab- lish whether regular transnational immigrant enterprises exist and whether they represent a form of economic adaptation distinct from more traditional forms. If so, we wish to identify the characteristics of these firms and of the entrepreneurs with an eye to clarify- ing the main determinants of involvement in these activities. The phenomenon of immi- grant transnationalism has received virtually no attention in the mainstream sociological literature. Is such indifference justified, or on the contrary, is it a significant topical oversight in need of redress?

THESTUDY

Data for this analysis come from the Com- parative Immigrant Entrepreneurship Project (CIEP), a collaborative effort by universities on the east and west coasts to gather data for identifying and studying entrepreneurial ac- tivities among three immigrant nationali- ties.' The project initially collected data on

The universities are Johns Hopkins Univer- sity, Brown University, and University of Cali- fornia, Davis. In 1998, the project's headquarters was moved to the Center for Migration and De- velopment at Princeton University.

samples of approximately 50 informants in areas of immigrant concentration in 1996 and used this information to guide a survey of adult family heads in these areas. The sur- vey, conducted in 1998, collected data on 1,202 respondents born in Colombia, the Dominican Republic, or El Salvador and who had subsequently migrated to the United States. Interviews were conducted in selected areas in Los Angeles, New York, Providence, and Washington, D.C.

The survey identified respondents via two sampling strategies. First, census tracts cor- responding to the immigrant community of interest in each city were identified and a two-stage cluster random sample was drawn using city blocks as primary sampling units (Kish 1965:155). Blocks were numbered and then selected through simple random sam- pling. Within selected blocks, a systematic random sampling of dwellings was conducted. The sampling fraction was constant within each area, being either 1-in-4 or l-in- 5 of a block's dwellings. A screener question on the place of birth of the household head served to determine eligibility. Dwellings in

1 which the household head was born in the United States or in locations outside the tar- get country of origin were replaced.

I This two-stage sample design maintains equal probability of selection within sampling sites because the sample of dwellings is self-weighting with respect to the first- stage clusters (Frankel 1983; Kish 1965: 154-61). However, probabilities of selection vary across sites because of the unequal size 1 of immigrant communities in each site. For I analysis using the total pooled sample, we adjusted these differences through case 1 weights. For each city and each immigrant 1 community, weights are the obverse of the I sampling fraction, defined as the ratio of the ' sample to the population of household heads of the target national origin in each area. The 1990 U.S. Census of Population and Hous- ing provided the necessary information to compute sampling fractions for each sam- pling site (Frankel 1983; Sudman 1983).

Second, the CIEP survey includes a non- random sample of entrepreneurs identified through informant referrals. The purpose of this sample was to compensate for the rela- tive dearth of entrepreneurs in the general population. A random sample, especially a

Sampling National Origin Number of Cases Fraction and Study Site Sample Populationa (Percent)

Colombian

Queens, NY 311 26,750 1.16

Dominican

Providence, RI 159 2,296 6.92

Washington 259 88,930 .29 Heights, NY

Subtotal 418 91,226

Salvadoran

Los Angeles, CA 240 57,076 .42
Washington, D.C. 233 12,176 1.91

Subtotal 473 69,252

TOTAL 1,202 187,228 .64

Percent Self-Employed CIEP 1990 Random Census Sample

8.0 11.5

4.0 21.3

7.8 8.2

7.7C 8.5

7.4 7.0

5.3 7.2

7.Oc 7.Oc

7.5 8.4

percent Transnational Entrepreneursb

4.3

13.6

3.7

4.8

5.5

5.3

5.3

5.1

Percent Trans-Percent national Domestic among Entre-the Self- preneursC Employed

7.2 37.5

7.7 64.0

4.5 45.6

3.7 52.5

1.5 78.5

1.9 74.0

1.7 76.2

3.3 57.9

a Working adult heads of households of the relevant nationality in target area, 1990 census. Weighted CIEP sample. Weighted average by population of adult heads of the relevant nationality in each site.

relatively small one, runs the risk of identi- fying too few entrepreneurs for analysis. Thus, it was necessary to supplement the two-stage sample to ensure sufficient cases. For this purpose, information was requested from the set of informants interviewed in the first phase of the study, which included com- munity leaders, politicians, priests, and lead- ing professionals. These informants were asked to identify all entrepreneurs of the same national origin that they knew in the community, whether transnational or not. We followed these leads and attempted to iden- tify and interview all persons mentioned.

This referral sample is treated as a special stratum (Kish 1965:408). It can be used as an estimate of the proportion of immigrant entrepreneurs in a community only on the assumption that informants did not miss a significant number, or in other words, that the probability of selection of such cases ap- proaches 1.This situation was approximated in some areas where the leads provided by informants were redundant, and the same persons were identified time and again. In these instances, the number of entrepreneurs in the sample approaches or exceeds the to- tal number of self-employed persons of the target nationality in the area, as reported by the 1990 census. This referral sample is used primarily to investigate the socioeconomic background and present characteristics of different types of entrepreneurs, as com- pared with other immigrants. For this pur- pose, we pooled both samples (cluster and referral) in a multivariate analysis, weight- ing referral cases by the proportion that the self-employed represented of the population of working adult household heads of the rel- evant nationality in each sampling area. This procedure avoids losing valuable informa- tion, while simultaneously avoiding biased estimates resulting from different probabili- ties of selection (Kish 1965:406-408; Sudman 1983). Samples, sampling fractions, and percentages of self-employed household heads in each target area are presented in Table 1.

DEFINITION OF TRANSNATIONALISM

Most immigrants occasionally engage in some kind of cross-border activity, such as sending remittances to kin or purchasing goods for resale in the home country on a casual basis. Although these activities ex- pand the transnational field, they are not the subject of investigation for two reasons. First, they are the least novel manifestation of transnationalism, as occasional return trips and remittances have been nearly uni- versal among immigrant groups, past and present (Foner 1997; Guarnizo and Smith 1998). Second, they do not represent an al- ternative mode of economic adaptation be- cause most persons engaged in them are still wage or salaried workers in the host soci- ety.

If the immigrant transnationalism celebrated in the recent literature means any- thing, it must amount to a distinctive form of economic adaptation. In other words, there must be a distinct class of immigrants who engage in these activities on a regular basis and who rely on them as their primary livelihood. These transnational entrepre- neurs should be expected to differ, in a number of ways, from both the mass of im- migrants engaged in wage labor and the more traditional ethnic entrepreneurs whose activities are limited to local markets of the host society (Landolt 2001; Levitt 2001).

