In and Out of The Ethnic Economy A Longitudinal Analysis of Ethnic Networks and Pathways To Economic Success Across Immigrant Categories

In and Out of the Ethnic Economy: A Longitudinal Analysis of Ethnic Networks and
Pathways to Economic Success across Immigrant Categories

Author(s): Wendy D. Roth, Marc-David L. Seidel, Dennis Ma and Eiston Lo
Source: The International Migration Review, Vol. 46, No. 2 (Summer 2012), pp. 310-361
Published by: Sage Publications, Inc. on behalf of the Center for Migration Studies of New
York, Inc.
Stable URL: https://www.jstor.org/stable/23279469
Accessed: 13-11-2018 10:07 UTC
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/23279469?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Center for Migration Studies of New York, Inc., Sage Publications, Inc. are collaborating
with JSTOR to digitize, preserve and extend access to The International Migration Review
This content downloaded from 111.68.97.170 on Tue, 13 Nov 2018 10:07:59 UTC
All use subject to https://about.jstor.org/terms
In and Out of the Ethnic Economy:
A Longitudinal Analysis of Ethnic
Networks and Pathways to Economic
Success across Immigrant Categories1
Wendy D. Roth
Marc-David L. Seidel
Dennis Ma
Eiston Lo
University of British Columbia
The economic benefits to immigrants of taking jobs in ethnic work

places, relative to the open economy, are heavily debated. We exam
ine longitudinally differences across immigrant categories in how the
choice of ethnic or non-ethnic workplace influences the ethnic com
position of social networks and how these factors impact immigrants'
economic success. Using the Longitudinal Survey of Immigrants to
Canada, with data 6 months, 2 years, and 4 years after arrival, we
find support for both sides of the ethnic economy debate when it is
qualified by immigrant category. While economic immigrants benefit
from non-ethnic workplaces, family immigrants face economic penal
ties in the open economy and do better in ethnic workplaces. We
argue that policies sorting immigrants into visa categories do much of
the work of leading them into segmented paths of incorporation.
There has been considerable debate on the economic benefits to immi

grants of taking jobs in ethnic workplaces (Sanders, 2002). While some
claim that the ethnic economy provides an alternate route to mobility for
immigrants who may be penalized by language and cultural barriers in
'This research was supported by grants from Metropolis British Columbia and the Social
Sciences and Humanities Research Council of Canada. It was made possible through Sta
tistics Canada providing access to the micro-level data through the Research Data Centre
program. The data for this study were accessed at the Inter-University Research Data Cen
tre at the University of British Columbia, with the kind support of Lee Grenon and
Cheryl Chunling Fu. We would like to thank Irene Bloemraad, Rich Carpiano, Barry
Edmonston, Sylvia Fuller, Tomâs Jiménez, Karen Kobayashi, Mark Leach, Sharon Lee,
and Gerry Veenstra for helpful comments on earlier drafts of the paper. All mistakes
remain the responsibility of the authors.
© 2012 by the Center for Migration Studies of New York. All rights reserved.
DOI: 10.111 l/j.!747-7379.2012.00889.x
310 IMR Volume 46 Number 2 (Summer 2012):310—361
In and Out of the Ethnic Economy 311
the primary labor market (Wilson and Portes, 1980; Portes and B
1985; Portes and Stepick, 1993), others argue that immigrant emplo
experience disadvantages in ethnic workplaces relative to their counterpar
in the open economy (Sanders and Nee, 1987; Nee, Sanders, and Sern
1994; Nee and Sanders, 2001). The strong sociological interest in
question speaks to its significance for contemporary immigrant incor
tion. A traditional model of assimilation, developed mainly around E
pean immigration, sees immigrants advancing economically as t
integrate socially and structurally with those outside the ethnic group
the possibility of economic mobility within the ethnic community
relying on ethnic ties and resources suggests that economic incorpora
does not necessitate the abandonment of ethnic connections (Fong a
Ooka, 2002).
Examining how initial movement into jobs in the ethnic or open
economy influences subsequent ethnic ties is crucial to understanding the
mechanisms behind immigrants' potentially disparate paths to integration.
Many immigrants rely on ethnic ties for scarce resources and information
to facilitate their settlement (Hagan, 1998; Zhou and Bankston, 1998;
Menjivar, 2000; Marger, 2001). However, because of network homophily,
this social support may discourage immigrants from building an ethnically
diverse social network and make them more dependent on these types of
ethnic resources in the future. This can reinforce a "segmented trajectory
of adaptation" (Nee and Sanders, 2001), where those relying on ethnic
based social capital for mobility are likely to become more isolated from
the economic and social mainstream, while those relying on investments
in human capital for mobility are more likely to integrate into it.
Because longitudinal data on immigration are rare, most studies have
limited ability to distinguish the causal effect of workplace types on new
social ties and vice versa. The few longitudinal studies on immigrant inte
gration focus on changes in occupational status and income (Jasso and
Rosenzweig, 1995; Chiswick, Lee, and Miller, 2003, 2005a) or how
human capital changes and affects economic outcomes (Li, 2001; Chis
wick, Lee, and Miller, 2004, 2006; Kler, 2006). None of this research con
siders the relationship between job choice and social network composition,
and this remains a crucial gap in the immigrant integration literature.
To examine whether different immigrants experience segmented
trajectories of adaptation through the open and ethnic economies, it is
crucial to consider the types of resources that immigrants bring with
them to a new country. In their forms-of-capital model, Nee and
312 International Migration Review
Sanders (2001) recognize that immigrants arrive with different typ

capital - human, financial, and social capital — which help them inte
and shape their trajectories into the ethnic or non-ethnic economy.
who arrive with or develop high stocks of social capital and lower l
of human capital are more likely to become self-employed or work i
ethnic economy. Those who arrive with higher levels of human capit
more likely to move into the open economy.
Yet a perspective that focuses on the resources immigrants bring
new country and subsequently develop must recognize how immigr
policy either selects for or enhances these different forms of capital in d
ferent categories of immigrants. Distinct immigrant categories - suc
economic, family reunification, and refugee - are associated with diff
motivations, skills, and forms of capital. Economic immigrants
selected for their ability to rely on human or financial capital as the
mary resource for integration. In contrast, family reunification immigra
by definition have ethnic social capital in the form of support from
tives and often an established ethnic community. In some cases, fa
sponsors bear direct responsibility for the immigrants' integration.
gees do not necessarily arrive with any forms of capital, yet they are oft
sponsored by the government or private groups or citizens, enhanc
their likelihood of forming social ties outside their ethnic commun
Several studies consider how ethnic ties and social capital lead immig
to participate in the ethnic economy (e.g., Zhou, 1992; Waldinger, 1
Nee and Sanders, 2001), yet most do not have a broad sample ac
immigrant categories to focus on how the resources associated with
category shape labor market trajectories.2
Immigrant categories are not simply synonymous with the differ
types of human, financial, or social capital. Economic immigrant
selected in part for their "labor market suitability," or fit with th
nomic climate. Government officials, and sometimes employers, weig
needs of the labor market with the particular skills and resources
immigrants provide. Even if family immigrants and refugees are a
highly motivated and have labor market skills, they may face struct
employment barriers because they are not primarily selected on the
of their fit with the economy (e.g., Anderson, 1990; Gold, 1992; Gr
2Nee and Sanders (2001) consider the variation across respondents in levels of so
financial, and human-cultural capital, but do not describe the visa categories of
respondents.
1999; Reitz, 2001). Thus, even when various forms of capital are con
trolled, we would still expect immigrants' visa categories to affect their
labor market outcomes through the selection criteria applied (Li, 2000).
The sorting of immigrants into different visa categories does much of the
work of leading them into segmented paths of incorporation.
We examine how the choice of workplace type (ethnic or non
ethnic3) influences the ethnic composition of social networks — what we
call ethnic network portfolios - and how these two factors impact immi
grants' economic success. We offer three major contributions to this liter
ature. First, we examine how different categories of immigrants transition
between workplace types in the initial years after settlement, exploring
whether family immigrants are more likely to become entrenched in eth
nic workplaces. Second, we identify the causal relationship between work
place type and the ethnicity of subsequent social ties for these different
categories of immigrants, to determine whether being in an ethnic or
non-ethnic workplace leads to developing ethnic or non-ethnic ties that
reinforce that integration pathway. Third, we evaluate the income gains
that different categories of immigrants reap from these workplace types
and ethnic network portfolios (ENP) over this period.
We take advantage of a unique dataset, the Longitudinal Survey of
Immigrants to Canada (LSIC), to analyze the ENPs and economic out
comes of immigrants at 6 months, 2 years, and 4 years after arrival. Much
of the existing literature focuses on immigrants who have been settled for
many years. However, immigrants' potentially disparate paths to integra
tion may be affected by the initial choice upon the arrival of ethnic or
open economy jobs. Although we would expect lower levels of movement
into the open community during this initial period than after longer set
tlement, investigating this largely understudied period of early integration
can provide a micro-level analysis of the forces that lead new immigrants
toward different integration pathways. The dataset also allows us to con
sider these early pathways across all ethnic groups and on a national scale
- a considerable advantage because many of the studies that have
^Throughout this paper, we use the term "ethnic" to mean "co-ethnic." Ethnic workplaces
and ethnic ties refer to those associated with the immigrant's own ethnic group. Thus,
non-ethnic ties and workplaces may be associated with other immigrant and ethnic groups
or with the mainstream group. Developing ties with or working with a different ethnic
group indicates movement away from reliance on one's own ethnic group and toward a
labor market open to workers regardless of ascriptive characteristics.
advanced this literature look only at specific ethnic groups in particula

urban settings (Wilson and Portes, 1980; Portes and Bach, 1985; Sanders
and Nee, 1987; Nee, Sanders, and Sernau, 1994; Fong and Ooka, 2002
Logan, Alba, and Stults, 2003; Li and Dong, 2007).
Our findings emphasize the need to consider immigrant category in
examining immigrants' workplace trajectories over time, and in the ethni
economy debate more generally. Family immigrants are more likely to
remain within the ethnic labor market, and working in the ethnic labor
market leads to building co-ethnic friendships. Yet there are significan
differences between categories of immigrants in success outside the ethni
economy. Over this period of initial settlement, family immigrants who
do move into the open economy fail to reap its benefits in the same way
as economic immigrants; they experience an income penalty relative to
employment in the ethnic economy. Thus, we find support for both side
of the ethnic economy debate when it is qualified by immigrant category
While the open economy provides the greatest benefits to economi
immigrants and allows them to follow a traditional path of assimilation,
the ethnic economy offers family immigrants the best economic options
given the structural barriers they face in the open labor market.
IMMIGRANT INCORPORA TION AND ETHNIC ECONOMIES
In the late 20th century, immigration to North America rose at remar

able rates, bringing a diverse population to Canadian shores. In 200
19.8 percent of the Canadian population were immigrants, making the
nation second only to Australia in its acceptance of foreigners (Statistic
Canada, 2006a). The European share of the immigrant populati
declined in the late 20th century, while the proportion from Asia, th
Middle East, and Africa steadily increased. In 2006, 58.3 percent of all
immigrants in Canada were born in Asia and the Middle East (Statistic
Canada, 2006b).
This influx of immigrants to North America and the growing diver
sity of their origins have led many scholars to question whether the tra
tional assimilation model represents the experiences of these newcomers
it did for earlier immigrants. Several alternative accounts have grown o
of critiques of the assimilation model, often focusing on economic inc
poration through the ethnic community (see Nee and Sanders, 2001
The segmented or dual labor market approach holds that two or more
labor markets develop, with a secondary labor market characterized by
low-paying, low-skill jobs offering little opportunity for mobility (Piore,

1979). With few options for movement between the primary and second
ary labor markets, immigrants become trapped in the secondary labor
market and are paid wages that provide low returns on their human capi
tal. Labor thus becomes ethnically divided, resulting in ethnic antagonism
(Bonacich, 1972) and making assimilation less likely for immigrants who
start down that path.
The concept of the ethnic enclave economy, a particular type of
ethnic economy, derives from this segmented labor market approach
(Light et al., 1994). In their enclave economy theory, Portes and col
leagues maintain that some immigrants avoid the disadvantages of the
secondary labor market by establishing an enclave economy in which
many ethnic enterprises are concentrated in a distinct geographic loca
tion. This economy provides similar returns to human capital invest
ments as those in the primary labor market (Wilson and Portes, 1980;
Portes and Bach, 1985; Portes and Jensen, 1989; Portes and Stepick,
1993). Within the enclave economy, immigrants are not confined by
their language abilities or cultural traits and their foreign credentials
may be recognized. Furthermore, reciprocal obligations of bounded sol
idarity and enforceable trust help ethnic businesses succeed (Portes and
Sensenbrenner, 1993). Recent immigrants who find work in the enclave
economy tend to meet more people of their own ethnicity and are
more likely to form ethnic ties. Such ethnic ties may provide economic
resources to help immigrants start businesses (Light, 1972, 1984;
Hum, 2001), and ethnic enclaves encourage the growth of entrepre
neurship (Bailey and Waldinger, 1991).
Mixed evidence persists on immigrants' labor market success in
ethnic enclaves versus open labor markets (Sanders, 2002). Some argue
that ethnic enclaves limit opportunities for social mobility. Sanders and
Nee (1987) find that employees in an ethnic enclave receive lower returns
to their human capital than immigrants in the primary labor market,
although the returns for entrepreneurs are commensurate in the different
labor markets. Other studies similarly find that ethnic enclaves provide
immigrant workers with lower returns to human capital, or have no
advantage over the mainstream economy (Li and Dong, 2007; Xie and
Gough, 2009). Based on a study of Asian immigrants in Los Angeles,
Nee, Sanders, and Sernau (1994) argue that substantial movement
between employment sectors indicates porous boundaries rather than a
dual labor market with immigrants remaining permanently in the ethni

