CONCENTRATING ON PARTICIPATION
Neighbourhood effects on the participation of ethnic minority women in the labour force
Wenda van der Laan Bouma-Doff
OTB Research Institute for Housing, Urban and Mobility Studies
Delft University of Technology
P.O. Box 5030, 2600 GA Delft, The Netherlands
w.vanderlaanbouma-doff@tudelft.nl
**** STILL UNDER CONSTRUCTION, MY APOLOGIES! ****
Abstract
For decades, urban social scientists have been interested in the relationship between place of
residence on the one hand and social achievements, health, exposure to crime etcetera on the
other. With regard to labour market participation, the literature poses that the neighbourhood
context can produce opportunities as well as constraints for individual residents. Recent
empirical studies found especially negative neighbourhood effects on labour market success.
However, the effect of the neighbourhood is likely to differ in its extent and operative
mechanism for different population categories. This paper explores spatial effects on the
labour market participation of ethnic minority women, utilizing data from the survey ‘Social
Position and Use of Public Utilities by immigrants 2002’ matched with neighbourhood
characteristics. The results show that the neighbourhood matters, but that further in-depth
research is needed to explain observed neighbourhood effects.
1
1. Introduction
The importance of labour market participation for getting by and getting ahead does not really
need an introduction. Besides breadwinning and financial security, for many individuals a job
is an essential part of their personal fulfilment. Moreover, for women participating in the
labour force has an emancipative dimension as well: with it she gains economic
independence, emancipation and probably self-confidence. Whether a woman, in the end,
decides to take part in the labour force, depends in particular on her human capital, such as
her education, and the household situation she’s in.
In the Netherlands, the average labour market participation of ethnic minority women
is lower than that of native Dutch, although large differences between the various ethnic
groups exist. When looking at women with a job consisting of at least 12 hours of work per
week, Surinam women have the highest labour market participation rate (64%). Next in line
are Antillean women and native Dutch women (59% each). The differences between those
groups and the Mediterranean women are large: a labour market participation rate of 32% for
Turkish women and 30% for Moroccan women. Differences with regard to human capital and
demographic features will probably explain much of the variance in participation rates.
However, participating in the labour force, which contributes to current and future
socioeconomic achievements, is shaped not only by individual features but also by the
geographic context in which life decisions are made (Galster, 2005, pp.5). Individuals behave
and make choices within a given spatial context and this context is characterised by unevenly
divided, segregated opportunities. Therefore, it should be recognised that individuals or
households do not have equal opportunities, even if there is a resemblance on personal
characteristics.
The independent influence of the spatial context on individual opportunity structures,
preserving and reinforcing social inequality in society, is the subject of this paper. It explores
2
neighbourhood effects on the labour market participation of ethnic minority women, utilizing
data from the survey ‘Social Position and Use of Public Utilities by immigrants 2002’
matched with neighbourhood characteristics. The research question is: are ethnic minority
women’s participation rates affected by spatial concentration and if so, how can we explain
this?
2. Theoretical background
Among the theories on determinants of immigrants’ economic success, the ethnic enclave
theory focuses explicitly on the significance of the place and the neighbourhood context in
which immigrants live (Musterd , Andersson, Galster & Kauppinen, 2005). The importance of
ethnic enclaves is emphasized by Portes and Rumbaut (1990), who argue that selfemployment within the ethnic community highly attributes to the employment chances of
immigrants. Two remarks can be made on this view. The first is that the employment chances
within ethnic enclaves are scale-dependent. In American cities, where Portes and Rumbaut
did their research, large-scale ethnic communities can be found where job opportunities are
correspondingly large. In most European cities the scale of ethnic communities is much
smaller, maybe with the exception of some cities in Great Britain and France. The latter may
also explain the high rates of self-employment in Great-Britain (e.g. Metcalf, Modood &
Virdee, 1996). The second remark is that not much information is available on women’s share
in this specific ethnic employment. In her ethnographic study on women’s social networks,
Van der Zwaard (2005) found that ethnic entrepreneurship hardly increases job opportunities
for (married) women, because amongst ethnic minorities this circuit is mainly considered to
be ‘a man’s world’. The business clustering aspect of enclaves, however, is only one aspect of
the ethnic enclave theory, the residential clustering being another. Residential concentration
of immigrants might produce ‘ethnic capital’ (Borjas, 1998), including intergenerational
3
transmissions of social and human capital, norms for educational attainment, educational and
job formation, and employment opportunities.