To identify such persons, we posed a se- ries of screening questions to all respon- dents. Entrepreneurs were operationally de- fined as respondents who indicated that they were proprietors of a firm or otherwise self-employed in response to questions ask- ing for their principal occupation and place of employment. Transnational entrepre- neurs were defined, in turn, as the subset of firm-owners who traveled abroad at least twice a year for business and who responded positively to one of the following questions:

The success of my firm depends on regular contact with foreign countries.

The success of my firm depends on regular contact with [Colombia/Dominican Repub- lic/El Salvador, according to respondent's country of origin].

A special questionnaire module was then administered to entrepreneurs, both domes- tic and transnational, to gain detailed infor- mation about their firms and the character of their activities. This procedure yielded three mutually exclusive employment categories: wage workers, domestic entrepreneurs, and transnational entrepreneurs. Our first ques- tion then, is: Are these three preliminary cat- egories actually distinctive in terms of back- ground characteristics, participation in vari- ous activities, and economic situation of their members?

SELECTION OF IMMIGRANT GROUPS

Colombian, Dominican, and Salvadoran im- migrants were the target nationalities and were selected for several reasons. First, they are all sizable immigrant groups, each com- prising half a million or more immigrants in 1996 (Farley 2001; U.S. Bureau of the Cen- sus 1999). Second, they are less well-known than larger immigrant populations, such as Mexicans, who have been studied exten- sively (Massey, Alarcon et al. 1987; Suarez- Orozco 1998). Third, their contexts of exit and reception are sufficiently diverse to pro- vide a basis for comparing different types of economic adaptation.

Dominicans are primarily economic immi- grants who come to escape difficult condi- tions at home while maintaining close ties with their families and communities of ori- gin. Over time, Dominican immigration has become increasingly diversified-encom- passing professionals as well as rural labor- ers-and has spread from the Washington Heights area of Manhattan into smaller cit- ies in the New York-Boston corridor, prima- rily Providence (Itzigsohn et al. 1999; Levitt 1997).

Salvadorans, in contrast, represent a po- litical emigration-they came to the United States to escape a violent civil war at home. Denied asylum by U.S. immigration authorities, most Salvadorans have subsisted as illegal immigrants in low-paid menial jobs. Most of the early arrivals were refu- gees from the countryside-small farmers and rural laborers who, despite difficult conditions of resettlement in the United States, have managed to retain strong ties with their communities of origin. Over time, the educational and occupational com- position of the group has diversified, and its legal situation has improved. Salvadorans cluster primarily in central-city areas of the Southwest, mainly Los Angeles. Over the years, they have also drifted east, establish- ing another large concentration in Washing- ton, D.C., primarily in the Adams Morgan section (Landolt et al. 1999; Mahler 1995; Repak 1995).

Colombians comprise a more recent im- migrant group that has grown rapidly be- cause of increasing political and drug-re- lated violence at home. Most Colombian immigration originates in cities and con-tains a sizable proportion of university-edu- cated professionals. Ties with the home country are weaker than those among other migration streams that originate in tightly knit rural areas. Several past studies report that the stigma associated with the drug trade and the perennial suspicion that others may be involved lead Colombians to be dis- trustful of each other and less willing to en- gage in cooperative activities (Diaz 1997; Guarnizo, Sanchez, and Roach 1999). For the same reason, Colombian immigrants are more dispersed than other groups, having few identifiable areas of residential concen- tration. The principal cluster is found in the Jackson Heights area of Queens, New York, with secondary clusters in Los Angeles and Miami.

Table 1 presents the characteristics of the CIEP sample by national origin and location. The 1,202 cases in the sample represent an estimated total population of over 180,000 immigrant adult household heads in the five target areas. For comparative purposes, we present 1990 census figures on the percent- age of self-employed working adults of the same nationality in these areas. With the ex- ception of Providence, Rhode Island, where our two-stage random sample yielded a per- centage of entrepreneurs five times greater than those counted by the census, the figures are of a similar magnitude. Taking into ac- count the almost 10-year difference in data collection and the rapid growth of immigra- tion during the 1990s, these results support the overall representativeness of the CIEP sample.

Census data do not differentiate the self- employed by type of entrepreneurship. We do this on the basis of the pooled sample and provide estimates of the percentage of transnational and domestic entrepreneurs for each national origin and in each site.

RESULTS

DISCRIMINANT

ANALYSIS

Table 1 makes clear two initial results of the survey. First, immigrant communities in- clude a class of persons who can be identi- fied as transnational entrepreneurs according to explicit criteria, such as those used in the screening questions. Second, this class is not dominant, but represents a fraction of the majority of immigrants who continue to be wage workers. This finding exposes a key limitation of the extensive qualitative litera- ture on the topic, namely its tendency to fo- cus on visible instances of transnationalism while overlooking the majority who are not involved in transnational activities on a regular basis. Simultaneously, we note that transnationals represent a sizable percentage of all immigrant entrepreneurs. In the pooled and weighted CIEP sample, transnationals represent only 5 percent of the total, but make up 58 percent of the self-employed. This is a significant finding as most of the past literature on immigrant entrepreneur- ship has neglected this distinction, suggest- ing, by default, that most such activities were of a purely domestic character (Nee and Sanders 2001; Portes and Zhou 1999).

Given these results, the next question is the extent to which immigrants who were identified as transnationals possess a distinc- tive social and economic profile. More pre- cisely, we seek to establish whether the three categories of economic adaptation defined on the basis of survey screening questions are actually different. For this purpose, we employ discriminant analysis, a method first introduced by Fisher, which assigns cases to a set of mutually exclusive categories on the basis of a linear function of predictor vari- ables. Weights assigned to each predictor are chosen so that the ratio of the between- groups sum of squares to the within-group sum of squares is maximized (Norusis 1990; Van de Geer 1971:242-43). The canonical discriminant function may be written as fol- lows:

where C,, is the centroid (average) of each mutually exclusive category i in discrimi- nant function j, the ps are unstandardized

Statistic/Variables

Statistic

Eigenvalue
Percent of common variance
Canonical correlation
Wilk's I.
Chi-square (d.f.)

Predictors a

Demographics:

Age
Sex (male)
Nationality: Colombian
Nationality: Salvadoran

Adaptation: Experienced discrimination in U.S.