economy.
Spatially concentrated ethnic enclaves are not found among all
immigrant groups. Logan, Alba, and McNulty (1994) found few econo
mies in the United States that meet the definition of an ethnic enclave
and argue that enclaves are a special case of the broader concept of ethnic
economies. "The key feature of an ethnic economy," they claim, "is that
it is bounded by race, ethnicity, or national origin. An ethnic economy can
be defined as any situation where common ethnicity provides an eco
nomic advantage: in relations among owners in the same or complimen
tary business sectors, between owners and workers, or even among
workers in the same firm or industry regardless of the owner's ethnicity"
(1994: 693). The ethnic economy concept derives from the literature on
middleman minorities, which posits that minority groups concentrated in
small businesses occupy an intermediate economic position between elites
and workers. These middlemen, who often bear the brunt of hostility
from both workers and elites, develop a high degree of internal solidarity,
which makes them resistant to assimilation (Bonacich, 1973; Light et al.,
1994). Although stemming from different conceptual origins, both ethnic
economies and ethnic enclaves emphasize shared ethnicity as the basis for
upward mobility (Sanders, 2002; Hou, 2009).
We focus on ethnic economies in this paper, as our goal is to exam
ine the experiences of all recent immigrants to Canada, rather than only
those in ethnically concentrated areas. There are multiple forms of ethnic
economies, depending on the ethnic concentration of workers, pattern of
ownership, and the interaction between the two (Light et al, 1994; Logan
and Alba, 1999; Logan, Alba, and Stults, 2003). Because of the varying
levels of ethnic concentration and workplace formality, Nee, Sanders, and
Semau (1994) claim that metropolitan labor markets should be viewed as
a continuum from the ethnic economy to the open economy.4 Because
ethnic economies can have different benefits for workers and entrepre
neurs (Sanders and Nee, 1987, 1992; Light and Gold, 2000; Logan, AJba,
4For example, based on the combination of worker concentration and ownership, Logan,
Alba, and Stults (2003) identify an "enclave economy" (primarily ethnic workers and
owners); an "employment niche" (primarily ethnic workers, with non-ethnic owners); an
"entrepreneurial niche," (non-ethnic workers, ethnic owners); and non-ethnic economic
sectors (non-ethnic workers and owners). As suggested by Nee, Sanders, and Sernau
(1994), we view these forms as ranging along a continuum from the more ethnic to the
fully open economy.
and Stults, 2003), we must consider the outcomes of both employees in

ethnic-concentrated workplaces and the self-employed.
Much as with ethnic enclaves, there is also mixed evidence on the
benefits to immigrants of participating in broader ethnic economies.
Logan, Alba, and Stults (2003) find that work in ethnic sectors of the
economy has no consistent effects. Hou (2009) finds no difference in
earnings between ethnically homogenous and heterogeneous workplaces
among Canadian immigrant women, but finds that Canadian immigrant
men earn 33 percent less in ethnically homogeneous workplaces, although
most of this gap is explained by differences in demographic and job char
acteristics. These and other authors conclude that ethnic economies can
accommodate immigrant workers who lack human capital or have diffi
culty finding work in the mainstream labor market (e.g., Zhou, 2004).
Such studies consider immigrants who may have been settled for many
years, but say little about how initial economic integration after arrival
sets the stage for subsequent workplace transitions.
IMMIGRANT CATEGORIES AND WORKPLACE TYPES
Canada's immigration policy, like those of many traditional immigrant

receiving states, defines three principal categories of entry, or immigration
categories: economic, family, and refugee. Economic immigrants are made
up of two broad groups: skilled workers and business-class immigrants.
Skilled workers are selected for their ability to participate in the labor
market and to establish themselves economically. They are assessed on
selection criteria that stress education, language ability, and skilled work
experience. Business-class immigrants are selected for their ability to esta
lish themselves economically through entrepreneurial activity, self-emplo
ment, or individual investment (Citizenship and Immigration Canada
2007). The investor and entrepreneur subcategories require a substantia
transfer of funds as a precondition of acceptance. Entrepreneurs ar
expected to establish or buy a business that will create at least two full
time jobs for non-family members, and investors must demonstrate a
given net worth and invest a set amount ($400,000 in 2004) in a Can
dian fund for 5 years (Fleras and Elliott, 2007).
Economic immigrants are assessed on a point system for permanent
residence eligibility. Points are awarded for criteria that are deemed to
enhance the applicant's chances of contributing economically and of inte
grating socially, including their education level, fluency in one of the
official languages, work experience, age, whether they have alread

arranged employment, and adaptability. Among the criteria considered
under "adaptability," points are given to applicants who have relatives in
Canada, signifying the government's belief in the value of local social
capital for assisting integration. However, only 37 percent of economic
immigrants in our sample had relatives in Canada before immigrating.
Family immigrants are sponsored by a relative who is a citizen or
permanent resident. Family sponsors have to agree to support their famil
members economically for between 3 and 10 years to avoid creating a
financial burden on the government (Fleras and Elliott, 2007). Whil
some family immigrants may be well suited for economic success, they ar
not selected by such criteria, and as such are structurally less likely on
average to be well matched to the labor market.
The final category of immigrants is refugees, who are accepted under
humanitarian and legal obligations. This includes government-assiste
refugees, who are pre-selected by government officials; privately sponsored
refugees, who are supported by agencies, individuals, clubs, or religious
groups which are obligated to provide support for them for up to 1
years; and successful refugee claimants, who are not sponsored but apply
for and are granted refugee status (Citizenship and Immigration Canada
2007; Fleras and Elliott, 2007). Among refugees in our sample, 65
percent had relatives in Canada before immigrating.
States reveal different priorities in the type of immigrants they
accept (Borjas, 1999). Out of the immigrants accepted to Canada each
year, approximately 60 percent are economic, 25 percent are family imm
grants, and 15 percent are refugees (Citizenship and Immigration Canada
2007). By contrast, in the United States, about 65 percent of immigrants
come through family reunification, while about 10-15 percent are eco
nomic immigrants (Batalova, 2009). Unlike the United States, Canad
does not have a lottery system to allow low-skilled individuals a chance t
immigrate without being sponsored by a family member. AH immigrant
classes are expected to have resources at their disposal to help them adapt
While most refugees receive support from the government or a private
sponsor, other immigrants are largely divided into those seen as having
social capital and those with human or financial capital.
Immigrant admission policies and immigration categories influence
the likelihood that immigrants will integrate socially and economically
in their ethnic community or outside of it. Economic immigrants are
more likely to speak an official language, facilitating the development of
friendships with those outside their ethnic group (Chiswick, 1

Chiswick, Lee, and Miller, 2003, 2006). They are also likely to h
higher education levels, which helps give them a common profession
identity with non-ethnics of similar professional background (Nee a
Sanders, 2001; Chiswick, Lee, and Miller, 2005b). But in addition
these measurable human capital factors, this category also represents
immigration authority's assessment of economic suitability for the d
nant labor market; those selected are seen as meeting the current need
the mainstream economy. We can view the economic immigrant categ
as capturing this otherwise unmeasured assessment of economic
Together, these factors suggest that economic immigrants will be m
likely to find non-ethnic work.
Refugees are also likely to interact with the host society, throu
their sponsorship by the government or by private individuals or agen
Because many are sponsored by charities, religious groups, and other
country organizations, refugees may be settled in diverse areas of the
country away from other immigrants (Krahn, Derwing and Abu-Lab
2003; Mclsaac, 2003). Refugees who are not employed by their spons
may have difficulty in securing jobs. They are not selected based up
labor market fit, and their low levels of skills and qualifications may
reduced the likelihood of their immigrating under normal circumsta
(Li, 2000). This constrained opportunity structure encourages them
start their own businesses (Knight, 1983) and to fulfill an "appetite
entrepreneurship" (Wauters and Lambrecht, 2006).
Family immigrants, and policy shifts toward family reunification, ar
seen by some as producing diminishing economic returns for the na
{e.g., DeVoretz, 1995; Borjas, 1999). Such arguments imply that hum
capital and financial capital facilitate immigrant economic integratio
more effectively than social capital. One reason why family immigra
may contribute less to the nation's economy is that the majority are
spouses or parents of their sponsors. While many do enter the labor
ket, age and gender norms may limit their labor market options. An
son (1990) finds that family-class immigrant women are concerned w
jeopardizing their chances for obtaining citizenship, and thus work
whatever job is available to support their families. Others may have i
grated without the intention of working, but rather to assist their fa
by caring for children and households, or to peacefully live out the
retirement. Family immigrants on average have the lowest official
guage skills (Chiswick, Lee, and Miller, 2006). Factors such as housewo
childcare, and survival jobs limit the ability to benefit from govern
language training (Anderson, 1990). This can affect friendship patt
(McPherson, Smith-Lovin, and Cook, 2001) and limit the potential f
working in the mainstream economy (Green, 1999). It is likely that
who do enter the labor market have a better chance of finding eth
work, owing to their pre-existing family connections. And because
ethnic economy has more low-paying jobs that require lower educa
levels (Nee, Sanders, and Sernau, 1994; Catanzarite and Aguilera, 20
categories of immigrants with lower educational attainment, such as
ily immigrants, will likely be sorted more toward the ethnic econo
These factors suggest:
HI: Immigrant category influences the choice of ethnic or non-ethnic work, with family
immigrants more likely than other categories of immigrants to initially enter the ethnic
workplace.
At the same time, Nee, Sanders, and Sernau (1994) find that the
immigrants they studied tended to move away from the ethnic economy
over time and that the boundaries between sectors of the labor market are
fairly porous. While we would expect there to be some mobility even
among family immigrants, the inertial stickiness of the labor market
proposed by research on job and occupational embeddedness (Mitchell
et al., 2001; Lee et al„ 2004; Ng and Feldman, 2007) would suggest that
the inertia from initial entry into the ethnic workplace predicted by HI
would continue to anchor later choices. Thus:
H2: Family immigrants will transition into the non-ethnic workplace over time, but are
still less likely to do so than other categories of immigrants.
WORKPLACE TYPE AND ETHNIC NETWORK PORTFOLIOS
A longitudinal framework permits direct examination of how workpla

type impacts the development of social networks over time. Scholars ofte
discuss social networks as channels for individuals to obtain information
about job opportunities (Granovetter, 1973; Petersen, Saporta, and Seidel
2000; Fernandez and Sosa, 2005; Fernandez and Fernandez-Mateo,
2006), but the literature focuses less on the role that workplace types play
in the future development of social networks. The dense ethnic networks
in which the ethnic economy is embedded help immigrants generate social

capital to acquire resources and start businesses, providing jobs for other
co-ethnics (Portes and Sensenbrenner, 1993). Participating in the ethnic
economy and having ethnic social ties are seen as mutually reinforcing
and may account for why people remain in the ethnic economy even if
those positions are not advantageous in the long run. Immigrants working
in the ethnic economy interact less with non-ethnics and are less likely to
obtain information about events outside their community (Fong and
Ooka, 2002).
An ethnic network may provide different opportunities from a non
ethnic one (McGuire, 2000). Many assume that developing connections
outside the ethnic group will lead to employment in the open economy
but these assumptions are seldom tested. A notable exception is Hagan
(1998), whose study of Mayan immigrants to Houston illustrates the neg
ative consequences over time of investing in ethnic ties. Mayan women
who become tightly encapsulated in ethnic networks through continuing
private-household domestic work are disadvantaged compared to Mayan
men who have more diverse social connections with the non-immigrant
community through work, neighborhood, and recreation. Those weak ties
provide the men with information about new job opportunities, while the
women's pool of economic resources diminishes over time.
In this paper, we are also interested in the reverse causal chain - the
role that workplace choices make in future social capital development. We
use the concept of an ENP to define different types of social networks
with regard to ethnic composition. The network portfolio concept was
originally discussed in terms of the composition of an individual's mixture
of strong and weak ties (Seidel, Polzer, and Stewart, 2000), but has also
been applied to other forms of network composition such as ethnicity
(Seidel et al., 2007). An individual's social network can be understood
as similar to a financial investment portfolio. In a common investment
portfolio, investors have different percentages of cash, bonds, and equities.
Some people may invest in greater amounts of stocks, while a more
risk-averse person may invest 100 percent in cash or bonds, and many
people fall somewhere in between. The same concept can be applied to
how an individual invests in his or her social network. With the ENP
concept, the question is how an individual diversifies his or her social
network based on the ethnic characteristics of social ties.
Some individuals have a diversified ENP, which is an ethnically diverse
social network. Others have a non-diversified ENP, or ethnically
homogeneous network, which is categorized by two subtypes: similar and

different. A similar non-diversified ENP is composed primarily of other
sharing the focal actor's ethnicity, while a different non-diversified ENP
consists mainly of individuals of a single ethnic group that differs from
the individual's own. For convenience, we refer to the similar non-divers
fied ENP as an ENP, while a non-ENP includes both diversified ENPs
and different non-diversified ENPs.
As we are interested in how workplace trajectories affect the ENP,
changes to the ENP over time are more important than its overall compo
sition, as this focuses on what social ties the immigrant adds within a
specific period. For this reason, we focus in this paper on the ethnic com
position of new friendships. In keeping with our portfolio analogy, we
can think of these as investments in an ethnic or non-ENP.
Many scholars of network homophily argue that individuals' social
network formation is influenced by the social composition of their sur
roundings and activities (Feld, 1982; McPherson and Smith-Lovin,
1987). We similarly argue that immigrants who work in an ethnic work
place spend more time interacting with co-ethnics and are more likely to
form ethnic ties. This suggests:
H3: The choice of ethnic or non-ethnic work influences investments in ENPs, with ethnic
jobs leading to new ethnic ties and non-ethnic jobs leading to non-ethnic ties.
ECONOMIC OUTCOMES OVER TIME
Pathways of immigrant incorporation have a temporal element by defi

tion. While some focus on intergenerational change, those concern
with the experiences of first-generation immigrants recognize that ec
nomic incorporation and social incorporation occur over time, particu
larly as immigrants move between economic sectors. Studies of Asi
immigrants in Los Angeles (Nee, Sanders, and Sernau, 1994) and C
nese immigrants in Toronto (Fong, Chan, and Cao, 2009) advance th
literature by gathering retrospective employment history data. Yet lon
tudinal data are generally preferable to retrospective data, which are su
ject to memory bias (Moss and Goldstein, 1979; Menard, 2008). Fo
instance, people are more likely to have faulty memories of the ethni
composition of their workplaces as time passes than they will abo
recent positions.
Changes from the early stages after immigration to periods when