For my purposes this theory is of interest when at the micro-level the immigrants'
social networks and their related social contacts are put forward as possible explanatory
variables of neighbourhood effects. That is, residents may benefit from resources and support
from spatial clustered social networks. Besides job opportunities and information on jobs, the
ethnic enclave might offer practical, social and emotional support (Portes & Sensenbrenner,
1993). Musterd et al. (2005) explored the relationship between ethnic clusters and economic
outcomes for immigrants in Sweden. Their general conclusion is that “both men and women
pay a substantial income penalty (…) if they reside (…) in areas where a substantial number
of their neighbours were members of the same ethnic group” (pp.31), this in spite of initially
positive effects of own group concentrations. Remarkably, the effects are stronger for women.
Which mechanisms are at work here has yet to be discovered. It could be argued that the
neighbourhood context has different functions for women in different household situations.
Mothers with young children in particular might profit from spatial clustering by gaining all
kinds of support from their own ethnic group, especially in relation to informal childcare.
Research on housing choices and neighbourhood satisfaction of residents of low-income
neighbourhoods, for example, showed that single mothers in concentration areas are more
content and less likely to move than their counterparts in other city neighbourhoods (Van der
Laan Bouma-Doff & Van der Land, 2007). It is plausible that local networks of kinship are
important in not leaving the neighbourhood as they might provide these vulnerable
households with practical, informative, social and emotional support. Indeed, the need for
these support networks determines women’s choices about their residential location (Vaiou &
Lykogianni, 2006).
4
Opposite to this positive view on spatial concentration, is the conception that
emphasizes the ‘dark side’ of networks and social bonds. The bright side of social networks
neglects the fact that social bonds may socialize residents in a way that hampers labour
market participation. With regard to this aspect of social bonds, members of the own ethnic
group living in the same neighbourhood could reject women’s labour market participation,
especially when the woman has a family to take care of. In-group social control may well
impede women's labour market chances (cf. De Vries, 1989). In this way, spatial
concentration influence a woman's labour market behaviour by the effect of in-group social
control and her own confirmation to social norms of the neighbourhood. Another drawback of
spatial concentration could be that women living in these circumstances lack the social
contacts that might be useful in the labour market outside the neighbourhood. Granovetter
(1974) stresses the significance of weak ties in finding a job. Those ties enable people to
participate in networks other than their own direct network, which could provide one with
useful (labour market) information. There are indications that contacts with native Dutch
operate as weak ties (Odé, 2002; Veenman, 2003) and the possibilities of establishing ethnic
bridges (contacts with native Dutch) are limited in multi-ethnic neighbourhoods to a certain
extent (Briggs, 2005; Van der Laan Bouma-Doff, 2007). Thus, in stead of stimulating job
opportunities thru ‘strong ties’, ethnic enclaves might reduce peoples’ opportunities by the
lack of ‘weak ties’.
In sum, spatial concentration may have positive as well as negative impacts on the
probability that a woman is active on the labour market. If we follow the positive view, the
network of kin and co-ethnic friends gaining from spatial proximity, positively influences
job opportunities and the possibility to participate. On the other hand, own group
concentrations could socialize women, resulting in more traditional views on women’s
liberation or hamper contacts with native Dutch, which might decrease the odds on labour
5
market participation. The next paragraph elaborates on the data, measurement and method,
used to unravel the spatial effects on women’s participation in the labour force.