U.S.
citizen
U.S.
monthly income
Ties with Home Country:
Hometown associations
Charity associations
Political organizations
Sport clubs
Attends hometown celebrations
Owns/invests in real estate
Money sent for hometown projects
Money sent for political campaigns

Group Centroids

Wagelsalaried worker
Domestic entrepreneur

Function 1 Function 2

Discriminant Function Coefficients

.233 --.277 ,638 (.26) -.233 ,000 (.45) ,000

-.583 --.040 ,799 -.728

Transnational entrepreneur 1.372 --.310

a See Appendix A for description of variables. Unstandardized coefficients Pooled within-group correlations are in parentheses. Correlations smaller

than 25 are omitted Dominicans are the reference category ***p< .001

discriminant function coefficients, and Oik are mean values in an array of k variables for each group i.

Predictor variables used in the analysis fall under three categories: (1) individual demographic traits, including age, sex, and national origin; (2) adaptation characteris- tics, including citizenship, monthly income, and perceptions of discrimination; and (3) relations with the home country. Appendix A presents means, medians, standard devia- tions, and ranges for all variables.

Unlike multiple regression, the purpose of discriminant analysis is not to examine causal effects, but to establish whether arbi- trarily constructed groups actually differ and, if so, in what ways. Table 2 presents the results. The first finding of note is that this array of predictors does empirically dis- criminate between the three groups. This is indicated by the eigenvalues (the ratio of be- tween-sum-of-squares to within-sum-of- squares) and the canonical correlations (the ratio of between-sum-of-squares to total variation) at the top of the table. With three groups, the analysis yields a maximum of two linear functions. The discriminating value of each can be tested with the chi- square distribution of Wilk's A,which is the obverse of the canonical correlation squared. As shown by the chi-square figures at the top of the table, both functions are significant, but the first accounts for the bulk of the dif- ference between groups. This is indicated by the wide gap in eigenvalues and the percent- age of common variance explained.

A look at the group centroids shows what kinds of difference the analysis identifies. The first, and by far the most powerful function, discriminates between all entre- preneurs whose mean values are positive and wage workers whose centroid is nega- tive. Among entrepreneurs, transnationals are far more distant in the canonical func- tion space from the wage-working majority, indicating that positive values in this func- tion are mainly associated with them. By contrast, the second, and weaker, function discriminates domestic entrepreneurs, whose centroid is positive, from the other two categories.

The contribution of each individual pre- dictor to the discriminant functions can be assessed in various ways. Table 2 includes the unstandardized function coefficients and the pooled within-group correlations be- tween each predictor and the respective function. Less important correlations (< .25) are excluded. Positive contributions to Func- tion 1-defined by transnational entrepre- neurship-are associated with monthly in- come, U.S. citizenship, and sex (males). Contrary to what might have been expected, immigrants who have become U.S. citizens are more likely to be classified in the trans- national category. The income correlation reflects the established finding in the immi- gration literature that entrepreneurs have higher average earnings than salaried work- ers (Bailey and Waldinger 1991; Nee and Sanders 2001).

Predictably, Function 1 is also associated with an array of continuing ties with the home country, a pattern that is made clearer by the variable-to-function canonical corre- lations. Participation in charity associations in the sending country, in hometown civic committees, and contributions to public works in sending communities are all strongly associated with transnational entre- preneurship. Function 2 is defined mainly by national origin. Relative to Dominicans, who serve as the reference category, Colombians are far more likely to be found among do- mestic entrepreneurs, as indicated by the positive sign of the discriminant function coefficient and the corresponding correla- tion; Salvadorans display precisely the op- posite pattern.*

The discriminant analysis successfully re- classifies three-fourths of the CIEP weighted sample into the originally assigned groups. Results, particularly those for the first dis- criminant function, indicate that the three modes of economic adaptation are not arbi- trary categories but possess a distinctive em- pirical profile. Despite their small numerical presence among the groups studied, trans- nationals form an identifiable category that differs from other entrepreneurs as well as from wage workers.

THEORY.Transnational entrepreneurs are self-employed immigrants whose business activities require frequent travel abroad and who depend for the success of their firms on their contacts and associates in another coun- try, primarily their country of origin. The pre- ceding analysis tells us that membership in this class is associated with high income, ac- quisition of U.S. citizenship, and the simul- taneous preservation of a number of ties with

Less predictably, domestic entrepreneurship is also associated with strong involvement in home country political parties as indicated by the variablelfunction correlations. Additional analy- sis reveals that this result is a function of the pat- tern of association between predictors that ren- der the value of coefficients relative to each other (Van de Geer 1971). With effects of other na- tional groups controlled, these coefficients reflect the stronger tendency for Dominican immigrants to become domestic entrepreneurs and to take an active part in their country's electoral politics. By contrast, Salvadorans tend toward both home- town civic activism and transnational enterprise.

the home country. We examine the potential determinants of economic transnationalism using several hypotheses suggested by socio- logical theory that are testable with our data.

A traditional assimilation approach toward immigration would suggest that transna- tional activities are transitional and bound to disappear over time as immigrants become better integrated in the host society (Alba and Nee 1997; Warner and Srole 1945). Thus, we expect that transnationalism will decline with years of U.S. residence and be most common among recent cohorts of mi- grants. Immigrants with a precarious economic foothold, particularly those who ex- perienced serious downward mobility upon arrival, should also be more likely to avail themselves of this option. Although transna- tional entrepreneurship is correlated with higher average incomes and U.S. citizenship in the preceding analysis and this result casts doubts on these predictions, the question re- quires additional investigation.

The recent literature on immigrant entre- preneurship has taken a different tack, iden- tifying gender, marital status, and human capital as important predictors. Past research consistently indicates that married males are overrepresented among entrepreneurs; gen- der, by itself, is the most powerful factor (Light and Gold 2000; Portes and Jensen 1989). Human capital, in the form of years of education and high occupational skills, has also been found to play a significant role in immigrant business success. We reason that transnational entrepreneurship would be even more likely to depend on these human capital characteristics, as it involves greater risks and complexity.

Lastly, recent writings on the economic sociology of immigration highlight two other relevant factors (Granovetter 1995; Roberts 1995). First, social networks play an important role in the process of immigrant adaptation based on their size and other characteristics. Other things being equal, we expect that individuals with more extensive and diverse social networks will be in a bet- ter position to initiate and sustain transna- tional enterprise (Kyle 1999; Poros 2001). Second, the social contexts in which particu- lar immigrant flows are embedded can also be expected to affect their economic options, regardless of individual characteristics.

Thus, groups who came to the United States to escape political upheaval and general vio- lence in their home country may have no transnational options, while those who are part of strong communities with multiple ties to a nation at peace can find numerous cross- border economic opportunities. The com- parative character of the CIEP data and the selection of immigrant nationalities in order to maximize differences in contexts of exit and reception allow us to test these notions.