immigrants are relatively more settled are also important to examine in
light of Portes and Jensen's (1987) assertion that their enclave hypothesis
"never claimed that perennial confinement within an ethnic community
would be the road to economic parity. The hypothesis simply stated that,
for newly arrived immigrants, participation in a pre-existing ethnic econ
omy can have positive economic consequences, including a greater oppor
tunity for self-employment [italics added]." While it is not clear how
these authors define "newly arrived," they nonetheless make a temporal
argument about early entry into an ethnic workplace leading to positive
later transitions, including higher rates of entrepreneurship. To test such
ideas, we must examine the differential returns to distinct pathways of
immigrant employment, both initially after arrival and over time.
Part of the return to human capital obtained from a job is the
nature of the match between an individual's skills and the demands and
rewards of the job. Newly arrived immigrants, who are likely to have
fewer contacts, foreign credentials, and language difficulties, may find
jobs that are a good match for their skills more easily within the ethnic
economy (Wilson and Portes, 1980). Those who move directly into the
open labor market may be relegated initially to jobs that are a poor
match for their skills:
H4: Initially after arrival, ethnic workplaces offer greater income returns to human capital
than jobs new immigrants find in the open labor market.
Even if ethnic jobs provide an immediate income advantage, immi

grants who find jobs through the open labor market should eventually catch
up, particularly if they are well suited to the labor market. Research suggests
that immigrants have the potential for upward economic mobility, particu
larly from the lower economic segments (Bean, Leach, and Lowell, 2004).
In Los Angeles, Asian immigrants eventually gravitate away from the ethnic
economy, and reduce their reliance on ethnic networks for job seeking, in
an effort to improve their job prospects (Nee, Sanders, and Semau, 1994).
We argue that immigrants who work in non-ethnic jobs, even if they are less
financially rewarding immediately, reap more benefits from diversifying
their social capital and eventually move to positions with higher earnings:
H5: Over time, non-ethnic workplaces will offer immigrants greater income returns to
human capital than ethnic workplaces.
Diversifying a social network leads to new sources of informat

experiences, and opportunities. Past network studies have warned o
hazards of being overembedded in networks where the focal actor c
gain access to new opportunities owing to the homogeneity of info
tion offered, and have stressed the importance of diversification (
Granovetter, 1973; Uzzi, 1999). While ethnic minorities may be att
to homophilous ties owing to the valuable resources of social suppo
non-homophilous network ties should yield more diverse informatio
opportunities (e.g., Mollica, Gray, and Trevino, 2003). Thus, we
that immigrants reap more economic benefits from diversifying t
social capital regardless of workplace type:
H6: The development of new non-ethnic ties will offer immigrants greater income returns
than the development of new ethnic ties.
DATA AND METHODS
We used the restricted-access micro-level data of the Longitudinal Surve

of Immigrants to Canada (LSIC) from Statistics Canada.5 The LSIC
target population is immigrants aged 15 and above who arrived betwee
October 1, 2000, and September 30, 2001, and who remain in the coun
try at each wave of data collection. The sampling frame comes from t
Field Operation Support System of Citizenship and Immigration Canad
the government's administrative database that contains all permanent re
dent immigrants. This target population at Wave 1 is approximate
164,200 individuals. Individuals who applied for immigrant status from
within Canada are not part of this target population, focusing the data o
immigrants' experiences when they first arrive.
Because recent immigrants are an extremely mobile population
Statistics Canada followed a rigorous procedure to track and locate the
respondents at each wave. Case files were forwarded to regional office
who forwarded them to interviewers. Respondents were tracked via feder
'Statistics Canada established the Research Data Centre Program to allow affiliate
researchers access to restricted-access micro-level data in several of its datasets. A condition
of using these data is that outputs are screened to ensure large enough cell sizes to preve
respondent identification. In some cases, this requires that categories be combined or th
small categories be omitted from published results. Analyses must also report weighted N
and provincial databases, such as provincial healthcare records, and were

verified by date of arrival and birth date matching. Potential respondents
that were not found by the interviewers were forwarded to a regional trac
ing team that utilized multiple methods to locate them, including elec
tronic phone books and administrative files from Citizenship and
Immigration Canada.
The survey data include interview waves conducted 6 months,
2 years, and 4 years after arrival.6 Following the lead of past research
(e.g., Green and Green, 1995; Chiswick, Lee, and Miller, 2005b), we
focus our analysis on principal applicants. This fits our interest in the role
of immigration policies in selecting immigrants with different forms of
capital, which accompanying partners and dependents may not have.7 We
include immigrants who are present in all three waves and are between
the ages of 18 and 65 (to focus on workplace experiences). For our popu
lation of interest, this provided a dataset of 4,451 respondents, represent
ing 92,936 weighted cases.
Statistics Canada conducted the interviews in 15 languages, with the
majority done face-to-face and the remainder conducted by phone. Each
respondent was assigned a weight designed by Statistics Canada to
represent the entire target population accurately. This final weight is built
on a series of cascading adjustments that takes into account design issues,
non-response, resolved case adjustment, and post-stratification. As
recommended by Statistics Canada, we utilized Stata BRR bootstrapping
procedures with proportional weighting using this set of final replicate
weights. This allows us to generalize results to the entire target popula
tion. Although weights are appropriate in regression models for complex
survey designs using stratification and clustering, we also ran our models
using unweighted data (see Winship and Radbill, 1994). The results were
substantively unchanged. As such, we present the bootstrapped propor
tional weighted models, following Statistics Canada's recommended analy
sis strategy for this complex sample design.
6While the time between the waves of data collection is short, past research that has used
the LSIC (e.g., Schellenberg and Maheux, 2007; Xue, 2008) has found enough variance
across waves for meaningful analysis.
7In the case of refugees, no distinction is made between principal applicants and accompa
nying family members. However, less consideration is given to the forms of capital refu
gees bring with them, and we expect that all members of the family are likely to
experience similar psychological processes.
Dependent Variables
We use the following dependent variables at each interview wave: workpla

type, ethnicity of new friends, and personal income. For our work
type variable, one advantage of our study is that we can measure the eth
composition of the respondent's workplace directly, rather than relyi
the ethnic group's overall concentration in a particular geographic are
nomic sector, or industry. Many studies using census data which adopt th
proxies for firm-level measures are prone to Type II error - that is, inclu
individuals that do not work in an ethnic economy (Sanders and Nee,
Portes and Jensen, 1989; Logan, Alba, and Stults, 2003). An ideal oper
alization of the ethnic economy using firm-level data is that used by Fon
his study of the Chinese ethnic economy in Toronto (Fong and O
2002; Fong, Chan, and Cao, 2009). Fong identifies individuals as parti
ing in the ethnic economy when they meet any one of the following
tions: (1) self-employed; (2) working in a firm where at least 40 perce
the management is co-ethnic; or (3) working in a firm where at least 40 p
cent of their co-workers are co-ethnic. A limitation of our data, however
that the LSIC only gathered information about the ethnicity of empl
managers or bosses in Waves 1 and 2. To keep our definition consis
across waves, we define individuals who meet Fong's conditions 1 or
particular wave as participating in an ethnic workplace. We include
ethnicity of the individual's managers or bosses as an independent variabl
the waves where the data are available. This approach, while reducing
II error, leaves us prone to Type I error — excluding from our definition
ethnic workplaces those who have mostly non-ethnic workers but
ethnic bosses or managers. However, only 1.9 percent of responden
Wave 1 and 1.5 percent at Wave 2 fall into this category. We ran addi
models including these individuals in our definition of ethnic workpla
Waves 1 and 2 and found little difference.8 We therefore view our an
as providing a conservative estimate of participating in an ethnic workpla
We define the workplace type variable for individuals who wor
at an income-producing job or business in each wave. For those who
income-producing work, we define those in an ethnic workplace as
the self-employed or (2) employees who said that "all" or "most" of t
co-workers were of the same ethnic or cultural group as themselve
others were coded as being in a non-ethnic workplace. When respond
8These analyses are available from the authors upon request.
had more than one job during the interview period, we coded workplace
type for the respondent's current or most recent job. In the event of two
or more current positions or multiple jobs that ended on the same date,
we selected the respondent's "main job," as determined by LSIC.9 We
then coded workplace type as a dummy variable with non-ethnic work
place coded as 1 and ethnic workplace coded as 0.
For the variable ethnicity of new friends, respondents were asked a
series of network questions about new friends made in a variety of ways
(detailed below) since arrival (in Wave 1) or since the last interview.
Those who made any new friends were asked: "how many of these new
friends belong to the same ethnic or cultural group as you?" We catego
rized those who answered "all of them" or "most of them" as having
made mainly ethnic new friends (coded as 0), and those who responded
"about half of them," "few of them," or "none of them" as mainly non
ethnic new friends (coded as 1).
For our final dependent variable, personal income, we take the log
of the respondent's personal income from all sources,10 plus one,11 for
those who had earned an income from working at a job or business in
the 12 months prior to the interview, or since arrival in the case of Wave
1. While personal earnings would be a preferable measure, unfortunately
Statistics Canada did not collect these data. Statistics Canada did report
household earnings; as such, we re-ran all of our models using this less
precise measure of personal earnings. While the majority of results
remained substantively unchanged, we found that some of the results for
sex and marriage changed, as might be expected, with the income penalty
for women disappearing when looking at total household earnings.
Considering that we are focusing on individual-level factors, we use the
best individual-level measure available in our reported models, while
recognizing our lack of a measure of personal earnings as a limitation of
the study.
9In the event of multiple jobs with the same characteristics, LSIC used the following
ordered criteria to select the main job until the tie was broken: (1) the job involving more
hours worked, (2) the job involving more weeks worked, and (3) the position listed first
on the job roster.
1 "Respondents were asked how much personal income they received from all sources,
inside and outside Canada, not including money withdrawn from savings.
"Because we cannot take the log of 0, we follow common practice and add 1 to each
respondent's income to accommodate for models including respondents with no income
(see Bartlett, 1947).
Independent Variables
Immigration category is coded as a set of dummy variables based u

respondents' CIC-determined immigration categories: family, economi
refugees and others.12 The economic category consists of skilled workers
business-class immigrants and is used as the reference group in our m
In our population of interest, 30.7 percent of respondents are family
grants, 62.4 percent are economic, and 6.9 percent are refugees and other
We also include dummy variables related to the characteristics
ethnic workplaces. As employees were asked at Waves 1 and 2 if th
supervisor or boss is of the same ethnic or cultural group as themsel
we include a dummy variable employed with an ethnic supervisor f
these waves. Because our definition of ethnic workplaces includes bot
self-employed and employees with mostly ethnic co-workers, w
include a dummy variable for self-employed, to distinguish the effe
being an entrepreneur versus employed in an ethnic workplace.
When using ethnicity of new friends as an independent variable
use a categorical variable that includes no new friends, mainly non-e
new friends, and mainly ethnic new friends, with the latter category
ted. Similarly, when using workplace type as an independent variable
define it with three categories: not working, working in non-ethnic
place, and working in ethnic workplace, with the latter omitted. We
similar specification in our analysis of transitions between workplace
(Table 2) except that there we divide the working in ethnic workpl
category into the finer-grained self-employed and employed in eth
workplace to examine transitions between these states as well. In the
vant multivariate models (Tables 4 and 5), we include workplace tra
tion variables, with variations in beginning states (not working,
ethnic workplace, and ethnic workplace) and end states (non-ethnic w
place or ethnic workplace) between waves; here, ethnic workplace in
ous wave and current wave is our reference category.
12"Refugees and others" includes the CIC subcategories government sponsored refug
privately sponsored refugees, other refugees abroad, and other immigrants abroad. T
ter category includes post-determination refugee claimants, deferred removal or
humanitarian and compassionate cases, sponsored humanitarian and compassionat
outside the family class, and people granted permanent resident status based on pub
icy considerations (Citizenship and Immigration Canada, 2007).
We use a series of dummy variables to indicate whether the respon

dent made new friends in the following ways during each interview wave:
through relatives or friends in Canada; through relatives or friends in the
old country; through an ethnic association or club; from a religious activ
ity; from an English-as-a-second-language or French-as-a-second-language
class; from other classes; from work; or from other means.
Control Variables
Age is directly related to labor market outcomes (Hansson et al„ 1997).

