the effects of 9/11 on intermarriage between natives and immigrants to the u.s
Post on 25-Aug-2016
213 Views
Preview:
TRANSCRIPT
The effects of 9/11 on intermarriage between nativesand immigrants to the U.S.
Chunbei Wang • Le Wang
Received: 27 March 2011 / Accepted: 20 September 2011 / Published online: 5 October 2011
� Springer Science+Business Media, LLC 2011
Abstract The existing literature generally finds a negative impact of the 9/11
tragedy on immigrants’ labor market performance, consistent with increased dis-
crimination in the labor market and stricter immigration policies. In this paper, we
examine the impact of this tragic event on a particular measure of immigrants’
social outcomes—marriage with a native or intermarriage. We find that the tragic
event actually increases Hispanic immigrants’ probability of being married to a
native. We suggest that our results could be explained by that after 9/11, the
deteriorated labor market conditions, along with tightened immigration policies,
may have led to increased incentives of immigrants to marry natives. This effect is
large relative to the potential discrimination effect, if any, that could reduce natives’
willingness to marry an immigrant. We also find that the magnitude of the effect is
much smaller in the years immediately following 9/11 and becomes larger over
time; and that there exists a large, statistically significant gender difference in the
effects of 9/11 on intermarriage outcomes. Finally, we conduct indirect tests of
proposed explanations; and our results imply existence of economic gains from
intermarriage, and that discrimination may indeed exist.
Keywords Marriage � Intermarriage � Discrimination � Immigrants � 9/11
JEL Classification J12 � J14 � J15 � J61
C. Wang
University of Massachusetts at Dartmouth, Dartmouth, MA, USA
L. Wang (&)
Department of Economics, University of New Hampshire, Durham, NH 03824, USA
e-mail: le.wang@unh.edu
123
Rev Econ Household (2012) 10:171–192
DOI 10.1007/s11150-011-9136-3
1 Introduction and background
The tragic attacks on 9/11, as carried out by foreign terrorists, have triggered an
increase in patriotic and anti-immigration sentiments over the country. These
sentiments, along with national security concerns, fueled the passing of more
stringent immigration policies and increased discrimination against immigrants in
the U.S., although Muslims and Arabs were the direct victims of the 9/11 backlash,
the impact has extended beyond them to other immigrant populations, especially
Hispanic immigrants who constitute the majority of illegal/undocumented immi-
grants in this country (Passel 2006; Passel and Cohn 2009). The literature on 9/11
has typically focused on how it has affected the labor market outcomes among
immigrants. However, the impact may stretch far beyond their labor market
outcomes into their social lives, particularly marriage outcomes. This paper
complements the literature by examining the impact of 9/11 on intermarriage
outcomes (i.e. marriage with a native) among Hispanic immigrants and provides a
fuller picture of its impact on immigrants’ economic and social lives.
The post-9/11 immigration policies at both federal and state levels may have
changed Hispanic immigrants’ living conditions and incentives to marry natives.
First, a number of federal laws (e.g. the USA Patriot Act, the Homeland Security
Act, and the Real ID Act) were passed in the few years after 9/11 to fight terrorism
and illegal immigration.1 In addition to enacting new laws, the government also put
more efforts into enforcing existing immigration policies, resulting in diminished
job opportunities for illegal immigrants. For example, strengthened enforcement of
the ‘‘no-match letter’’ program alone led to thousands of Hispanic immigrants losing
their jobs after 9/11 (Sheridan 2002; Mehta et al. 2003; Orrenius and Zavodny
2009).2 Second, some states have also taken on their own immigration policies.
Actions were taken to limit welfare benefits, ban in-state tuition, prohibit rental
housing, or disallow state driver’s licenses for illegal immigrants. State and local
law-enforcement agencies also started joining the 287(g) program of the Immigra-
tion and Nationality Act after 9/11, which led to 81,000 illegal immigrants being
arrested from 2006 to 2008 (Vaughan and Edwards 2009).3
These collective efforts of federal and state governments mostly targeted illegal
immigrants. Since the majority of illegal immigrants in the U.S. are Hispanic
immigrants, these policies may have drastically harmed their job market
1 See Orrenius and Zavodny (2009) for a detailed summary. For example, the USA Patriot Act was
signed immediately after 9/11 to expand law enforcement access to private information and to allow them
greater authority to detain or deport immigrants. The Real ID Act (2005) set new federal standards that
require documentation of legal status for immigrants to obtain state driver’s licenses and ID cards.2 The ‘‘no-match letter’’ program reports to employers of employees whose social security number does
not match that on record. Other examples include more frequent large-scale illegal immigration raids
(Orrenius and Zavodny 2009). The E-verify program, an internet-based free program provided by the
government that allows employers to voluntarily check employees’ work eligibility, has also gained
popularity after 9/11. And in fact, several states made it mandatory that some or all companies use
E-verify for new hires.3 The 287(g) program allows state and local law-enforcement agencies to obtain training and
authorization from the Immigration and Customs Enforcement (ICE) office to identify and detain illegal
immigrants during their daily duty.