3. Data, measurement & method
The individual data used in this paper were taken from the 2002 survey ‘Social Position and
Use of Public Utilities by immigrants’ (SPVA), administered by the Institute for Social
Economic Research (ISEO) of the Erasmus University Rotterdam and the Social and Cultural
Planning Office (SCP). The SPVA respondents were selected by a stratified sample based on
city and ethnic origin. From municipal registers of thirteen cities (including the four major
cities Amsterdam, Rotterdam, The Hague and Utrecht ) heads of households were selected
randomly and data were collected by means of face-to-face interviews (using structured
questionnaires). Besides the heads of households, all other persons in the household above the
age of 12 were asked to participate. The final SPVA data file contains data derived from
4,199 households, of which 1,173 Turkish, 1,056 Moroccan, 1,101 Surinamese and 869
Antillean/Aruban households. These ethnic groups constitute the four largest ethnic minority
groups in the Netherlands, and make up 67% of the entire group of non-western ethnic
minorities (CBS, 2005). From this data file, ethnic minority woman in the age of 15-50 were
selected (the age selection relates to participation in education on the one hand and to the
cohort-effect, as a result of which participation significantly reduces after the age of 50 on the
other hand). The remaining dataset contains data on 783 Turkish women, 650 Moroccan
women, 605 Surinamese women and 460 Antillean women. In the multivariate analysis the
Antillean women were dropped, since they hardly live concentrated, as we will see later on.
The zip code for each household in the SPVA is linked to population data from Statistics
Netherlands (CBS) for the year 2002. As a result, we know the ethnic composition of
neighbourhoods for each respondent and therefore the level of ethnic concentration. As many
6
scholars have pointed out so far, such administrative boundaries are not the most perfect
operational definitions of ‘the neighbourhood’(Dietz, 2002; Sampson et al, 2002; Lupton,
2003). On average, 1,740 people reside in each of the 327 postcode areas included in the
SPVA. Such areas might be too large in scale to accurately measure the variables of local
neighbourhood affecting residents (Friedrichs, Galster & Musterd, 2003). I nevertheless have
to rely on postcode areas since the information on lower spatial levels is lacking.
The dependent variable is female labour market participation, measured as having an
employment contract of at least 12 hours per week. From the survey data I also derive
information on age, education (highest diploma), household situation and motherhood. Spatial
concentration is measured by the percentage of the own ethnic group living in the
neighbourhood. To investigate thresholds, I divided this percentage into three categories:
neighbourhoods containing less than 10% of the own ethnic group, neighbourhoods
containing 10-20% of the own ethnic group and neighbourhoods containing more than 20% of
the own ethnic group. As Galster (2003) argues, the group has to reach some critical mass of
density over an area that is likely to become effective in shaping the behaviour of others (see
also Crane, 1991; Buck, 2001). Regrettably, my focus on effects of spatial concentration of
the own ethnic group implies that Antilleans have to be dropped from the multivariate
analyses. As a result of an evenly distribution over the cities, almost the entire Antillean group
(99%) fits into the first category. Consequently, the analyses in the following paragraphs are
on the three largest Dutch immigrant groups: Turks, Moroccans and Surinamese.
In the multivariate analyses, the probability of a woman participating on the labour
market constitutes the dependent variable, and the independent variable is the degree of
spatial concentration (own ethnic group concentrations). In order to find out whether or not
the level of spatial ethnic concentration is indeed influencing labour market participation, we
need to check for differences in individual background characteristics. If the participation
7
rates of ethnic minority women living in neighbourhoods with higher own group
concentrations differ from participation rates of their counterparts in non-concentration areas,
the question arises whether this is due to compositional effects, or is it, in fact, a concentration
effect. Perhaps the educational level differs between residents of different neighbourhoods,
and this characteristic has an effect on participation, not spatial concentration itself. For this
reason a multivariate model needs to be constructed in order to test for additional contextual
effects (Buck, 2001; Sampson et al, 2002; Galster, 2003). The association between spatial
concentration and labour market participation is controlled for by adding variables known to
affect (women’s) participation on the labour market, namely: age, household situation,
motherhood, education and length of stay in the Netherlands.
However, it is not possible to draw any firm conclusions about the causal effects of
spatial concentration. We need longitudinal data or experimental data to unravel the causality
in more detail. Moreover, it is not possible to entirely account for the potential problem of
neighbourhood selection, because individuals have unmeasured motivations which affect the
choice of residence. In other words, observed associations between ethnic concentration and
labour market success might be biased because of this spatial selection, even when all the
observable characteristics are controlled (Manski, 1995). On the other hand, recent studies of
Musterd, Andersson, Galster and Kauppinen (forthcoming) demonstrate that this particular
variable bias might not be very large. With these limitations of this study in mind, we now
continue with the results of the multivariate analyses.