As such, we coded respondents' age continuously in years. Marriage
impacts social networks (McPherson, Smith-Lovin, and Cook, 2001) and
labor market outcomes (Waite, 1995). We treat married as a binary vari
able at each wave, coded as 1 for married and common-law marriage and
0 for divorced, separated, never married, and widowed. Language skills
are directly linked to labor market outcomes (Dustmann and Fabbri,
2003) and also are likely to influence the choices people make regarding
new friends (McPherson, Smith-Lovin, and Cook, 2001). Speaks official
language is coded as 1 when Quebec-based respondents report speaking
French well or very well or when respondents outside Quebec report
speaking English well or very well, and 0 otherwise.
Job class variables are coded at each wave for the respondent's cur
rent/most recent job (or "main" job if they have two or more cur
rent/most recent jobs), using the LSIC Employment Roster & Details
data. Because the types of jobs available in ethnic economies are more
likely to be low-skilled (Piore, 1979), immigrants in professional jobs may
be more likely to be in the open economy than those in low-status jobs,
and may develop more diverse social ties in their workplace interactions
(Roth, 2012). We recoded the LSIC job classifications into three catego
ries: low status, medium status, and high status,13 and use low status as
our reference group. Hours worked is a continuous variable of the num
ber of hours worked in an average week.
13High-status occupations include managerial and professional occupations. Medium-status

positions include technical, sales, administrative support, and military occupations. Low
status jobs include service occupations; precision production, craft, and repair occupations;
and operators, fabricators, and laborers. This division parallels research using the occupa
tional category groupings provided in the variable documentation of U.S. Census micro
level data (Roth, 2008, 2009). The coding map is available from the authors upon
request.
Static control variables that have been shown to be linked to labor

market outcomes and/or social networks (McPherson, Smith-Lovin, and
Cook, 2001; Chiswick, Lee, and Miller, 2005a) are measured at the first
interview. These include female (a binary variable coded as 1 for women
and 0 for men); ethnicity (a set of dummy variables based upon ethnic back
grounds: Chinese, South Asian, Black, Filipino and Southeast Asian, Latin
American, Korean, Arab, and Other Ethnicities, with White as the reference
category); education (a set of dummy variables for high school, college
level,14 bachelor's degree, and graduate degree, with less than high school as
the reference category)15; contacts in Canada before arrival (a dummy
variable coded as 1 if respondents had relatives and/or friends residing in
Canada before they immigrated); and CMA ethnic concentration (a contin
uous measure of the proportion of co-ethnics in the census area that they
arrive in). To create the CMA ethnic concentration variable, we obtained
information from the 2001 Census on the ethnic composition of Census
Metropolitan Areas and Census Agglomerations (Statistics Canada, 2001)
and tabulated the proportion of each landing CMA/CA that shares the focal
respondent's ethnicity. We summarize the descriptive statistics for our static
and longitudinal variables in Table la and b, respectively.16
Sample Attrition
Our study represents the population of immigrants that arrive and remain
for at least 4 years, rather than all immigrants. Some immigrants left the
country prior to each wave of data collection, which results in a smaller
target population. At Wave 1, the target population was approximately
164,200; at Wave 2, it was 160,800; and at Wave 3, it was 157,600
(Statistics Canada, 2007). From the original Wave 1 sample of 12,040
immigrants, 9,322 were re-interviewed in Wave 2 and 7,716 were
re-interviewed in Wave 3.
Between Waves 2 and 3, 17.2 percent of all 18- to 65-year
olds were lost because of attrition. It is common for attrition rates in
14Canadian colleges typically offer vocationally oriented programs and grant diplomas and
certificates, not degrees.
15We use education level at the first interview for all waves of analysis to control for
human capital upon arrival.
I6We believe the large percentage of refugees in the "Other ethnicity" category reflects
Iranians, as the majority conducted the interview in Persian.
^ \U >U sU sU sV sW \U xSr1 \V \W
a 0s- o~~ o~~ o^ o^ oN a- o~~ o~~
O O CMA
(N (N pcOCN^r^t\fNfNCs
(N CN N (N \6 <7\ rifNCNforifNCNiAvd
(N rN cr> rN — |\ vo
CC) © 00 \Q
K OfSfi^CNNinri o
(N (N (N n M CNCNCO
v« vU sV si) \U \W \U \U vU yU
0s ©•* 0s o" ff- a~ o~- 0s 0~- O" 0~
inqoo^fsooqooxf
r-H
Xfco CNNrifONOVONH d(N(N-!rno6vovO(N

rri(N -OCOiTNrriiAiACNÔ^ — ia m oo po vo
TABLE 1
sP vP sP vP \P vP sP vp \y nH \U >p \P \P vU \U -P \P
ff- 0s ©"* 0s- 0s 0s 0s 0s 0s o"- u1 o~~ 0s- d- 0s O1'
Longitudinal Variables NcniAxrOMOiAin O^lNiTiOOVnCNCOO
OCN\£)vAcdfAfO\Ovd
H u
3
on
• o
y i
c/i e
« °
c £
rt O
U "O
_c o ^
<
s
u
>^"0
u C
I u' £ "2
o <2 u
3Jj rr u H
o .
 o
.S 3
si S -5 -m ■§ ■s u> SPji_ -d c c
rt o°
_2 JI
C V o _0,,
DescriptvSasicforWkingAe(18-65)PrincpalA nts,byImigrant()Cegory:StaicVrbles;()CatgoryndWave: =3 "3 o rt " s
W)
'31 (j aJ to rt
^ £ JP u<lsE.3£i=6 J
*2
W
TJ
PJ 211#
International Migration Review
^ ^ ^ - •P -p >p l vP Vp sO
o- o ' 0s 0s o^
(S N > co co Nf
vP vP vP vO vP O
o~- e- p- ©•- o
m © vo fs
t t ^ ^ ^ o
s"3
~a sO
c?- o~- o
^"00
N Cn fO
— CO
c? ^ s? 8=
O 00 (N O CO oo oo xj" I/N Cv ^ "
K c\ cn tn xr ON N rj CO •—< C\ «/"N
-«r co ,-. vo co - rfiNf N in H
<n cn
^ -S
xC xO0s vO sp
xr ^r »n o oo o
| ri ^
•S^.s
•P >P vP vP ' vP vP vP v© u -
©- cf~ <— •- CTv «~)
CO 00 CN CN CN . C\ vj- 00 O >. <N f~
^ oo o xf
i/~\ cr> ir\ co ffcT2
tudinal Variables
TABLE 1 (Continued) C o '~
SF 2? #^ ^ ,«1 m h>
U f\ o
<n ^q <n ir\ —
vq© © >o o -_
o^xr
cci -i !A VO N K co cs o .3 !r^ ri
M CO OO CO —• ^
■g 11 ^
2^g
^ -o ^
o a .c
&C~bb O
0s tf ^ ^ ^ ^ .£ *c >5
N IA q °o n co co vo 5 £A
co K oo' —• © © K oo ( "
XJ- s^J. — r-. ~ (
Jî
■- -£ "2
# # # # # # # #
— co oo ■ — o cN o o
60'C
l||
3 f!
•3 « -§
U (5 r
<-E T3 -rt ^ V
^ > I-M
¥S
a>-g
Jl ~Z u
c _2
5 rr
"O
8^1
ftb-O rt
i •* u t g
. 2 Si too
IB . o -£ o JO U cjs.c
■? c £ CL
■•& § u E
J3 £ T? "hn"
js SP.S o-s-fj
I "S » a s^s-S •!•!_§■_ _c i±: _c n>
DescriptvSasicforWkingAe(18-65)PrincpalA nts,byImigrant()Cegory:StaicVrbles;()CatgoryndWave:Logi
OU _£J jSiW gsss szzS^
2
longitudinal data to be relatively large (Borjas, 1989). We ran a binary

logistic regression to determine what influenced the likelihood of disap
pearing from the dataset between Waves 2 and 3.17 We found that immi
grants who are married and who are older have significantly lower odds of
dropping out of the survey. Several of the factors that lead to immigrants'
disappearance suggest that many leave because they fail to achieve eco
nomic or social integration. Groups that may experience discrimination are
more likely to drop out of the survey: The group with the greatest increase
in the odds of disappearing (57%) is Blacks (relative to Whites). South
Asian immigrants also have 33 percent greater odds of disappearing. Those
self-employed in Wave 2 are also more likely to drop out of the survey in
Wave 3 (our multivariate results support the notion that this group may
experience an economic disadvantage at Wave 2).
These findings are consistent with studies of return migration; immi
grants who are not attached to paid employment, who have poorer labor
market outcomes, and who have not cultivated strong ties in the host
society are more likely to return to their country of origin (Borjas, 1989;
Constant and Massey, 2003). Many immigrants have and do exercise the
option of return as a strategy in the face of social and economic barriers.
We believe it is likely that many of the people who drop out of the data
set do so by leaving the country because of the considerable efforts and
strategies Statistics Canada undertakes to track and locate immigrants,
and because the survey is mandatory. Those who do not leave the coun
try, but become untraceable by Statistics Canada are likely to have
dropped out of the workforce. As such, those on the lower ends of the
income distribution in our analysis may still be better off economically
than those who were lost to attrition. This would make it more difficult
to find a statistically significant income difference among the remaining
respondents, and makes our tests for income disparities more conservative.
The Relationship between Human Capital and Immigrant Category
We have suggested that economic immigrants are selected on the basis

of their human capital, while family immigrants are selected because of
17It was not possible to determine the percent of attrition between Waves 1 and 3 or to
analyze the likelihood of disappearing between Waves 1 and 2, because the original Wave
1 data file containing the full sample was not provided by Statistics Canada; Wave 1 vari
ables were included in the Wave 2 data file for those who remained in Wave 2.
their social capital, which raises the question of whether immigrant ca

egories are so strongly associated with these concepts that they cannot
be analyzed separately. Economic immigrants do have very differen
skill profiles overall than family immigrants and refugees, yet we found
that there is still notable diversity in human capital measures withi
each immigration category. Members of each immigrant category are
represented at every outcome level in the education and speaks off
cial language variables. Just as there are economic immigrants who
do not speak English or French and have less than a high-schoo
education, there are family immigrants and refugees with graduat
degrees.
In Wave 1, for example, 82 percent of economic immigrants are
proficient in an official language, followed by 50 percent of family
immigrants and 36 percent of refugees and others; although the differ
ences are apparent, they do not represent complete association. By wave
3, these numbers are 87, 63, and 61 percent, respectively. It is not sur
prising that human capital differences exist between immigrant catego
ries, as these are key evaluation criteria for the admission of economic
immigrants. However, the three immigrant categories receive sizeable
representation across the range of each human capital component. There
are likely still unobserved labor market suitability differences between
the visa categories, above and beyond the observed human capital fac
tors. These differences are captured by the dummy variables included
for each visa category.
Models
We first present descriptive analysis of immigrants' transitions between

not working, self-employment, being employed in an ethnic workplace,
and being employed in a non-ethnic workplace. We then present multi
variate analysis with sets of weighted regressions for the dependent vari
ables, each starting with Wave 1, followed by Wave 2 and Wave 3.
Therefore, each dependent variable is predicted three times, once per
wave, using a comparable set of independent variables. We update some
time-varying independent variables for each wave, as appropriate. For
example, we control for marital status in Wave 2 when predicting a
ISAmong the economic immigrants, there are both skilled workers and business-class
immigrants at every level in these variables.
dependent variable in Wave 2. Other static variables, such as immigrant

category, do not change across waves, and we use the answer supplied
in Wave 1. Where independent variables change across waves, indicators
such as Wl, W2, and W3 are placed before coefficients in the tables.
Some independent variables are lagged by one wave where appropriate
to clarify the causal direction. For example, our analysis of being in a
non-ethnic workplace (Table 3) in Wave 2 uses independent variables
for mainly non-ethnic new friends made in Wave 1. Unfortunately, the
data do not report information at the time of arrival, so the Wave 1
models are not able to be lagged comprehensively. We ran the Wave 2
and 3 models using non-lagged data, and the results were not substan
tively different. Our models of workplace type and income (Tables 3
and 5) are restricted to those currendy working to focus on the compar
ison between ethnic and non-ethnic workplaces, while our analysis of
the type of new friends made (Table 4) focuses only on those who
made new friends in that wave.
DESCRIPTIVE RESULTS
To understand the patterns in how immigrants move from one type of

workplace to another over time, we present descriptive transition tables
for workplace type (Table 2). We use proportional weights to have our
sample represent our population of interest. However, owing to Statistics
Canada's privacy restrictions, we are unable to include table cells where
the unweighted sample size is below 10 cases. Therefore, we can only
display data for family immigrants and economic immigrants and must
omit rows or columns featuring cells with too few cases.19 This means
that overall percentages in the various quadrants of Table 2 cannot be
directly compared when the rows or cells present are not consistent. To
mitigate this reporting challenge, we present the weighted numbers in
the table and draw out and calculate appropriate comparisons in the
text.
Workplace transitions of family and economic immigrants are

shown from Wave 1 to Wave 2 (a and c) and from Wave 2 to Wave 3
(b and d). Out of all immigrants working for someone else {i.e., not
self-employed), a larger proportion of family immigrants work in ethnic
workplaces compared to economic immigrants in all waves, including
19As a result, some margin percentages for Wave 2 differ slightly across transition tables.
TABLE 2
TransitobewnWorkplaceTyforWkingAe(18-65)PrincpalAicnts,byImigrantCegory,%(N)
Wave 1, which supports HI and H2. In successive waves, about 38,20 33,
and 27 percent of such family immigrants are in ethnic workplaces, com
pared to around 18, 13, and 11 percent of economic immigrants, respec
tively.21 Also supporting H2, the ratio of family immigrants employed in
a non-ethnic workplace to those in an ethnic workplace increases over
time (around 1.60 in Wave 1; 2.03 in Wave 2; and 2.65 in Wave 3), but
it also remains consistently lower than the ratio for economic immigrants
(around 4.53 in Wave 1; 7.03 in Wave 2; and 8.05 in Wave 3), who
show much greater movement into the open economy over time. The per
centage change in this ratio for family immigrants is around 27 percent
[(2.03-1.60)/1.60] from Wave 1 to Wave 2 and 30 percent [(2.65
2.03)/2.03] from Wave 2 to Wave 3. For economic immigrants, these fig
ures are around 55 and 15 percent, respectively. This suggests that at the
2-year mark, economic immigrants' workplace trajectories plateau and sta
bilize to a large degree, while family immigrants are still gradually moving
into non-ethnic workplaces, revealing their slower movement into the
open economy.
Analysis of the transitions in Table 2 from ethnic to non-ethnic
workplaces reveals a considerable difference between family and economic
immigrants. If we look at the proportions of employed immigrants that
start in an ethnic workplace and either stay there or transition into a non
ethnic workplace, around 33 percent [1782/(1782+3564)] of family
immigrants transition from an ethnic to a non-ethnic workplace from
Wave 1 to Wave 2; of such family immigrants in Wave 2, 39 percent
[2189/(2189+3487)] transition into the non-ethnic economy in Wave 3.
For such economic immigrants however, 61 percent transition from an
ethnic to a non-ethnic workplace from Wave 1 to Wave 2 and 58 percent
from Wave 2 to Wave 3. This is even more transition from the ethnic to
non-ethnic economy than was found by Nee, Sanders, and Sernau (1994)
20These proportions are calculated by dividing those employed in an ethnic workplace by