FINDINGS.

There are two ways of concep- tualizing the probability of transnational en- trepreneurship: (1) in relation to the other forms of economic adaptation combined; (2) in relation to the dominant category of wage employment in the host country's labor mar- ket. We consider both options using logistic regression models with different reference categories and compute the net increaselde- crease in the probabilities of transnational- ism using the method proposed by Petersen (1985).~

Table 3 presents the results. As expected, immigrant businesses of any kind are prima- rily the business of married males-both sex and marital status bear strongly on the pur- suit of this economic path. This result is no different from that reported in the literature on immigrant entrepreneurship, except that the gender effect is stronger for the probabil- ity of transnational enterprise than for do- mestic enterprise. Measures of socioeco- nomic background-education and profes- sionallexecutive experience-have the posi- tive effects anticipated by the same litera- ture: Both increase the probability of self- employment, but the effects are stronger on transnational enterprise than on domestic en- terprise. Based on model coefficients, a mar- ried male with a college education and a pro- fessional background has a 37 percent greater probability of becoming a trans- national entrepreneur; the figure increases to 45 percent if wage worker is the reference category. Because all of these predictors re- fer to characteristics brought by immigrants from their home countries, the direction of causality is unambiguous.

For multinomial regression models, Petersen's (1985) method is modified to take into account the multiple categories of the dependent variable.

The notion that transnational activities are a transitional pursuit, to be abandoned as as- similation takes hold, is not supported by these data. Long periods of residence in the United States increase the probability of en- gaging in both domestic and transnational enterpri~e.~

The coefficient associated with the most recent immigrant cohort, that arriv- ing during the 1990s, is nonsignificant and negative. Nor are experiences of early down- ward mobility associated with transnational- ism. Our mobility indicator is the ratio of the last occupational status in the country of ori- gin to the first occupational status in the United States (see Appendix A). A higher score on this variable thus signifies greater downward mobility and would be associated with increased transnationalism. The ob- served effect is actually negative, indicating that each point drop in status reduces the probability of transnational enterprise by 2 percent.

Rejection of this set of hypotheses indi- cates that, contrary to what may be expected from an assimilationist standpoint, it is not recency of arrival or the experience of occu- pational failure that leads immigrants to be- come transnational entrepreneurs. Our find- ings indicate that this route is mainly open to immigrants who have established a secure foothold in the United States. While those experiencing serious downward mobility may be motivated to follow the same route, they seemingly lack the experiences, re- sources, and stability to succeed at it.

Results pertaining to social networks point in the same general direction. The CIEP sur- vey contains data on attributes of the respon- dents' networks, including their size, den- sity, breadth, and multiplexity. Early model specifications (not shown) indicated that the key characteristic of immigrant networks with respect to entrepreneurship is their size. Business owners have more numerous social ties than do wage workers, and transnational

In other model specifications, we substituted acquisition of U.S. citizenship for years of U.S. residence. The two variables cannot be added si- multaneously because of collinearity. The direc- tion of effects is the same as those observed here, with U.S. naturalization increasing the probabil- ity of transnationalism by 10 percent after con- trolling for other variables.

entrepreneurs have more than do domestic entrepreneurs. As shown in Table 3, the co- efficient for size of social networks is very strong, exceeding five times its standard er- ror in both the binomial and multinomial models. Each additional social contact in- creases the probability of transnational en- terprise by 1.5 percent. A second character- istic of social networks is their scope, de- fined as the ratio of extra-local contacts (out- side the city of residence) to purely local ones. The effect of network scope on domes- tic entrepreneurship is significant but nega- tive: Each unit increase in the ratio of out- side-to-local contacts reduces the probabil- ity of engaging in this form of economic ad- aptation. This result reinforces the view of domestic enterprise as an economic path en- gaged in by immigrants whose ties do not reach beyond the local comm~nity.~

With Dominicans as the reference cat- egory, results in both models confirm the sharp differences between the three immi- grant nationalities compared here. All else equal, Salvadorans are 7 to 9 percent more likely to engage in transnational business activities than Dominicans, while Colombi- ans are less likely to do so by about half that figure. Both coefficients are reliable, indicat- ing the resilience of national differences af- ter controlling for other factors. These results fit the known contexts of exit and reception under which each of these migrant flows has taken place: Salvadoran transnationalism is supported by strong bonds of solidarity with origin communities that were forged during the country's civil war; these bonds were subsequently put to economic use once the country returned to political democracy and internal peace. In contrast, Colombian trans- nationalism is weakened by continuing po-

The causal relationship between network size and composition and entrepreneurship may run in the direction opposite to that posited by the model. With cross-sectional data and in the absence of suitable instruments to model network character- istics independently, it is not possible to satisfac- torily address this question. We note, however, that theory in this area has consistently given so- cial networks a primary causal role both in the onset of immigration and in the economic success of immigrant enterprise (Massey 1987; Massey and Espinoza, 1997; Portes and Jensen 1989; Rob- erts, Frank, and Lozano-Ascencio 1999).

Table 3. Determinants of Transnational Enterprise among Colombian, Dominican, and Salvadoran Immigrants: CIEP Weighted Sample, 1998

Predictor    Binomial TrEntrCoef.    Lansnational epreneurship (S.E.)    ogistic Regression Aa    Multinomial Logistic Regression Transnational Domestic Entrepreneurship Entrepreneurship Coef. (S.E.) Aa Coef. (S.E.)    Aa
Demographic Age ,017 Sex (male) 1.035"' Marital status (married) .440* Number of children -.049        (.012) (.231) (.215) (.070)    -.08 .03 -            
Human Capital Education (years) Professional/executive background    .I 14*** (.026) 1.191*    ** (.33 1)    .01 .10        
Assimilation Years of U.S. residence ,036: Post- 1989 arrival -.437        (.017) (.338)    ,003 -            
Downward mobilityb Experienced discrimi- nation in the U.S.    -.402** .308    (.167) (.207)        -.03 -            
Social Networks Size    .I11*    ** (.022)        .01            
Scope    .226        (.121)    -            
Nationality Colombian    -1.519"'        (.387)        -.05            
Salvadoran    1.097"'        (.279)    .09            
Constant    -6.235    (.686)    -
Chi-square (d.f.) Pseudo R2 Number of cases    14    1.67": (14) .256 1,096    
Note: See Appendix A for description of variable measurement.
a Change in the net probability of each outcome per unit change in significant predictors, evaluated at the

mean of the weighted sample distribution. Nonsignificant effects are omitted.