4. Results
Table 1 shows the results of the analyses for the three groups separately. All three groups
show negative neighbourhood effects, although only Moroccan women are significantly
affected by spatial concentration. Furthermore, the results show the importance of
8
distinguishing thresholds. Only when the own group concentration rises above 20%,
Moroccan women show lower participation rates, strongly decreasing the odds on having an
employment contract of at least 12 hours per week.
Table 1 Logistic regression analysis of female labour market participation, effect of spatial concentration, B’s
and odds ratios
Turkish women
Moroccan women
Surinamese women
B
Exp(B) B
Exp(B) B
Exp(B)
15-24 years (ref)
25-34 years
35-44 years
45-54 years
Child(ren) living at home
No partner living at home
No formal education (ref)
Vbo/mavo
Mbo/havo/vwo
Hbo/wo (higher education)
Born in the Netherlands (ref)
≥10 years in the Netherlands
<10 years in the Netherlands
<10% own ethnic group in NH (ref)
10-20% own ethnic group
>20% own ethnic group
Constant
N
Nagelkerke R2
*= p<.05 **= p<.01 ***=p<.001
source: SPVA 2002 (ISEO/SCP)
-0.69 *
-0.80 *
-1.11 *
-0.13
0.35
0.50
0.45
0.33
0.88
1.42
0.64 **
0.92 ***
1.35 **
**
0.01
-0.80 *
1.90
2.52
3.88
1.04 **
1.53 ***
2.37 ***
2.84
4.64
10.75
1.01
0.45
-0.27
-1.60 ***
0.76
0.20
0.20
0.07
1.23
1.07
-0.09
-0.39
-0.26
0.92
0.68
0.77
-0.32
-1.29 **
0.20
0.72 0.19
0.28 -0.35
1.22 0.10
1.21
0.71
1.11
699
10%
-0.68
-1.18 **
-1.61 **
-0.50
0.65 *
0.51 -0.06
0.31 0.44
0.20 0.03
0.61 -0.47
1.92 -0.33
573
34%
0.64 **
1.67 ***
2.43 ***
0.94
1.55
1.03
0.63
0.72
1.89
5.30
11.34
557
19%
Thus, with regard to labour market participation, the effect of high spatial concentration is
quite strong for Moroccan women. But the question still remains why. Is it socialization, lack
of weak ties or another mechanism that is at work in the neighbourhood?
First, the socializing power of the neighbourhood which may hamper labour market
participation of women is examined. In this connection, the neighbourhood's social norms on
women's liberation should be concluded in the analyses. Unfortunately, this information is
lacking, which makes it impossible to test the assumption that ethnic minority women in high
9
concentration areas are affected by the social norms of their own-ethnic neighbours. However,
I do know the respondents’ own view on female liberation. Respondents were asked to react
on several statements about women's liberation and on the basis of their positive or negative
reactions a scale has been constructed (1-5) in which a lower score corresponds to a more
traditional view on women's liberation. A further disadvantage of this concept is that not all
female respondents were asked to give their opinion on this issue, since in many households
the head of the household, often a male, was asked to reply. Nevertheless, I will examine to
what extent the inclusion of views on women’s liberation in the model affects the influence of
spatial concentration.
Table 2: Logistic regression analysis of female labour market participation, effect of spatial concentration (I)
and inclusion of view on women’s liberation (II) (controlled for age, household situation, motherhood, education
and length of stay in the Netherlands), odds ratios
Turkish women Moroccan women Surinamese women
I
II
I
II
I
II
<10% own ethnic group in NH (ref)
10-20%
0.92 1.13
0.72
0.81
1.21
1.09
>20%
0.68 1.10
0.28** 0.26
0.71
0.71
Women’s liberation (scale)
2.30***
3.41***
1.95**
N
699
339
573
289
557
413
2
Nagelkerke R
10% 21%
34%
38%
19%
27%
*= p<.05 **= p<.01 ***=p<.001
source: SPVA 2002 (ISEO/SCP)
When the individual attitude to women’s liberation is accounted for, the effect of spatial
concentration is not significant any longer. However, the effect of spatial concentration does
not change that much. Therefore, with the available data I cannot trace an intermediate
relation between spatial concentration, social norms and female labour market participation.