the sum of those employed in an ethnic workplace and those employed in a non-ethnic
workplace for each wave, for example for family immigrants in Wave 1
[5346/(5346+8542)]=38%.
2'Because some rows with small cell counts are omitted, these figures are slightly over
reported with the exception of the figures for family immigrants in Wave 2 and economic
immigrants in Wave 3 (because these are based on a full set of cells). However, because
the omitted cells are so small, they provide little alteration to the final proportion reported
(<0.5%). We assure that the overall scope and trend of these figures remains consistent
with and without the cells omitted in this table.
in their study of Asian immigrants*in Los Angeles, supporting their

ment of porous boundaries between these labor markets. However,
data also show that immigrant category matters and economic immig
are much more likely to transition from an ethnic to non-ethnic w
place than family immigrants, a finding that generally supports th
ment that immigrants with greater human capital are more lik
transition from the ethnic to the mainstream economy (Fong, Chan
Cao, 2009).
Looking at immigrants who start in a non-ethnic workplace and eith
stay there or transition to an ethnic workplace, only 16 perc
[1337/(1337+7206)] of family immigrants who are employed in
ethnic workplace in Wave 1 move to an ethnic workplace in Wave 2;
figure decreases to 11 percent between Waves 2 and 3. Out of all co
rable economic immigrants, even fewer move from non-ethnic to et
work: 7 percent [2033/(2033+27378)] from Wave 1 to Wave 2 a
percent from Wave 2 to Wave 3. Immigrants in both categories are
likely to move toward non-ethnic workplaces and stay there, altho
economic immigrants seem to be more successful at it.22
Very few immigrants are already self-employed by Wave 1. By W
2, however, approximately 7 percent [3422/49597] of economic
grants who are working are self-employed (compared to 12% emplo
in an ethnic workplace and 81% in a non-ethnic workplace). T
increases to 11 percent by Wave 3 (with 10% in an ethnic workplac
79% in a non-ethnic workplace). Among family immigrants, remar
few become self-employed in any wave. Even in Wave 2, the percen
self-employed is too small to report. By Wave 3, only 4 pe
[865/(865+5298+14028)] of family immigrants who are working are
employed (compared to 26% in an ethnic workplace and 69% in
ethnic workplace). The higher rate of entrepreneurship among econ
immigrants is not too surprising; government policies and officials
some economic immigrants based on their ability to start a business.
22It is more common for immigrants who are not employed at one wave to move i
non-ethnic workplace in the next wave than for them to move into an ethnic work
Out of the group that goes from not working in Wave 1 to being employed in W
74 percent [4068/(1433+4068)] of family immigrants and 86 percent
[9176/(1488+9176)] of economic immigrants move into employment in a non-ethnic
workplace (versus 26% and 14% into an ethnic workplace). From Wave 2 to Wave 3, 72
percent of family and 88 percent of economic immigrants move into a non-ethnic work
place (versus 28% and 12% into an ethnic workplace).
Of all economic immigrants who start out employed in ethnic wo

places in Wave 1 and continue working, approximately 9 perce
[595/6696] of them move into self-employment in Wave 2, while onl
percent [942/30354] of economic immigrants who start out in non-et
workplaces transition into self-employment from Wave 1 to Wave 2.
most common, however, for economic immigrants who are not work
in Wave 1 and find work in Wave 2 to become self-employed in Wave
at 15 percent [1885/12548] of those not working at Wave 1. From W
2 to Wave 3, economic immigrants who start out self-employed
remain working are most likely to stay self-employed (72% [2982/4
of economic immigrants self-employed at Wave 2). Of economic imm
grants not working in Wave 2 that later work in Wave 3, those that e
into self-employment represent a relatively large proportion
[533/3728] of those not working). Among those who are employed,
ever, it is slightly more common to move from an ethnic workplace
self-employment (8% [479/5644] of those in an ethnic workplace) th
from a non-ethnic workplace (5% [1970/39671]). Among family i
grants from Wave 2 to Wave 3, it is only slightly more common to t
sition into self-employment from an ethnic workplace (5% [324/6459
those in an ethnic workplace at Wave 2) than from a non-ethnic wo
place (2% [324/13135]) or from not working (3% [216/7405
Although self-employment is fairly uncommon, especially among fam
immigrants, ethnic workplaces are generally more likely to serve as ju
ing-off points for self-employment than non-ethnic workplaces.
MULTIVARIATE RESULTS
Workplace Type
Table 3 shows logistic regressions predicting, among those working

being employed in a non-ethnic workplace, relative to an ethnic work
place.23 The coefficients are odds ratios; those <1 signify the percentage
decrease in the odds of being in the outcome category, relative to the
omitted category, and coefficients >1 indicate the percentage increase in
the odds.
23Owing to software limitations for logistic regressions with this replicate weight structure
we report the pseudo-/?2 from the unweighted models. The /"-statistics presented are from
the weighted models.
TABLE 3
Logistic Regressions Predicting Non-ethnic Workplace, for Employed Working Age (18-65)
Principal Applicants, Waves 1-3
Wave 1 Wave 2 Wave 3
Age in W1 0.99*
0.01 W2 0.97*** 0.01 W3 1.00 0.01
Female 0.98 0.13 0.89 0.11 1.33* 0.18
Married in W1 W2
0.870.15 0.920.14
W30.720.13
Ethnicity (Omitted = White)
Chinese 0.470.23 0.890.48 0.33*
0.17
South Asian 0.95 0.46 1.20 0.66 0.35* 0.18
Black 3.35 2.20 2.36 1.54 0.50 0.30
Filipino and Southeast Asian 1.01 0.56 1.38 0.82 0.45 0.26
Latin American 3.42 2.65 2.07 1.38 1.04 0.81
Korean 0.20* 0.13 0.36 0.23 0.09*** 0.05
Arab 2.26 1.51 3.65 2.45 0.42 0.26
Other ethnicity 1.82 1.14 1.56 0.96 0.44 0.26
Immigrant category (Omitted = Economic)
Refugees and others 1.26 0.40 1.45 0.39 1.05 0.28
Family immigrants 0.56*** 0.09 0.79 0.13 0.58*** 0.09

Education (Omitted = <High School)
High School 1.64* 0.39 1.25 0.31 1.15 0.27
College Level 2.09** 0.52 2.00** 0.49 1.42 0.32
Bachelor's Degree 2.64*** 0.64 2.33*** 0.57 1.60* 0.38
Graduate Degree 4.49*** 1.21 3.14*** 0.83 1.62 0.42
Job class (Omitted = Low job class)
Medium job class in W1 1.46" 0.20 W2 0.74* 0.10 W3 1.04 0.16
High job class in W1 0.98 0.16 W2 0.71* 0.11 W3 0.89 0.15
Speaks official language in 0.23
1.78*** W1W1 1.96*** 0.26 W2 1.58*** 0.22
CMA ethnic concentration 0.88 0.63 1.28 1.02 0.24 0.18
Contacts in Canada before arrival 0.94 0.20 0.73 0.13 1.10 0.21
Hours worked in W1 1.00 0.00 W2 0.99*** 0.00 W3 0.99 0.00
New friends (Omitted = Mainly ethnic)
No new friends in W1 0.78 0.13 W2 0.92 0.24
Mainly non-ethnic new friends in W1 1.13 0.15 W2 1.18 0.15
Workplace (Omitted = Ethnic workplace)
Not working in W1 1.72** 0.34 W2 1.11 0.27
Non-ethnic workplace in W1 3.54*** 0.65 W2 3.26*** 0.61
Employed with ethnic supervisor in W1 0.44*** 0.08 W2 0.32*** 0.06
Self-employed in W1 0.27*** 0.10 W2 0.10*** 0.03
Constant 2.034 1.41 4.119 3.11 6.413* 4.63
Unweighted Pseudo-i? 0.139 0.228 0.273
F-test 9.987 16.20 18.03
Weighted N 52,347 64,944 68,260
Source: Longitudinal Study of Immigrants to Canada; */k0.05; **/k0.01; ***/k0.001.
Family immigrants who work are less likely to enter a non-ethnic

workplace initially after arrival and in Wave 3, providing additional sup
port for HI. There is some support for H2 in that refugees and others'
workplace sorting does not differ significantly from that of economic
immigrants, revealing that they also move into the non-ethnic workplace
faster than family immigrants. In Wave 2, family immigrants do not
differ from other immigrant groups, suggesting that some dabble in the
non-ethnic workplace before reverting back to the ethnic economy.24
Having worked in a non-ethnic workplace in a previous wave greatly
increases the odds of working in a non-ethnic workplace in subsequent
waves, suggesting path dependence. Not working in Wave 1 also leads to
higher odds of work in a non-ethnic workplace in Wave 2, relative to
working in an ethnic workplace in Wave 1. Both being self-employed and
being employed with an ethnic supervisor in one wave reduce the odds of
working in a non-ethnic workplace in the subsequent wave. Immigrant
entrepreneurs are particularly unlikely to switch into the non-ethnic work
place over time, and the decrease in their odds from Wave 2 to Wave 3
suggests greater stability in their workplace types.
Neither the ethnic concentration of the CMA nor having contacts
in Canada before arrival impacts the odds of working in a non-ethnic
workplace. Higher education and proficiency in an official language tend
to increase the odds of immigrants being in a non-ethnic workplace,
rather than an ethnic workplace, through all waves. In Wave 1, working
in a medium-status job (mostly technical, sales, and support jobs)
increases the odds of working in a non-ethnic workplace, relative to low
status jobs. However, by Wave 2, both high- and medium-status jobs
decrease the odds of being in a non-ethnic workplace, suggesting that
people in low-status jobs (e.g., laboring and manufacturing) are more
likely to move into the non-ethnic workplace in Wave 2 only. This lends
further support to the notion that lower-skilled immigrants may try out
the non-ethnic workplace, but later revert back to the ethnic economy.
Interestingly, by Wave 3, women become more likely than men to work
in a non-ethnic workplace. Finally, Korean immigrants are the only eth
nic group more likely to initially enter an ethnic workplace than White
immigrants. While there are no Wave 2 ethnicity effects, by Wave 3 Chi
nese, South Asian, and Korean immigrants are more likely to settle into
the ethnic economy. In fact, the coefficient predicting non-ethnic work
place for Korean immigrants drops even lower in Wave 3 than its already
low Wave 1 level.
24To assess the differential change in the gap across the categories, we initially included an
interaction between immigrant category and the lagged non-ethnic workplace variable, but
found no significant results.
Ethnicity of New Friendships
Many immigrants do not make any new friends. Overall, 13 percen

our immigrants do not make new friends in Wave 1, but this patte
associated with immigrant category. Twenty-five percent of family
grants and 7 percent of economic immigrants do not make new fr
within 6 months of arrival. Table 4 shows logistic regressions predic
for those who did make new friends, the odds of making mainly n
ethnic new friends, compared to mainly ethnic new friends, in each
The ethnicity of new friends that immigrants make differs sign
cantly across immigration categories. Refugees and others are more
than economic immigrants to make mostly non-ethnic friends in W
and Wave 3. This suggests that refugees' integration outside their et
community is facilitated early on by their sponsorship by the govern
or private groups. The pattern for family immigrants is more surpr
In Wave 1, family immigrants have significantly higher odds of ma
non-ethnic friends than economic immigrants. This suggests that ini
after arrival, family immigrants are less ensconced within ethnic co
nities than subsequently, and that economic immigrants may have
culty developing non-ethnic ties shortly after arrival.
Those who made mainly non-ethnic friends in a previous wave
tinue to do so in the next wave. More strikingly, those who made no
friends in an earlier period are more likely to make new non-et
friends than those who invested previously in ethnic friendships. T
patterns of early investment in ethnic friendships continue and self
force going forward.
The workplace transitions that immigrants make also influence t
types of friendships they develop. Compared to those who stay in et
workplaces between waves, those who move from an ethnic workpl
(EW) to a non-ethnic workplace (NEW) have much higher odds of m
ing mostly non-ethnic new friends. Those who stay in non-ethnic
places and those who move from not working to working in a non-e
workplace have the highest odds, in both Waves 2 and 3, of investi
non-ethnic friendships. In fact, every type of transition increases the
of making non-ethnic friends compared to staying in an ethnic work
over time, lending support to H3. Immigrants who start and remai
ethnic workplaces are most embedded in the ethnic community, ev
more so than those who transition into an ethnic workplace after 4 y
O (N O oo ir\ (N i—' NfooNt-mcOO xfco