Ratio of occupational status in the country of origin to status of the first U.S. occupation.

Ratio of number of contacts outside city of residence to local contacts.

Dominican immigrants are the reference category. *p< .05 **p < .01 '*'p < ,001

litical convulsions and widespread violence that is neither marginal nor associated with in the home country and by the atmosphere poverty or recency of arrival. On the con- of distrust, reflecting the pervasive influence trary, it is the better qualified, more experi- of the drug trade (Guamizo et al. 1999). enced, and more secure immigrants who are

Thus, we find that social contexts, along overrepresented in these economic activities. with individual differences, matter. To sum- Yet the opportunity to engage in them is marize, transnational enterprise is an excep- heavily conditioned by the sociopolitical tional mode of economic adaptation, but one conditions in the country of origin and the

Table 4. Multinomial Logistic Regression Coefficients Showing the Interaction Effects of National Origin on the Probability of Transnational Entrepreneurship: CIEP Weighted Sample, 1998

Predictor    Coef.    Model 1 S.E.    Aa    Coef.    Model 2 S.E.    Aa
Age Sex (male) Marital status (married) Number of children                        
Education (years) Professionallexecutive background Years of U.S. residence                        
Post- 1989 arrival                        
Downward mobility Experienced discrimination in U.S. Network size                        
Network scope                        
Nationality Colombian                        
Salvadoran                        
Interaction Effects Salvadoran x Years in U.S.                        
Salvadoran x Education                        
Salvadoran x Professionallexecutive background Salvadoran x Network size                        
Colombian x Years in U.S.                        
Colombian x Education                        
Colombian x Professionallexecutive background Colombian x Network size                        
Constant                        
Chi-square (d.0 Pseudo R2                        
Note: Effects other than those on transnational entrepreneurship are omitted.
a Change in the net probability of transnational entrepreneurship associated with a unit change in each

predictor. Nonsignificant effects are omitted. Ratio of occupational status in the country of origin to status of the first U.S. occupation. Ratio of number of contacts outside city of residence to local contacts. Dominican immigrants are the reference category.

* p < .05 **p< .01 ***p< ,001 (two-tailed tests)

characteristics of the immigrant community ternative and become channeled toward ex- itself. Depending on these conditions, mi- clusively domestic pursuits. grants may avail themselves of this option, INTERACTION EFFECTS. Based on our re- launching successful enterprises across na- sults, the next logical question is the extent tional borders, or they may lack such an al- to which the hypothesized determinants of

Table 5. Percentage Distribution of Characteristics of Transnational Enterprises, by National Origin

National Origin (Per cent) Total Characteristic Colombian Dominican Salvadoran (Percent)

Sector of Transnational Firm

Construction

Manufacturing

Retail sales

Credit, finance, real estate

Personal services

Business services/telecommunications

Health services

Others

Total

Transnational Activitya

Imports goods from abroad

Exports goods abroad

Invests in home country businesseslreal estate

Hires at least one employee in home country

Has been an international courier

Travels abroad at least twice a year for business

Percent transnational firms in weighted sampleb

Mean monthly income

a Activities conducted by the firm or the entrepreneur at the time of the interview. There were a total of 452 transnational firms in the sample.

immigrant transnationalism and the charac- The rest of the interaction effects, though not ter of these enterprises vary among the three reaching statistical significance, are consis- groups studied. Table 4 presents the interac- tently positive indicating a tendency for the tion effects of national origin with selected more qualified and better connected among predictors of transnationalism. Because of Salvadoran immigrants to engage in trans- high collinearity, it is impossible to consider national enterprise. all interaction effects simultaneously. In- The pattern for Colombian immigrants is stead, Table 4 presents two separate models exactly the opposite: All interaction effects that incorporate selected interactions be- are negative and once they are controlled the tween individual characteristics and national main effect of Colombian origin (negative in origin. This analysis reveals several impor- Table 3) changes signs. Interactions of Co- tant results. First, multinomial coefficients lombian nationality with years of U.S. resi- in Model 1 indicate that Salvadoran transna- dence and network size are statistically sig- tional enterprise is the preserve of the more nificant, showing that more established and educated members of this group. With this better connected Colombians are less likely interaction controlled, the original positive to engage in transnational enterprise. Their effect of Salvadoran origin on the probabil- tendency, as already suggested by previous ity of transnationalism becomes negative. results, is to follow the route of salaried em- ployment as their principal means of eco- nomic adaptation.

These findings lend additional support to the hypothesis of contextual effects on the onset of transnational entrepreneurship. By the same token, they qualify previous con- clusions reached on the basis of additive ef- fects: Not always are the more educated and better-established migrants more inclined to pursue this option. While this tendency holds in the aggregate, there are exceptions dic- tated by historical factors. In the absence of a socially supportive context, skilled and well-connected immigrants may be more in- clined to pursue upward mobility through salaried employment, rather than to seek ad- vantage in cross-border business ventures.

The observed variations by nationality are robust, holding across many different model specifications. Table 5 further fleshes out these differences by showing what transna- tional entrepreneurs actually do. The princi- pal difference is quantitative: Over two- thirds of transnational businesses in the CIEP weighted sample are owned by Salvadoran immigrants; 24 percent are owned by Dominicans, and less than 10 percent are owned by Colombians. The pattern of distribution across business sectors is broadly similar, indicating that immigrant transnational firms concentrate in retail merchandising and on the provision of business and telecommuni- cation services to the respective home coun- try. Transnational enterprises are also found in a broad array of other sectors that include manufacturing, construction, real estate, and health services.

The number of sectors in which these firms are found and the manifold activities in which they and their owners engage point to a highly dynamic phenomenon. As Table 5 shows, transnationals buy and sell abroad, travel internationally on business, invest in enterprises, and even hire person- nel in their home countries. The payoff for these activities is substantial: Transnational entrepreneurs had an average monthly in- come of $3,855 in 1998, a figure signifi- cantly higher than that for workers ($1,299) and even for domestic entrepreneurs ($3,031). Transnational firms initiated by Salvadoran immigrants are not only the most numerous, but they are also the largest in terms of monthly sales and net incomes.