Other research methods are probably more capable of bringing to light the socialization
mechanism, such as in-depth interviews and ethnographic studies.
Second, the lack of weak ties or bridging capital may be the operative mechanism
behind the observed negative effect of own group concentration. In the SPVA respondents
were asked if they have contact with native (‘white’) Dutch in their spare time. Those who
10
indicated not to have any contacts were coded 0 and those who said to have some or many
contacts were coded 1. Table 3 shows whether the effect of spatial concentration changes or
not.
Table 3: Logistic regression analysis of female labour market participation, effect of spatial concentration (I)
and inclusion of contact with native Dutch (II) (controlled for age, household situation, motherhood, education
and length of stay in the Netherlands), odds ratios
Turkish women Moroccan women Surinamese women
I
II
I
II
I
II
<10% own ethnic group in NH (ref)
10-20%
0.92 0.97
0.72
0.76
1.21
1.20
>20%
0.68 0.78
0.28** 0.28**
0.71
0.69
Contact with native Dutch
1.62*
2.06**
0.92
N
699
689
573
566
557
556
2
Nagelkerke R
10% 12%
34%
36%
19%
19%
*= p<.05 **= p<.01 ***=p<.001
source: SPVA 2002 (ISEO/SCP)
Ethnic bridges are important with regard to Turkish and Moroccan labour market
participation, yet it does not decrease the effect of spatial concentration at all.
To conclude, without further information on social norms and social networks it is
hard to tell what the reason is for negative concentration effects. Therefore, the results of this
paper are somewhat unsatisfactorily and further in-depth research should be conducted to
explain observed neighbourhood effects. In future research attention should be paid to other
neighbourhood mechanisms causing lower participation rates as well, such as lacking
resources and services in neighbourhoods and possible stigmatization of residents. The
relevance of the results presented here is that the geographic context matters when trying to
find explanations for low participation rates.
5. Conclusion and discussion
This paper dealt with differences in labour market participation of ethnic minority women and
the way they are affected by spatial concentration. From the theoretical overview on the
11
significance of the neighbourhood context, opposite hypotheses on the effects of own group
concentrations on female labour market participation could be extracted. The ethnic enclave
theory expects positive concentration effects as spatial clustered networks provide
employment opportunities for immigrants. However, the employment opportunities from
ethnic enclaves are probably not to the advantage of ethnic minority women. Another positive
feature of neighbourhood networks, though, is the linked practical, informative, emotional and
social support. In the theoretical section I also elaborated on the possible ‘dark side’ of social
bonds: the socialization mechanisms of the neighbourhood. Social norms of co-ethnic
residents may hinder labour market participation, because women are obligated to abide by
these norms or because they themselves have internalized these norms.
The results show that in general, spatial concentration (own group concentrations)
lower the probability that ethnic minority women participate on the labour market, but only
significantly for Moroccan women. I tried to uncover the mechanisms resulting in negative
concentration effects, but did not really succeed. The influence of spatial concentration does
not decrease when social norms (held by the woman herself) are accounted for. Nevertheless,
I expect social norms held by the residents (male or female) within the neighbourhood are
meaningful, but further research on these topics should be conducted. In any case, the
thresholds used in the analysis seem to point towards changes in attitudes and behaviour due
to living in concentrated neighbourhoods. Ethnic bridges did not change the effect of spatial
concentration either. Contact with native Dutch are important with regard to Turkish and
Moroccan labour market participation, yet it does not decreases the effect of spatial
concentration at all. In future research the networks of ethnic minority women – which is
something else than ‘having contact’ with native Dutch – should be studied in relation to the
spatial context. It would quite be interesting, for instance, to examine in more detail to what
extent the neighbourhood offers getting-by support and/or getting-ahead leverage (Briggs
12
1998). Finally, we should also find out more about the provision and quality of childcare
within the neighbourhood, in relation to the ability and the wish to participate on the labour
market.
References (not complete)
BORJAS, G.J. (1998) To Ghetto or Not to Ghetto: Ethnicity and Residential Segregation,
Journal of Urban Economics, 44, pp. 228-253.