O -j (S tn (fi co | cn co ■—< o rn rT)
ooo o o o o o o ■ o o o o o o d o d d o
fCOCNMNOOfSfO
lA00xJ<m(N00(N>—>
dd^d—^d(N^'
SO ^ JO £2 Q C!
^ ^ ^ ^
VD O
oooooooo d o o o o o o o o o o o
fOCN'-'O'-'OOrrifrj
rONfOmCNN^N 00 00
00-^0^0 -h'o o o
(NN-HCO'Or)(NCN oo oo cn (N m t\ m ^ ' o
rHfoq^mcnooiA (N (O fS ^ CN CN <N CO —; O
TABLE 4 ddHOHodo ddod odd do
Waves 1-3
N O to O N h —< —t (N(NOh
c
*c3
a £
3 a t ii Z
 w __, n .a ^PS. -a .a .a
3 C w ^ -
QJ ■ t t tyj _ OD I c y v o
JB v2< O
nj ..
~So -a
s £
£
-a fc" § -g J* w r, s5 our: 2 - iT w
1^-§
-o <2 S S .2 2
■a -S111
3
u
nJ §-g .sf><£
^ — .2 a «,sp= "5
tt£siu C/3 s 'Hi, < O E Cd OS 8 ^ _3 ?,
O _Q "S Ix! X z:
U c
CO
LogistcReronsPedictgMakn ilyNon-ethicwFrends,foWrkingAe(18-65)PrincpalA ntswhoMadeNwFrinds, <f tx! 2 w s a AS S UIZ
ooooooooo
nf CO vo cn(ncoo^ôiaiti(n
vq <n (n C) o v/^vo'-.f^cnhîrnoqq(n cr) m
po co ^ rn r4 odôd^hodd ^6 oo
—i t\
fnfsfnrofoforororf) rocriftirorofrirriro
>j>sss
■o O 00 ^ (t\ n vonooo^nncnoo
vocncn 17; n m in qq<no^—;—;qq
o o oooho 000000 0 0 0*
* * * * * mo
00 t\ cn on —< c\n-^>a\of^(niaxro
GO -h on fn on vo^-'noo^voo^fs
fn rn ^
•hrhrhfs(s(nn(n(n (n(nfn(nfn(n(nfn
jsssss js^ss^
OOOOOOOOO
* * * * * m vo
(n -_ OO rh (fi m - cf! 00 vo
vci ^ r^. [\ _h co 00 »h r
o o o' o ^ ^ ~ -j o' o
5s>s^>s^>s
w
£
T
w -a
«
ii
i fe c
rt
is ~5b ^
2 ,et !
w o -S
c ^
•- M .= i
uos
g TT o 3 -3 • ch
&£.£ v rv<
jm vjo rt V
.s *
v —' <-i [EW in EW in ►-3 u fe;
\ 3 oo % c S "§
"0 _o c 3 3 c -
t T
sl>
q. 2 oj -a -o y u* js
s2
^ .2£ 6 -g t k. i i
•c c c c -s 8 ■§
u -s
•c-5<y c/3 i- o "s „ k b£>
■i*« § i
fet
«-. oh
;±- rt £ ss w o o u 1 u« w ff5 w
V3
z uj co h Z w Z zzz o ud
Working in a job with an ethnic supervisor in a previous wave has

no effect on the type of friends one develops in Waves 2 or 3. Being self
employed in a previous wave, however, significantly increases the odds of
forming new non-ethnic friendships in Waves 2 and 3; even though
entrepreneurship is considered part of the ethnic economy, those who
work for themselves develop more friendships outside their ethnic group,
which may relate to past findings of why entrepreneurs have better out
comes than those employed in the ethnic economy (Sanders and Nee,
1987). In all waves, neither the ethnic concentration of the CMA nor
having contacts in Canada before arrival affects the ethnicity of new
friendships. However, people who make new friends specifically through
their relatives or friends in Canada, and through contacts in their old
country, are more likely to make mainly ethnic new friends. The same
is true of those who make new friends through religious activities
(e.g., church, synagogue, mosque) through all waves. Taking English- or
French-as-a-second language classes has no effect, but taking other (non
ESL/FSL) classes leads to making mainly non-ethnic friends in all waves.
Making friends through work also leads to new non-ethnic friendships
across waves. Surprisingly, making new friends from an ethnic association
does not affect the ethnicity of those friendships.
Human capital - as measured through education, job class, and lan
guage proficiency - also influences friendship patterns. Compared to those
with less than a high-school education, immigrants with a college-level
education, bachelor's degree, or graduate degree are generally more likely
to develop non-ethnic friendships. Speaking an official language increases
the odds of making mainly non-ethnic friends in all waves. A higher job
class also elevates the likelihood of forming mainly non-ethnic friendships
in Wave 1; however in Waves 2 and 3, a greater likelihood of forming
new non-ethnic ties is associated with professional- and management-level
jobs only.
Other background characteristics also play a role in the ethnicity of
new friendships. Women are more likely to build primarily non-ethnic
friendships in Wave 1. People who are married are more likely to develop
ethnic friendships through the first 2 years after immigration. There are
also ethnic differences that are not explained by other factors. While
Chinese immigrants are initially less likely to develop non-ethnic friend
ships than Whites, they acquire statistically similar levels after the first 2
years. Filipinos and Southeast Asians are also more likely to form mainly
ethnic new ties, although only in Wave 1. Latin Americans and Blacks are
the more likely than White immigrants to make non-ethnic fri

although only during the first wave.
Income
The final set of models, addressing H4, H5, and H6, use weighted li
regression to predict the logged personal income of respondents
earned any income through a job or business (Table 5). We present
models for each wave: the first with our primary variables of interes
the second with interactions between immigrant category and non-e
workplace, and between immigrant category and self-employed. Lo
incomes from the previous wave are used as independent variab
Waves 2 and 3.
We do not find support for H4: working in a non-ethnic workpl
initially after arrival (Wave 1) does not show any significant differen
working in an ethnic workplace. By Wave 2, we do see in the trans
variables that those who were not working in Wave 1 and transitio
into either an ethnic or non-ethnic workplace have lower incomes
those who entered and stayed in an ethnic workplace between Wav
and 2. While entering the ethnic workplace right after arrival does
offer greater immediate income returns than jobs in the open econo
starting work in the ethnic economy provides a later income gain rel
to delaying work.
H5 predicts that non-ethnic workplaces will offer higher incom
than ethnic ones over time. By Wave 2, in the model with interact
terms, the non-ethnic workplace is associated with higher incomes
those who worked in a non-ethnic workplace through Waves 1 and
Wave 3, the positive impact of working continuously in a non-e
workplace through Waves 2 and 3 appears in both models, and transi
ing from ethnic workplace to non-ethnic workplace in Wave 3 is si
cant in the interaction model. These results at first seem to support
However, closer inspection of the interactions between non-eth
workplace and immigrant category, in Model 2 of Waves 2 and 3, s
that for family immigrants, working in a non-ethnic workplace, r
than an ethnic workplace, actually reduces income. For refugees and
immigrants, working in a non-ethnic workplace has no incremental e
on income. These combined results lend support to H5, but not for
immigrant categories. Once the non-ethnic workplace/immigrant cat
interactions are included, working in a non-ethnic workplace does
0 CO ON 00 r-H O 1—1 CO T—I 1—1 CN 00 r^ 00 co •^r vo VT\ 0 ,—1
0 CN <N CN CN 0 0 0 0 0
q q >—• *—1 CN CN •—• O 0 0 CN O 0 q
d d 0 d 0 O O O O d d d d d d 0 d 0 0 d O d d d
OOiA^NCOl"^
q cn —<
odd oooooooo o o o o o o o o o -
I I III I I I I I
0 co xr CN 00 <N O co _1 _| *0 tn 00 00 Nf ^0 ir\ 0
0 O 0 CN CN CN CN CN O O
CN q q q 0 0 O 0 CN
q 0
q
>—< »—1
d O 0 O 0 O d d O d O 0 O 0 0 0 d O O 0 d d 0 d
* * * * * *
* * *
* * *
0 CN r-H 0 h co CN l/~\ 1r\ cn 00 0 m CO t\ C\ xr
0 CN CN 0 cn rs 0 0 0 <N 0 CO 0 0
q q q
•—< <—< •—1
0 O O 0 d O d 0 d O O O O 0 0 0 O 0 O 0 0 0 d d
CO (N co CN
>* $ $
0 ^ U~\ NNOCNCNNOO fS CN 00 00 CN CN ^
0 0 0 —< —' CN <N CS (N -H 0 OOOO O
0 0 0 00000000 do Oddd O 0.05 0.05 0.25 0.05
O X}- \0
o co o
d d o ooooooo oo o o o o ooooooo
I I I I I I I I I II I I
IA ^ IA IA O --H
o q o cn o o o
oooooooo o o o o 0000000
OCNOCNOOiTnCNON ts N fO 00 h h n n tn
r* o q q
o o ■ oooooooo oo dodo d o o d o
I I I I I I I I I I II I I
CN CN «-H CN i-i
£
TABLE 5 O vo vo ON 00 00 00 nj ITN ir\ VTN *r\ IT\ 0 00 0
O 0 CO CO CO CO CO NT CO CO O O O ITN 0 0
q
O d d d 0 0 O d O d d 0 0 O d O d O d 0 d d d
*
O CS CN "t O CN 00 00 t\ t\ 0 CN CN vr\ 0
O CN <0 0 0 •^r CN 0 CN O O 0 0 [\ 0 0 0
d d d d d d d d 0 O 0 d d d 0 O d 0 d d d d d
I I I I I
0 CN 00 00 00 0 00 ITN vo 0 00 0
0 CO CO CO CO CO 0 0 0 0 0
q q CO CO >—1 1r\
q
d d d d 0 0 d O O d d 0 0 d d d d d d 0 d d d
* * *
* * * *
0 CN CN 0 CN 00 00 t\ 00 00 v/N 0 ITN
0 CN T—> CO
q 0 •—> •^r (N 0 «-* CN 0 >—1 0 0 0 0 0 0
dd d d 0* d 0" O* d d 0 d d 0 d d 0 0 d d d d d
o
o
1
_c
£ I V
O hJ SJ
£
O
-£ 3 C <* o
3 c
e
& 8 X 3 -—rsi
O « £{ w » bX) ■ —
- s _ _£}
c ft
3 §> p •§ -c -if _c -c u
d bi &_2
| .2. ® ^ ^ o
•s -a -g JS 43 £ § "2 "2 •="€ I 3 -SP"3 "I 2 "8 3 -SP-i _< w

s> 13 o E«u, siusUjjSils
£ 3 o .e-n
Q M
<£ S
oooooooooo
00
\5(7\n\001ama\0^'-'>00
nôqfnhnir;qqfnovo
dddddodddcndôn
ii ^ vo
vo
o o o o o
* ON 00
it\ vo (n v0 O t\ o o o
^ o o O co r~l ***} ^
ddddd cn o N c\
i i fo \0
<n cr> focofîcrifnc^fof^fo
& >s >s ^
OOOOOOOOOO
<n t\
niaftiooi-'monfoaiavom
h-hqotfihnqq-;(fiho
o o ddcjcjoooo'oonokv:^
i i i i i i i ^ v0
00 on <tn m ^
o o O o (N
o o* o o o o
* —11
— ctn vo u-n 00
0 o o m fnt^too
© o o o ooooo c\ o o v*
1 i i i i ^ \o
(n m (s (n jn {n fs fs jn
co^vo ON
<—> •—1 co ^ v^h
oo'ooo
o
CO 00 vo
s P p —;
ooo OOOOOOOVOri
i i
CO o 00 m
$&
■s.e£ S 8
w
z
z 5? 00 «
"0 zw
2 c . *x*.x
11
' "t
Tt
o
1i £^
p. i^p, g> e1
■14
_ .2^— .2f
t3 >
tu
cl, ^ ifl T
tt | £
gl: 8 Oh %
% H
— g £ _
5£ >, C
c^.2 >
- -o -a
-5 -c ^
1:1^
-, rt s
"t 0 •a v .a u a w o o
vsl-sl g-cl's
oc
£z s !±! rt z z z
z w 00 h
(u «-<
(2,2 2
translate into higher incomes for family immigrants. This finding suggests
that the open economy creates barriers to long-term success for family
immigrants. This is perhaps why family immigrants did not differ from
economic immigrants in their likelihood of being in a non-ethnic work
place in Wave 2, while they were more likely to be in an ethnic workplace
in Waves 1 and 3 (Table 3). Family immigrants likely sampled the open
economy and, finding it not a good fit, move back to the ethnic econ
omy. In fact, additional post hoc analyses (not shown) reveal that when
controlling for human capital, family immigrants in an ethnic workplace
have no difference in income from economic immigrants working in a
non-ethnic workplace in Waves 1 and 2. They do start to suffer an
income penalty by Wave 3, however, suggesting that the relative advanta
ges of working in a non-ethnic workplace take time to materialize for
economic immigrants.
Is self-employment, then, a route to greater economic success for
family immigrants if the open economy is not? The first model in Wave
1 shows that, on its own, being self-employed has no effect on income.
By Waves 2 and 3, being self-employed in general leads to a lower
income, relative to being an employee in an ethnic workplace. When we
add the interactions between self-employed and immigrant category in the
second models, the negative base effect for self-employment remains in
both Waves 2 and 3. The net effect of self-employment on income is thus
negative. The interaction terms reveal no differential effect for family
immigrants in self-employment; thus, it is not an alternate route to eco
nomic success for them. In Wave 3, refugees who are self-employed do
have an income advantage relative to refugees in an ethnic workplace;
however, this finding should be treated with caution as the number of
self-employed refugees in Wave 3 is fairly small. As an economic strategy,
self-employment does not provide income benefits to family or economic
immigrants within the first 4 years after immigration.
Confirming H6, forming mainly non-ethnic new friends in previous
waves leads to higher-income levels in subsequent waves. With a coeffi
cient around 0.09 in both waves, those who form new friendships mainly
with non-ethnics earn around 9 percent higher incomes than their coun
terparts who develop mainly ethnic friendships. This echoes past scholars'
contention that homogeneous networks that do not provide access to
unique information create a disadvantage {e.g., Uzzi, 1999).
Several relevant control variables show statistical significance; nota
bly, while education has a positive effect on income in Waves 2 and 3,
there is no effect in Wave 1. In Wave 2, those with a high-school degree

earn more than those without one. Those with graduate-level degrees earn
higher incomes in Waves 2 and 3 as well. Oddly, those with bachelor's
degrees do not earn significantly more than those without high-school
degrees. It is possible that while those with high-school degrees are bette
at finding low-paying jobs than those without high-school degrees, imm
grants with bachelor's degrees from their home countries try to find
higher-paying jobs in the open economy but their credentials are not re
ognized, while those with graduate degrees have more success. These find
ings support the idea that immigrants face barriers to getting their skill
and qualifications recognized by Canadian employers, but it shows that
those with significantly more schooling eventually receive increasin
returns to graduate education.
Proficiency in an official language increases income in Wave 2 only.
Respondents' incomes from previous waves predict income in the curren
wave. High job class is also associated with higher incomes across al
waves, although the coefficient steadily decreases from wave to wave. As
in the other models, neither the ethnic concentration of the CMA nor
having contacts in Canada before arrival has any effect. Being employe
with an ethnic supervisor also has no effect during the first two waves,
when this variable is included. Female immigrants have lower incomes
than males throughout all waves, but being married is associated with
higher income in Wave 3.
DISCUSSION AND CONCLUSIONS
Does the ethnic economy provide an alternate route to mobility fo