CONCLUSION

We have investigated transnational entrepre- neurship, a phenomenon that has gained in- creasing attention on the basis of qualitative studies of immigration. We ask if it repre- sents a truly significant development merit- ing further study. Our overall conclusion is a qualified yes. The CIEP data leave little doubt that both a diversified set of transna- tional enterprises and a class of immigrants operating them exist. On the other hand, to identify a sufficient number of such entre- preneurs for analysis, we had to look for them as they represent a small proportion of the respective immigrant populations. When properly weighted, our data indicate that the percentages of self-employed persons regu- larly involved in transnational activities are in the single digits for all three immigrant groups studied.

On the other hand, three findings call at- tention to the theoretical and practical sig- nificance of the phenomenon. First, trans- national entrepreneurs represent a large pro- portion, often the majority, of the self-em- ployed persons in immigrant communities. Thus, our results uncover an important di- mension overlooked by prior studies of im- migrant entrepreneurship, namely the reli- ance of many such firms on continuing ties with their home countries. Second, multi- variate analysis shows that transnational economic activities are not associated with recency of arrival or with marginal eco- nomic status. Overall, transnational entre- preneurs are part of the elite in their respec- tive communities in terms of education and legal standing, and they derive from these activities higher-than-average incomes com- pared with the wagelsalaried majority. Third, within this overall picture, there is consider- able variability depending on the historical context under which particular migrant flows take place. This opens the possibility that, under certain favorable circumstances, transnational activities can become far more prevalent than observed so far.

Even when limited numerically, the diver- sity of transnational activities documented by our survey and the relative prominence of immigrants involved in them suggest that their impact goes well beyond the entrepre- neurs themselves. At the very least, custom- ers of these firms and their home country counterparts are brought into transnational circles on a repeated basis. Transnational firms can be viewed as bridges helping to keep ties alive with the home countries and even strengthening them over time. To the extent that transnational entrepreneurs are economically successful, they may also stimulate others to follow their example, thus expanding this mode of economic ad- aptation.

Three auestions that remain unanswered and need additional investigation are: the relationship of present-day transnationalism to its predecessors, its variability across im- migrant nationalities and contexts of recep- tion, and its evolution over time. There are numerous precedents to the phenomenon among earlier immigrant flows, and indeed, the concept of transnationalism has led to the reinterpretation of these historical in- stances. In their study of grassroots trans- national activists, Keck and Sikink (1998) note that the issue-oriented global coali- tions of today also have several predeces- sors. Yet there are at present no quantitative estimates of the incidence of earlier trans- national activities, or of their range. While it is implausible on sheer technological grounds that earlier instances of transna- tionalism could reach the levels observed today (Levitt 1997, 2001), progress in this field requires more precise estimates than the anecdotal material available so far.

Second, the forms of transnationalism can be expected to vary significantly by immi- grant nationality and context of reception in ways that are not well understood today. A growing ethnographic literature shows that the phenomenon is present not only in the United States, but also in Western Europe, and not only among Latin American immi- grants, but among immigrants from Asia, the Middle East, and Africa (Marques et al. 200 1 ;Ostergaard-Nielsen 200 1 ;Poros 200 1 ; Zhou and Kim forthcoming). However, no quantitative study of transnationalism among these nationalities has been attem~ted. Given the significant variations in this phenomenon uncovered by the CIEP survey among immi- grants of similar cultural background, we expect that future quantitative studies will document even broader differences in trans- national practices and their determinants.

Third, longitudinal research is needed to determine the proportion of immigrants in- volved in these activities who eventually re- turn home versus those who translate the eco- nomic resources acquired through these ac- tivities into a more rapid incorporation in the host society. Similarly, longitudinal research will be required to elucidate the question of generational transmissibility. Although the evidence indicates that the bulk of the immi- grant second generation is assimilating at a fast pace (Waldinger and Perlman 1998; Zhou and Bankston 1998), no studies have focused on the children of transnationals. It is not known whether most of these children will follow in their parents' wake, expanding the size and scope of their firms, or whether they will use these firms as a stepping stone for more complete and successful integration to the host society.

Transnationalism has been described in the anthropological and cultural studies lit- erature as an alternative to assimilation (Appadurai 1990, 1993; Basch et al. 1994). In reality, it is not at all clear that the two processes are at odds. As an escape from low-paid, menial occupations, transnational enterprise can offer first-generation immi- grants a more desirable path, including the economic means to support the successful adaptation of their offspring. The past socio- logical literature on immigrant entrepreneur- ship repeatedly notes how middleman enter- prises and ethnic enclaves created by first- generation immigrants financed the educa- tional and occupational advancement of the second and succeeding generations (Bailey and Waldinger 199 1 ;Nee and Sanders 200 1 ; Zhou 1992;). Transnational enterprise may offer a novel twist to the same story, accel- erating rather than retarding the long-term integration of contemporary immigrants. The finding that those engaged in these ac- tivities are, more often than not, long-term

U.S. residents, naturalized citizens, and eco- nomically successful points in this direction.

Alejandro Portes is Howard Harrison and Gabrielle Snyder Beck Professor of Sociology and Director of the Center for Migration and Development at Princeton University. His last books (with Rubtn G. Rumbaut) are Legacies: The Story of the Immigrant Second Generation

(University of California Press, ZOO/), and

Ethnicities: Children of Immigrants in America

(University of California Press, 2001). He con- tinues his longitudinal study of the adaptation of the immigrant second generation and is about to launch a new comparative study of Latin Ameri- can cities in the era of open markets and neo- liberal economic policies.

William J. Huller is Postdoctoral Research As- sociate with the Office of Population Research at Princeton University. He specializes in stratifi- cation, migration and socioeconomic change. His research includes studying the relation ship be- tween shifts in metropolitan labor markets and urban poverty. He is also directing the third round of data collection for the Miami sample of the Children of Immigrants Longitudinal Study. This data will enable research on school-to-work transitions and other early adulthood outcomes among the immigrant second generation.

APPENDIXA

Luis Eduardo Guarnizo is Associate Professor of Sociology at the Department of Human and Community Development at the University of California, Davis. His research focuses primarily on transnational processes linking Latin Ameri- can migrants in the United States to their home- lands. He has published on transnational prac- tices among Colombians, Dominicans, Mexicans, and Salvadorans in the United States. In his new research project, undertaken jointly with Ninna Nyberg Slrensen, he compares transnational po- litical, sociocultural, and economic practices of

Colombian, Dominican, and Peruvian immi- grants residing in four European countries. This new project is a logical continuation of the com- parative project presented here and will provide an empirically grounded analysis of transna- tional migration with a global scope.