BRIGGS, X. de SOUZA. (2005) Who bridges and how? Race, Friendships, and Segregation
in American Communities, unpublished, MIT, Cambridge, MA.
BUCK, N. (2001) Identifying neighbourhood effects on social exclusion, Urban Studies, 38,
pp. 2251-2275.
CRANE, J. (1991) The epidemic theory of ghettos and neighbourhood effects on dropping out
and teenage childbearing, American Journal of Sociology, 96, pp. 1226-1259.
DIETZ, R.D. (2002) The estimation of neighbourhood effects in the social sciences: an
interdisciplinary approach, Social Science Research, 31, pp. 539-575.
DREVER, A.I. (2004) Separate spaces, separate outcomes? Neighbourhood impacts on
minorities in Germany, Urban Studies, 41, pp. 1423-1439.
FRIEDRICHS, J., G. GALSTER & S. MUSTERD (2003) Neighbourhood effects on social
opportunities: the European and American research and policy context, Housing Studies, 18,
pp. 797-806.
GALSTER, G. (2003) Investigating behavioural impacts of poor neighbourhoods: towards
new data and analytic strategies, Housing Studies, 18, pp. 893-914.
GRANOVETTER, M.S. (1974) Getting a job: A study of contacts and careers. Cambridge:
Harvard University Press.
13
LAAN BOUMA-DOFF, W. van der & M. VAN DER LAND (2007) Hobson’s Choice:
Housing Choice and Neighbourhood Satisfaction in Low-income Neighbourhoods, paper to
be presented at the ENHR conference 2007.
LAAN BOUMA-DOFF, W. van der (2007) Confined contact. Residential segregation and
ethnic bridges in the Netherlands, Urban Studies, 44.
LAAN BOUMA-DOFF, W. van der (2005) De buurt als belemmering? [The neighbourhood
as an obstacle?] Assen: Van Gorcum.
LUPTON, R. (2003) ‘Neighbourhood effects’: can we measure them and does it matter?,
CASE paper 73, Centre for Analysis of Social Exclusion. London: London School of
Economics.
MANSKI, C.F. (1993) Identification Problems in the Social Science, Sociological
Methodology, 23, pp. 1-56.
MONTGOMERY, J.D. (1991) Social networks and labour-market outcomes: toward an
economic analysis, American Economic Review, 81, pp. 1408-1418.
MUSTERD, S., R. ANDERSSON, G. GALSTER, & T. KAUPPINEN (forthcoming) Are
immigrants’ earnings influenced by the characteristics of their neighbours?, Environment and
Planning A.
MUSTERD, S. & R. ANDERSSON (2006) Employment, Social Mobility and
Neighbourhood Effects: The Case of Sweden, International Journal of Urban and Regional
Research, 30, pp. 120-140.
ODÉ, A. (2002) Ethnic-cultural and socio-economic integration in the Netherlands. A
comparative study of Meditterarean and Carribean minority groups. Assen: Van Gorcum.
PORTES, A. & R.G. RUMBAUT (1990) Immigrant America: A Portrait. Berkeley:
University California Press.
14
PORTES, A. (ed.) (1995) The economic Sociology of immigration. New York: Russell Sage
Foundation.
PUTMAN, R. D. (2000) Bowling Alone. New York: Simon and Schuster.
SAMPSON, R.J., J. D. MORENOFF & T. GANNON-ROWLEY (2002) Assessing
“neighbourhood effects”: social processes and new directions in research, Annual Review of
Sociology, 28, pp. 443-478.
SIBLEY, D. (1995) Geographies of exclusion. Society and difference in the West.
London/New York: Routledge.
SQUIRES, G.D. & C.E. KUBRIN (2005) Privileged places: race, uneven development and
the geography of opportunity in urban America, Urban Studies, 42, pp. 47-68.
VAIOU, D. & R. LYKOGIANNI (2006). Women, Neighbourhoods and Everyday Life,
Urban Studies, 43, pp. 731-743.
VEENMAN, J. (2003) Allochtonen en hun sociaal kapitaal [Ethnic minorities and their social
capital], ESB-Dossier Sociaal Kapitaal, 88, pp. 24-25.
15