grants facing barriers that hurt them in the primary labor mark
analysis of the initial period after immigration shows that this c
question in the sociology of immigration must be qualified by cons
the pathways followed and barriers faced by different categories o
grants. Immigrant category has an important influence on sorting
grants into different labor market trajectories, and on financial return
human capital investments. Family immigrants are less likely tha
nomic immigrants to work in the open economy initially after a
Although they do transition into non-ethnic workplaces over time
remain more concentrated in ethnic workplaces than other types o
grants even 4 years after they arrive. Economic immigrants tran
more frequently from the ethnic economy into the non-ethnic ec
Strikingly, when family immigrants do venture into the non-ethnic

economy, they fail to achieve the same income gains as other immigrants,
giving them incentives to continue working outside of it. Yet while they
initially do just as well in the ethnic economy as economic immigrants do
outside of it, by 4 years after arrival, economic immigrants have pulled
ahead. The ethnic economy literature developed out of a concern that a
dual labor market was developing — offering different pathways to integra
tion inside and outside of the ethnic community. Our findings support
the existence of a dual labor market - one divided not just by entrepre
neurs and employees, but by family immigrants and economic immi
grants. Immigrant entrepreneurship in particular has long been discussed
as a potential path for immigrants whose skills and credentials may not
be recognized in the mainstream economy. Consistent with past research
suggesting that entrepreneurship may be a beneficial alternative to refu
gees, given the barriers they face in mainstream employment {e.g., Knight,
1983; Wauters and Lambrecht, 2006), our findings tentatively suggest
that refugees are able to benefit from self-employment in the longer term,
relative to being employed in an ethnic workplace. These findings should
be replicated in a study with a larger sample of refugee entrepreneurs,
however. Yet for family immigrants, who also face barriers in the open
economy, entrepreneurship is not an avenue to economic advancement, at
least not within the first 4 years after arrival in Canada. Nor do economic
immigrants, including those who enter Canada as business-class immi
grants because of their ability to create new jobs, reap financial rewards
from self-employment. Although much of the literature portrays entrepre
neurship as potentially offering higher returns on immigrants' skills, entre
preneurial endeavors entail considerable risks that do not always pay off.
Immigrant entrepreneurship also includes day labor and other precarious
work offering little pay, benefits, or employment security, as well as little
opportunity for collective bargaining (Valenzuela, 2001; Cranford et al.,
2005). In short, self-employment is not necessarily advantageous for new
immigrants.
Social networks play an important role in how and where immi
grants integrate. With the concept of ENPs, we illustrate how immi
grants invest in different types of networks. There is considerable path
dependence here, with investments in ethnic ties leading to the devel
opment of even more ethnic ties in the future. Yet while many assume
that investing primarily in non-ethnic ties will lead to future employ
ment in a non-ethnic workplace, we found no evidence to support this
assumption. Investment in a non-ENP produces economic benefits t

are separate from movement into the open economy. Non-ENPs —
a diversified ENP and a different non-diversified ENP - likely prov
sources of new information, opportunities, and experiences that pa
no matter where an immigrant works.
Immigrants who arrive under economic, family, and refugee im
gration categories differ in terms of their resources, their forms of cap
and their perceived fit with the demands of a particular economy. I
gration policies largely sort them into distinct pathways of incorpor
While economic immigrants diversify their ENPs, move into an
labor market, and typically stay there, family immigrants are more
to become embedded within ethnic communities and workplaces an
face economic barriers to succeeding elsewhere. If refugees experie
greater returns from entrepreneurship than other immigrants, it l
reflects the institutional and organizational support they receive f
private and government sponsors.
Our focus on immigrant categories is particularly relevant to po
discussions, because immigration policies in Canada have recently shi
toward greater acceptance of economic immigrants. Indeed, many o
developed countries such as New Zealand and Australia also use a po
system to evaluate potential immigrants, and there is an increasing
for other nations, such as the United States, to adopt a similar point
tem to raise the "quality" of immigrants (Borjas, 1999). Yet re
research on immigrants to Canada presents a paradox: Over the pas
decades while the share of economic immigrants - selected for
human capital and ability to meet current labor market demands -
been increasing, the economic performance of recent immigrant coh
has been deteriorating (e.g., Bloom, Grenier, and Gunderson,
Green and Green, 1995; Wanner, 2003; Picot, Hou, and Coulombe,
2007). There are several possible explanations — such as declining eco
nomic conditions at the time of immigrants' landing; differing human
capital of the immigrant cohorts; or decreasing economic returns to for
eign work experience (Chui and Zietsma, 2003; Picot and Sweetman,
2005). Yet our study suggests that the declining economic performance is
not attributable to the shift from family immigrants to economic immi
grants or economic immigrants' lack of established social capital in Can
ada (see Picot, Hou, and Coulombe, 2007). Four years after arrival,
economic immigrants do better economically than family immigrants,
even family immigrants employed in ethnic workplaces. Furthermore,
established social capital in Canada before arrival has no impact on

economic outcomes during this period. We suspect that larger structural
economic factors do a better job of accounting for declines over time than
the changing policy emphasis on economic immigrants.
However, we caution against overestimating the short-term cost of
immigration and missing the longer-term consequences of immigrants'
economic integration over time (Green and Worswick, 2004), as well as
their families' outcomes over generations (Aydemir, Chen, and Corak,
2009). Although our findings show lower economic success for family
immigrants, we retain a healthy inquisitiveness about the impacts of fam
ily reunification that are not observable at the individual level. Our data
are unable to show the indirect economic benefits attributable to having a
supportive family unit. For instance, if a grandparent immigrates to help
care for children, this may improve the economic opportunities of the
parent, a benefit that would not be captured effectively with individual
level analysis. Future research on differences between immigrant categories
would benefit from viewing families and/or communities as the unit of
analysis.
What is needed on a policy level is a system that not only considers
economic immigrants' human capital, but one that helps immigrants of
all categories to succeed economically. For family immigrants, this can
include helping them diversify their network portfolios, for instance
through classes that are unrelated to language acquisition. Classes and
training can also help them acquire skills that are suited to the current
economic climate, for which economic immigrants are selected. Policies
with such a focus can help maximize the benefits of a pro-immigration
regime.
Our study shows that immigrant categories, or the different types of
immigrants selected by contemporary immigration policies, need to be
factored into considerations of the economic benefits of the ethnic econ
omy. It is not simply a matter of whether "immigrants" can achieve
mobility through their ethnic communities to avoid the structural barriers
in the primary labor market, or whether the open economy provides
greater advantages in the long term. Rather, certain immigrants face pen
alties in parts of the economy that other immigrants do not. And while
family immigrants may initially receive the same returns to their human
capital in the ethnic economy as economic immigrants do outside of it,
these benefits do not last over time. Immigration policies, by sorting
immigrants into different visa categories based on their resources and
economic suitability, help to create segmented trajectories of incor

tion, with different categories of immigrants doing better in or out of t
ethnic economy.
REFERENCES
Anderson, J. M.
1990 "Work and Health: Considerations for Immigrant Women." Anthropology of Work
Review ll(3):3-7.
Aydemir, A., W.-H. Chen, and M. Corak
2009 "Intergenerational Earnings Mobility among the Children of Canadian Immi
grants." Review of Economics and Statistics 91(2):377-397.
Bailey, T., and R. Waldinger
1991 "Primary, Secondary, and Enclave Labor-Markets: A Training Systems-Approach."
American Sociological Review 56(4):432î45.
Bartlett, M. S.
1947 "The Use of Transformations." Biometrics 3(1):39—52.
Batalova, J.
2009 "U.S. In Focus: Spotlight on Legal Immigration to the United States." Migration
Policy Institute. <http://www.migrationinformation.org/USfocus/dis
play.cfm?id=730>.
Bean, F. D., M. Leach, and B. L. Lowell
2004 "Immigrant Job Quality and Mobility in the United States." Work and Occupations
31 (4) :499—5 8.
Bloom, D. E., G. Grenier, and M. Gunderson
1995 "The Changing Labour Market Position of Canadian Immigrants." Canadian Jour
nal of Economics 28(4b):987-1005.
Bonacich, E.
1972 "A Theory of Ethnic Antagonism: The Split Labor Market." American Sociological
Review 37:547-559.
1973 "A Theory of Middleman Minorities." American Sociological Review 38(5):583-594.

Borjas, G. J.
1989 "Immigrant and Emigrant Earnings: A Longitudinal Study." Economic Inquiry
27(l):21-37.
1999 Heaven's Door: Immigration Policy and the American Economy. Princeton, NJ:
Princeton University Press.
Catanzarite, L., and M. B. Aguilera
2002 "Working with Co-Ethnics: Earnings Penalties for Latino Immigrants at Latino
Jobsites." Social Problems 49(1):101-127'.
Chiswick, B. R.
1994 "The Performance of Immigrants in the United States
Aspects of International Migration. Ed. H. Giersch. New Y
Pp. 95-114.
, Y. L. Lee1, and P. W. Miller
2003 "Patterns of Immigrant Occupational Attainment in a Longitudinal Survey." Inter
national Migration 41 (4):47—69.
, , and
2004 "Immigrants' Language Skills: The Australian Experience in a Lon
International Migration Review 38(2):611-654.
, , and
2005a "Immigrant Earnings: A Longitudinal Analysis." Review of In
51 (4):485—503.
, , and
2005b "A Longitudinal Analysis of Immigrant Occupational Mo
Immigrant Assimilation Hypothesis." International Migratio
353.
, , and
2006 "Immigrants' Language Skills and Visa Category." Internat
40(2):4l9-450.
Chui, T., and D. Zietsma
2003 "Earnings of Immigrants in the 1990s." Canadian Social Tr
alogue 11-008(70):24.
Citizenship and Immigration Canada
2007 "Immigration Overview: Permanent and Temporary
www.cic.gc.ca/english/resources/statistics/facts2007/glossary.asp
Constant, A., and D. S. Massey
2003 "Self-Selection, Earnings, and out-Migration: A Longitudin
to Germany." Journal of Population Economics 16(4):631—653.
Cranford, C. J. et al.
2005 Self-Employed Workers Organize: Law, Policy, and Unions. M
ON: McGill-Queens University Press.
DeVoretz, D. J.
1995 Diminishing Returns: The Economics of Canada's Recent Imm
ON: C.D. Howe Institute.
Dustmann, C„ and F. Fabbri

2003 "Language Proficiency and Labour Market Performance of Immigrants in the
UK*." The Economic Journal 113(489):695—717.
Feld, S. L.
1982 "Social Structural Determinants of Similarity among Associates." American Sociolog
ical Review 47(6):797-801.
Fernandez, R. M., and I. Fernandez-Mateo
2006 "Networks, Race, and Hiring." American Sociological Review 71(1):42—71.
, and M. L. Sosa
2005 "Gendering the Job: Networks and Recruitment at a Call Center." American Jour
nal of Sociology 111 (3):859—904.
Fleras, A., and J. L. Elliott

2007 Unequal Relations: An Introduction to Race, Ethnic, and Aboriginal Dynamics i
ada, 5th ed. Toronto, ON: Prentice Hall.
Fong, E., E. Chan, and X. Cao
2009 "Moving out and Staying in the Ethnic Economy." International Migration
10.1111 /j. 1468-2435.2009.00534.X.
, and E. Ooka
2002 "The Social Consequences of Participating in The Ethnic Economy." Intern
Migration Review 36(1): 125—146.
Gold, S. J.
1992 "The Employment Potential of Refugee Entrepreneurship: Soviet Jews and Viet
namese in California." Review of Policy Research 11 (2): 176—186.
Granovetter, M. S.
1973 "Strength of Weak Ties." American Journal of Sociology 78(6):1360-1380.
Green, D. A.
1999 "Immigrant Occupational Attainment: Assimilation and Mobility over Time."
Journal of Labor Economics 17( 1) :49—79.
Green, A. G., and D. A. Green
1995 "Canadian Immigration Policy: The Effectiveness of the Point System and Other
Instruments." Canadian Journal of Economics 28(4b):1006—1041.
Green, D. A., and C. W. Worswick
2004 "Immigrant Earnings Profiles in the Presence of Human Capital Investment: Mea
suring Cohort and Macro Effects." Institute for Fiscal Studies, http://econpapers.re
pec.org/RePEc:ifs:ifsewp:04/13.
Hagan, J. M.
1998 "Social Networks, Gender, and Immigrant Incorporation: Resources and Con
straints." American Sociological Review 63(1):55—67.
Hansson, R. O. et al.
1997 "Successful Aging at Work: Annual Review, 1992-1996: The Older Worker and
Transitions to Retirement." Journal of Vocational Behavior 51(2):202-233.
Hou, F.
2009 "Immigrants Working with Co-Ethnics: Who Are They and How Do They Fare?"
International Migration 47(2):69-100.
Hum, T.
2001 "The Promises and Dilemmas of Immigrant Ethnic Economies." In Asian and
Latino Immigrants in a Restructuring Economy: The Metamorphosis of Southern Cali
fornia. Ed. M. C. Lopez-Garza and D. R. Diaz. Palo Alto, CA: Stanford University
Press. Pp. 77-101.
Jasso, G., and M.R. Rosenzweig
1995 "Do Immigrants Screened for Skills Do Better Than Family Reunification Immi
grants?" International Migration Review 29( 1) :85—111.
Kler, P.
2006 "Graduate Overeducation and Its Effects among Recently Arrived Immigrants to
Australia: A Longitudinal Survey." International Migration 44(5):93—128.
Knight, R.
1983 "Entrepreneurship in Canada." Journal of Small Busin
1(1):9—15.
Krahn, H., T. M. Derwing, and B. Abu-Laban
2003 "The Retention of Newcomers in Second and Third Tier Cities in Canada." Prairie
Centre of Excellence for Research on Immigration and Integration, Working Paper
No. WP01-03.
Lee, T. W. et al.
2004 "The Effects of Job Embeddedness on Organizational Citizenship, Job Performance,
Volitional Absences, and Voluntary Turnover." Academy of Management Journal
47(5):711-722.
Li, P. S.
2000 "Economic Returns of Immigrants' Self-Employment." Canadian Journal of Sociol
ogy 25(1): 1-34.
2001 "Immigrants' Propensity to Self-Employment: Evidence from Canada." Interna