Descriptive Statistics for Variables Used in the Analysis: CIEP Weighted Sample, 1998

Variable

Endogenous Variable

Immigrant Economic Adaptation:
Percent transnational entrepreneur
Percent domestic entrepreneur
Percent wage or salary worker

Exogenous Variable

Age (in years)
Sex (male = 1, otherwise 0)
Marital status (married = 1, otherwise 0)
Number of children

Nationality

Percent Colombian
Percent Dominican
Percent Salvadoran

Prior Status

Years of education

Occupation in country of origin (Professionallexecutive = 1, otherwise 0)

Assimilation

Years in the U.S."
Downward mobilityb
Experienced discrimination in U.S. (yes = 1, otherwise 0)

U.S.
citizen (yes = 1)
U.S.
monthly income
Mean    Median    S.D.    Range
21.06    -    -    -
12.92    -    -    -
66.03    -    -    -
39.26    -    -    -
45.83    -    -    -
.19    .OO    .39    1.00
15.10    14.00    8.35    51.00
1.13    1.00    .78    4.80
.46    .OO    .50    1.00
.31    .OO    .46    1.00
$1,913    $1,500    $3,194    $30,000
Continued on next page)
(AppendixA continued)

Variable

Ties with Home Countq

Hometown associations (participates regularly = 1, otherwise 0)
Charity organizations (participates regularly = 1, otherwise 0)
Political organizations (participates regularly = 1, otherwise 0)
Sports clubs (participates regularly = 1, otherwise 0)
Attends hometown celebrations (regularly = 1, otherwise 0)
Ownslinvests in real estate (invests regularly = 1, otherwise 0)
Sends money for hometown projects (yes = 1, otherwise 0)
Sends money for political campaigns (yes = 1, otherwise 0)

Social Networks

Size of respondent's social network (number of ties)
Scope of respondent's social networkC
" Year of CIEP survey minus year of arrival in U.S.

Mean Median S.D. Ranee

Ratio of status of country-of-origin occupation (coded 1 to 5) to first U.S. occupation (coded 1 to 5).

Ratio of nonlocal contacts to local contacts.

REFERENCES

Alba, Richard and Victor Nee. 1997. "Rethink- ing Assimilation Theory for a New Era of Im- migration." International Migration Review 3 1:826-74.

Appadurai, Arjun. 1990. "Disjuncture and Differ- ence in the Global Cultural Economy." Public Culture 2: 1-24. . 1993. "Patriotism and Its Futures." Public Culture 5:411-29.

Bailey, Thomas and Roger Waldinger. 1991. "Primary, Secondary, and Enclave Labor Mar- kets: A Training System Approach." American Sociological Review 56:432-45.

Basch, Linda G., Nina Glick Schiller, and Cristina Blanc-Szanton. 1994. Nations Un- bound: Transnational Projects, Post-Colonial Predicaments, and De-Territorialized Nation States. Langhorne, PA: Gordon and Breach.

Diaz, Luz M. 1997. "Comunidades Transna- cionales: El Caso Colombiano" (Transnational communities: The case of Colombia). Paper presented at the seminar on Comparative Re- search on International Migration, March, Santo Domingo, Dominican Republic.

Farley, Reynolds. 2001. "Immigrants and Their Children: Evidence from the Census Bureau's Recent Survey." Paper presented at the annual meeting of the American Sociological Asso- ciation, August 18-2 1, Anaheim, CA.

Foner, Nancy. 1997. "What's New about Trans- nationalism? New York Immigrants Today and at the Turn of the Century." Diaspora 6:355-75.

Frankel, Martin. 1983. "Sampling Theory." Pp.

21-67 in Handbook of Survey Research, edited by P. H. Rossi, J. D. Wright, and A. B. Ander- son. New York: Academic Press.

Glick Schiller, Nina. 1999. "Transmigrants and Nation States: Something Old and Something New in the U.S. Immigrant Experience." Pp. 94-1 19 in Handbook of International Migra- tion, the American Experience, edited by C. Hirschman, P. Kasinitz, and J. DeWind. New York: Russell Sage.

Granovetter, Mark. 1995. "The Economic Soci- ology of Firms and Entrepreneurs." Pp. 128- 65 in The Economic Sociology of Immigration: Essays in Networks, Ethnicity, and Entrepre- neurship, edited by A. Portes. New York: Russell Sage.

Guarnizo, Luis E., Arturo I. Sanchez, and Eliza- beth M. Roach. 1999. "Mistrust, Fragmented Solidarity, and Transnational Migration." Ethnic and Racial Studies 22:367-96.

Guarnizo, Luis E. and Michael P. Smith. 1998. "The Locations of Transnationalism." Pp. 3- 34 in Transnationalism from Below, edited by

M. P. Smith and L. E. Guarnizo. New Brunswick, NJ: Transaction.

Itzigsohn, Jose, Carlos Dore, Esther Hernandez, and Obed Vazquez. 1999. "Mapping Domini- can Transnationalism." Ethnic and Racial Studies 22:3 16-39.

Keck, Margaret E. and Kathryn Sikkink. 1998. Activists beyond Borders. Ithaca, NY: Cornell University Press.

Kish, Leslie. 1965. Survey Sampling. New York: Wiley.

Kyle, David. 1999. "The Otavalo Trade Diaspora: Social Capital and Transnational En- trepreneurship." Ethnic and Racial Studies 22:422-46. . 2001. The Transnational Peasant: The Social Construction of International Economic Migration and Transcommunities from the Ecuadoran Andes. Baltimore, MD: Johns Hopkins University Press.

Landolt, Patricia. 2001. "Salvadoran Economic Transnationalism: Embedded Strategies for Household Maintenance, Immigrant Incorpo- ration, and Entrepreneurial Expansion." Global Networks 1 :217-42.

Landolt, Patricia, Lilian Autler, and Sonia Baires. 1999. "From 'Hermano Lejano' to 'Hermano Mayor': The Dialectics of Salvadoran Trans- nationalism." Ethnic and Racial Studies 2: 290-3 15.

Levitt, Peggy. 1997. "Transnationalizing Community Development: The Case of Migration between Boston and the Dominican Republic." Voluntary Sector Quarterly 26:509-26. . 2001. "Transnational Migration: Taking Stock and Future Directions." Global Net- works 1:195-216.

Light, Ivan and Steven J. Gold. 2000. Ethnic Economies. San Diego, CA: Academic Press.

Lopez, David. 2001. "Los Angeles: Transnational City or MClange of Transnational Communities?" Paper presented at the annual meeting of the American Sociological Assocation, August 18-2 1, Anaheim, CA.