tional Migration Review 35(4): 1106-1128.
, and C. H. Dong
2007 "Earnings of Chinese Immigrants in the Enclave and Mainstream Economy." Cana
dian Review of Sociology and Anthropology 44(l):65-99.
Light, I. H.
1972 Ethnic Enterprise in America: Business and Welfare among Chinese, Japanese, and
Blacks. Berkeley, CA: University of California Press.
1984 "Immigrant and Ethnic Enterprise in North America." Ethnic and Racial Studies
7(2): 195-216.
, and S. J. Gold
2000 Ethnic Economies. San Diego, CA: Academic Press.
Light, I. et al.
1994 "Beyond the Ethnic Enclave Economy." Social Problems 4l(l):65-80.
Logan, J. R., and R. Alba
1999 "Minority Niches and Immigrant Enclaves in New York and Los Angeles: Trends
and Impacts." In Immigration and Opportunity: Race, Ethnicity, and Employment in
the United States. Ed. F. Bean and S. Bell-Rose. New York, NY: Russell Sage Foun
dation. Pp. 172-193.
, R. D. Alba, and T. L. McNulty
1994 "Ethnic Economies in Metropolitan Regions: Miami and Beyond." Social Forces
72(3):691-724.
, R. Alba, and B. J. Stults
2003 "Enclaves and Entrepreneurs: Assessing the Payoff for Immigrants and Minorities."
International Migration Review 37(2):344-388.
Marger, M. N.
2001 "The Use of Social and Human Capital among Canadian Business Immigrants."
Journal of Ethnic and Migration Studies 27(3):439-453.
McGuire, G. M.
2000 "Gender, Race, Ethnicity, and Networks: The Factors Affecting the Status o
Employees' Network Members." Work and Occupations 27(4):500-523.
Mclsaac, E.
2003 "Nation Building Through Cities: A New Deal for Immigrant Settlement in
Canada." Ottawa, ON: Caledon Institute of Social Policy.
McPherson, J. M., and L. Smith-Lovin
1987 "Homophily in Voluntary Organizations: Status Distance and the Composition o
Face-to-Face Groups." American Sociological Review 52(3):370-379.
McPherson, M., L. Smith-Lovin, and J. M. Cook
2001 "Birds of a Feather: Homophily in Social Networks." Annual Review of Sociology
27:415-444.
Menard, S. W.
2008 Handbook of Longitudinal Research: Design, Measurement and Analysis. Boston, MA:
Elsevier Academic Press.
Menjivar, C.
2000 Fragmented Ties: Salvadoran Immigrant Networks in America. Berkeley, CA: Univer
sity of California Press.
Mitchell, T. R. et al.
2001 "Why People Stay: Using Job Embeddedness to Predict Voluntary Turnover."
Academy of Management Journal 44(6): 1102—1121.
Mollica, K. A., B. Gray, and L. K. Trevino
2003 "Racial Homophily and Its Persistence in Newcomers' Social Networks." Organiza
tion Science 14 (2): 123-136.
Moss, L. B., and H. Goldstein
1979 The Recall Method in Social Surveys. London: University of London, Institute of
Education.
Nee, V., and J. Sanders

2001 "Understanding the Diversity of Immigrant Incorporation: A Forms-of-Capital
Model." Ethnic and Racial Studies 24(3):386-411.
, J. M. Sanders and S. Sernau
1994 "Job Transitions in an Immigrant Metropolis - Ethnic Boundaries and the Mixed
Economy." American Sociological Review 59(6):849-872.
Ng, T. W. H., and D. C. Feldman
2007 "Organizational Embeddedness and Occupational Embeddedness across Career
Stages." Journal of Vocational Behavior 70(2):336—351.
Petersen, T., I. Saporta, and M.-D. L. Seidel
2000 "Offering a Job: Meritocracy and Social Networks." American Journal of Sociology
106(3):763-816.
Picot, G., and A. Sweetman.
2005 "The Deteriorating Economic Welfare of Immigrants and Possible Causes: Update
2005." Analytical Studies Research Paper Series. Catalogue No.
11F0019MIE2005262. Ottawa, ON: Statistics Canada.
Picot, G., F. Hou, and S. Coulombe
2007 "Chronic Low Income and Low-Income Dynamics among Recent Immigrants."
Statistics Canada, Analytical Studies Branch.
Piore, M. J.
1979 Birds of Passage: Migrant Labor and Industrial Societies. New York, NY: Cambridge
University Press.
Portes, A., and R. L. Bach
1985 Latin Journey : Cuban and Mexican Immigrants in the United States. Berkeley, CA:
University of California Press.
, and L. Jensen
1987 "What is an Ethnic Enclave? The Case for Conceptual Clarity." American Sociologi
cal Review 52(6):768-771.
, and
1989 "The Enclave and the Entrants: Patterns of Ethnic Enterprise in Miami before and
after Mariel." American Sociological Review 54(92):9—949.
, and J. Sensenbrenner
1993 "Embeddedness and Immigration: Notes on the Social Determinants of Economic
Action." The American Journal of Sociology 98(6):1320-1350.
, and A. Stepick
1993 City on the Edge: The Transformation of Miami. Berkeley, CA: University of Califor
nia Press.
Reitz, J. G.
2001 "Immigrant Success in the Knowledge Economy: Institutional Change and the
Immigrant Experience in Canada, 1970-1995." Journal of Social Issues 57(3):579
613.
Roth, W. D.
2008 "There is no Discrimination Here': Understanding Latinos' Perceptions of Color
Discrimination through Sending-Receiving Society Comparison." In Racism in the
21st Century: An Empirical Analysis of Skin Color. Ed. R. E. Hall. New York:
Springer Press. Pp. 205-234.
2009 "Transnational Racializations: The Extension of Racial Boundaries from Receiving

to Sending Societies." In How the United States Racializes Latinos: White Hegemony
and its Consequences. Eds. J. A. Cobas, J. Duany, and J. R. Feagin. Boulder, CO:
Paradigm Publishers. Pp. 228-244.
2012 "Race Migrations: Latinos and the Cultural Transformation of Race." Palo Alto,
CA: Stanford University Press.
Sanders, J. M.
2002 "Ethnic Boundaries and Identity in Plural Societies." Annual Review of Sociology
28(l):327-357.
, and V. Nee
1987 "Limits of Ethnic Solidarity in the Enclave Economy." American Sociological Review
52(74):5—773.
, and
1992 "Problems in Resolving the Enclave Economy Debate." American Sociological
Review 57(3):4l5—418.
Schellenberg, G., and H. Maheux

2007 Immigrants' Perspectives on Their First Four Years in Canada: Highlights from
Waves of the Longitudinal Survey of Immigrants to Canada, Canadian Social T
Catalogue no. 11-008. Ottawa, ON: Statistics Canada.
Seidel, M.-D. L., J. T. Polzer, and K. J. Stewart
2000 "Friends in High Places: The Effects of Social Networks on Discrimination
ary Negotiations." Administrative Science Quarterly 45(1): 1—24.
Seidel, M., Z. Barsness, E. Lo and J. Wong
2007 "The Role of the Ethnic Network Portfolio in Discrimination for Canadian Immi
grants and Minorities of Work." Administrative Science Association of Canada.
Ottawa, ON.
Statistics Canada
2006a "Number and share of the foreign-born population in Canada, 1901 to 2006."
http://wwwl2.statcan.gc.ca/census-recensement/2006/as-sa/97-557/figures/cl-eng.
cfm.
2006b "Region of birth of recent immigrants to Canada, 1971 to 2006." http://

wwwl2.statcan.gc.ca/census-recensement/2006/as-sa/97-557/figures/c2-eng.cfm.
2007 Longitudinal Survey of Immigrants to Canada: Wave 3, Microdata User Guide.

Ottawa, ON: Statistics Canada, Special Surveys Division.
Uzzi, B.
1999 "Embeddedness in the Making of Financial Capital: How Social Relations and
Networks Benefit Firms Seeking Financing." American Sociological Review 64(4):
481-505.
Valenzuela, A.
2001 "Day Labourers as Entrepreneurs?" Journal of Ethnic & Migration Studies
27(2):335-352.
Waite, L. J.
1995 "Does Marriage Matter?" Demography 32(4):483-507.
Waldinger, R.
1994 "The Making of an Immigrant Niche." International Migration Review 28(l):3-30.
Wanner, R. A.
2003 "Entry Class and the Earnings Attainment of Immigrants to Canada, 1980-1995."
Canadian Public Policy/Analyse de Politiques 29(1):53—71.
Wauters, B., and J. Lambrecht
2006 "Refugee Entrepreneurship in Belgium: Potential and Practice." International Entre
preneurship and Management Journal 2(4):509-525.
Wilson, K. L„ and A. Portes
1980 "Immigrant Enclaves: An Analysis of the Labor Market Experiences of Cubans in
Miami." American Journal of Sociology 86(2):295—319.
Winship, C., and L. Radbill
1994 "Sampling Weights and Regression Analysis." Sociological Methods & Research
23(2):230-257.
Xie, Y„ and M. Gough

2009 "Ethnic Enclaves and the Earnings of Immigrants." P
Research Report 09-685, Institute for Social Research, Universit
Xue, L.
2008 "Social Capital and Labour Market Outcomes of Recent Immigrants to Canada:
Employment Entry, Wages and Duration of Access to the First Job in Intended
Occupation." Ph.D. dissertation, University of Ottawa.
Zhou, M.
1992 Chinatown: The Socioeconomic Potential of an Urban Enclave. Philadelphia, PA:
Temple University Press.
2004 "Revisiting Ethnic Entrepreneurship: Convergencies, Controversies, and Conceptual

Advancements." International Migration Review 38(3):1040-1074.
, and C. L. Bankston
1998 Growing up American: How Vietnamese Children Adapt to Life in the United States.
New York, NY: Russell Sage Foundation.

In and Out of The Ethnic Economy A Longitudinal Analysis of Ethnic Networks and Pathways To Economic Success Across Immigrant Categories

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

In and Out of The Ethnic Economy A Longitudinal Analysis of Ethnic Networks and Pathways To Economic Success Across Immigrant Categories

Uploaded by

Copyright:

Available Formats

In and Out of the Ethnic Economy: A Longitudinal Analysis of Ethnic Networks and

Pathways to Economic Success across Immigrant Categories

University of British Columbia

The economic benefits to immigrants of taking jobs in ethnic work

There has been considerable debate on the economic benefits to immi

310 IMR Volume 46 Number 2 (Summer 2012):310—361

Sanders (2001) recognize that immigrants arrive with different typ

advanced this literature look only at specific ethnic groups in particula

IMMIGRANT INCORPORA TION AND ETHNIC ECONOMIES

In the late 20th century, immigration to North America rose at remar

low-paying, low-skill jobs offering little opportunity for mobility (Piore,

dual labor market with immigrants remaining permanently in the ethni

and Stults, 2003), we must consider the outcomes of both employees in

IMMIGRANT CATEGORIES AND WORKPLACE TYPES

Canada's immigration policy, like those of many traditional immigrant

official languages, work experience, age, whether they have alread

friendships with those outside their ethnic group (Chiswick, 1

WORKPLACE TYPE AND ETHNIC NETWORK PORTFOLIOS

A longitudinal framework permits direct examination of how workpla

in which the ethnic economy is embedded help immigrants generate social

homogeneous network, which is categorized by two subtypes: similar and

ECONOMIC OUTCOMES OVER TIME

Pathways of immigrant incorporation have a temporal element by defi

Changes from the early stages after immigration to periods when

Even if ethnic jobs provide an immediate income advantage, immi

Diversifying a social network leads to new sources of informat

DATA AND METHODS

We used the restricted-access micro-level data of the Longitudinal Surve

and provincial databases, such as provincial healthcare records, and were

We use the following dependent variables at each interview wave: workpla

8These analyses are available from the authors upon request.

Immigration category is coded as a set of dummy variables based u

We use a series of dummy variables to indicate whether the respon

Age is directly related to labor market outcomes (Hansson et al„ 1997).

13High-status occupations include managerial and professional occupations. Medium-status

Static control variables that have been shown to be linked to labor

Xfco CNNrifONOVONH d(N(N-!rno6vovO(N

longitudinal data to be relatively large (Borjas, 1989). We ran a binary

The Relationship between Human Capital and Immigrant Category

We have suggested that economic immigrants are selected on the basis

their social capital, which raises the question of whether immigrant ca

We first present descriptive analysis of immigrants' transitions between

dependent variable in Wave 2. Other static variables, such as immigrant

To understand the patterns in how immigrants move from one type of

Workplace transitions of family and economic immigrants are

20These proportions are calculated by dividing those employed in an ethnic workplace by

in their study of Asian immigrants*in Los Angeles, supporting their

Of all economic immigrants who start out employed in ethnic wo

Table 3 shows logistic regressions predicting, among those working

Wave 1 Wave 2 Wave 3

Family immigrants 0.56*** 0.09 0.79 0.13 0.58*** 0.09

Family immigrants who work are less likely to enter a non-ethnic

Ethnicity of New Friendships

Many immigrants do not make any new friends. Overall, 13 percen

O (N O oo ir\ (N i—' NfooNt-mcOO xfco

Working in a job with an ethnic supervisor in a previous wave has

the more likely than White immigrants to make non-ethnic fri

0 CO ON 00 r-H O 1—1 CO T—I 1—1 CN 00 r^ 00 co •^r vo VT\ 0 ,—1

•s -a -g JS 43 £ § "2 "2 •="€ I 3 -SP"3 "I 2 "8 3 -SP-i _< w

<n cr> focof^icrifnc^fof^fo

& >s >s ^

there is no effect in Wave 1. In Wave 2, those with a high-school degree

DISCUSSION AND CONCLUSIONS

Family immigrants 0.56* 0.09 0.79 0.13 0.58* 0.09