Mahler, Sarah J. 1995. American Dreaming, Im- migrant Life on the Margins. Princeton, NJ: Princeton University Press.

Marques, Maria M., Rui Santos, and Fernanda Aranjo. 2001. "Ariadne's Thread: Cape Verdean Women in Transnational Webs." Global Networks 1:283-306.

Massey, Douglas S. 1987. "Understanding Mexi- can Migration to the United States." American Journal of Sociology 92: 1372-403.

Massey, Douglas S., Rafael Alarcon, Jorge Durand, and Humberto Gonzalez. 1987. Return to Atzlan: The Social Process of International Migration from Western Mexico. Berkeley, CA: University of California Press.

Massey, Douglas S., Joaquin Arango, Graeme Hugo, Ali Kouaouci, Adela Pellegrino, and J. Edward Taylor. 1998. Worlds in Motion: Un- derstanding International Migration at the End of the Millennium. Oxford, England: Clarendon.

Massey, Douglas S. and Kristin E. Espinosa. 1997. "What's Driving Mexico-U.S. Migration? A Theoretical, Empirical, and Policy Analysis." American Journal of Sociology 102:939-99.

Merton, Robert K. 1968. "On the History and Systematics of Sociological Theory." Pp. 1-38

in Social Theory and Social Structure. New York: Free Press.

Nee, Victor and Jimy Sanders. 2001. "Under- standing the Diversity of Modes of Incorpora- tion: A Forms-of-Capital Model." Ethnic and Racial Studies 24:386-412.

Norusis, Marija J. 1990. SPSS Advanced Statis- tics User's Guide. Chicago, IL: SPSS, Inc.

Ostergaard-Nielsen, Eva. 2001. "Transnational Political Practices and the Receiving State: Turks and Kurds in Germany and the Nether- lands." Global Networks 3:261-8 1.

Park, Robert E. 1928. "Human Migration and the Marginal Man." American Journal of Sociol- ogy 33:881-93.

Petersen, Trond. 1985. "A Comment on Present- ing Results from Logit and Probit Models." American Sociological Review 50: 130-3 1.

Poros, Maritsa V. 2001. "The Role of Migrant Networks in Linking Local Labour Markets: The Case of Asian Indian Migration to New York and London." Global Networks 1:243

60.

Portes, Alejandro and Leif Jensen. 1989. "The Enclave and the Entrants: Patterns of Ethnic Enterprise in Miami before and after Mariel." American Sociological Review 54:929-49.

Portes, Alejandro and RubCn G. Rumbaut. 1996. Immigrant America. Berkeley, CA: University of California Press.

Portes, Alejandro and Julia Sensenbrenner. 1993. "Embeddedness and Immigration: Notes on the Social Determinants of Economic Action." American Journal of Sociology 98: 1320-50.

Poms, Alejandro and Min Zhou. 1999. "Entrepre- neurship and Economic Progress in the Nine- ties: A Comparative Analysis of Immigrants and African Americans." Pp. 143-71 in Immigration and Opportunity: Race, Ethnicity, and Employment in the United States, edited by F.

D. Bean and S. Bell-Rose. New York: Russell Sage.

Ravenstein, E. G. 1885. "The Laws of Migra- tion." Journal of the Royal Statistical Society

48: 167-227.

Repak, Terry A. 1995. Waiting on Washington, Central American Workers in the Nation's Capital. Philadelphia, PA: Temple University Press.

Roberts, Bryan R. 1995. "Socially Expected Du- rations and the Economic Adjustment of Im- migrants." Pp. 42-86 in The Economic Sociol- ogy of Immigration, edited by A. Portes. New York: Russell Sage.

Roberts, Bryan R., Reanne Frank, and Fernando Lozano-Ascencio. 1999. "Transnational Migrant Communities and Mexican Migration to the US." Ethnic and Racial Studies 22:238

66. Smith, Robert C. 1998. "Mexican Immigrants, the Mexican State, and Transnational Practice of Mexican Politics and Membership." LASA Forum 24: 19-24.

Stark, Oded. 1984. "Migration Decision Mak- ing." Journal of Development Economics 14: 251-59. . 1991. The Migration of Labour. Cambridge, England: Basil Blackwell.

Suarez-Orozco, Marcelo. 1998. Crossings: Mexi- can Immigration in Interdisciplinary Perspec- tives. Cambridge, MA: David Rockefeller Cen- ter for Latin American Studies, Harvard Uni- versity.

Sudman, Seymour. 1983. "Applied Sampling." Pp. 145-94 in Handbook of Survey Research, edited by P. H. Rossi, J. D. Wright, and A. B. Anderson. New York: Academic Press.

Thomas, Brinley. 1973. Migration and Economic Growth: A Study of Great Britain and the At- lantic Economy. Cambridge, England: Cam- bridge University Press.

Todaro, Michael. P. 1969. "A Model of Labor Migration and Urban Unemployment in Less Developed Countries." American Economic Review 59: 138-48. . 1976. Internal Migration in Developing Countries. Geneva, Switzerland: International Labor Office.

U.S. Bureau of the Census. 1999. Profile of the Foreign-Born Population of the United States:

1997. Current Population Reports, Special Studies #P23-195. Washington, DC: U.S. De- partment of Commerce.

Van de Geer, John P. 1971. Introduction to Mul- tivariate Analysis for the Social Sciences. San Francisco, CA: W. H. Freeman.

Waldinger, Roger. 1998. "Comment." Presented at the International Conference on National- ism, Transnationalism, and the Crises of Citi- zenship, April DATE, University of Califor- nia, Davis, CA.

Waldinger, Roger and Joel Perlmann. 1998. "Sec- ond Generations: Past, Present, Future." Journal of Ethnic and Migration Studies 245-24.

Warner, W. Lloyd and Leo Srole. 1945. The So- cial Systems of American Ethnic Groups. New Haven, CT: Yale University Press.

Zhou, Min. 1992. New York's Chinatown: The Socioeconomic Potential of an Urban Enclave. Philadelphia, PA: Temple University Press.

Zhou, Min and Carl Bankston. 1998. Growing Up American: How Vietnamese Immigrants Adapt to Life in the United States. New York: Russell Sage.

Zhou, Min and Rebecca Kim. Forthcoming. "A Tale of Two Metropolises: Immigrant Chinese Communities in New York and Los Angeles." In Los Angeles and New York in the New Mil- lennium, edited by D. Halle. Chicago, IL: Uni- versity of Chicago Press.

Comments
  • Recommend Us