172 C. Wang, L. Wang
123
opportunities and various aspects of their everyday lives. Not only may illegal
Hispanic immigrants be affected, but legal Hispanic immigrants and the Hispanic
community as a whole may also be harmed because of potential discrimination (for
example, they may be more likely to be perceived as being illegal and thus
experience a decline in job opportunities). For example, Orrenius and Zavodny
(2009) show that recent male Hispanic immigrants indeed experienced deteriorated
labor market outcomes, including a decrease in wages, hours worked, and
employment opportunities.
The deteriorated labor market and general living conditions for Hispanic
immigrants after 9/11 may have created more incentives for them to seek stability in
alternative ways. Marrying a native (i.e., intermarriage) could be such an
alternative. Intermarriage has become more prevalent in the past century, and it
has long been recognized as an important way to improve immigrants’ economic
and social well-beings. Numerous studies (e.g. Baker and Benjamin 1997; Meng
and Gregory 2005; Meng and Meurs 2006; Kantarevic 2004; Furtado and
Theodoropoulos 2009, 2010) have shown that immigrants who are married to
natives tend to have higher earnings and better employment outcomes.4 Given the
large benefits of intermarriage, along with the worsened economic status of
Hispanic immigrants after 9/11, we hypothesize that 9/11 may have increased the
intermarriage rate among Hispanic immigrants.
On the other hand, an intermarriage decision is formed by both parties in a
marriage, and thus natives’ attitudes and preferences are also essential to the final
outcome. According to Becker’s theory of marriage, marriages tend to exhibit a
positive assortative mating pattern—people tend to marry individuals with similar
traits such as education, ethnicity, and religion (Becker 1974). In that sense, the
intermarriage phenomenon demonstrates a society’s attitudes toward the accep-
tance of various ethnic groups (Kalmijn 1998; Kalmijn and Van Tubergen 2010).
The catastrophic event of 9/11 has increased the hostile sentiments and
discrimination among natives toward immigrants. Unlike in other areas such as
labor markets, there are no laws regulating discrimination against minorities in the
marriage markets; preferences toward minorities may therefore be revealed more
fully in the marriage markets than in other places. The increased hostile
sentiments and discrimination toward immigrants after 9/11 may thus lead to a
decrease in the willingness of natives to marry immigrants, thereby decreasing the
intermarriage rates.
Even though our discussions seem to suggest that there are two opposing effects
of 9/11 on intermarriage among Hispanic immigrants, we believe that the increased
incentives among immigrants to marry natives (due to increased economic gains
from it) may be more important. We assess the empirical relevance of this
4 The positive effect of intermarriage on immigrants’ labor market performance comes from several
sources. First, immigrants may benefit from the expanded social network through intermarriage, which
could facilitate immigrants’ job search and locating better paid jobs. Second, immigrants may acquire
more human capital through intermarriage, such as improving language and communications skills. As a
result, intermarried immigrants tend to have higher productivity and thus higher wages and better
employment opportunities. Third, intermarriage allows immigrants to obtain permanent residence or
citizenship status which has been shown to have a positive effect on immigrants’ labor market outcomes.
The effects of 9/11 on intermarriage 173
123
hypothesis in this paper. In particular, we use the March Current Population Survey
1995–2010 and a difference-in-differences approach to estimate the impact of 9/11
on intermarriage among Hispanic immigrants. First, we find a positive effect of 9/11
on Hispanic immigrants’ intermarriages with natives, using other immigrants as the
control group. Further robustness checks imply that our results are not driven by any
particular subgroups or measures of intermarriage. Second, we find that the
magnitude of the effect is much smaller in the years immediately following 9/11 and
becomes larger over time. This result is consistent with the fact that discrimination
or hostile attitudes may be greater at the beginning, offsetting the positive effect of
9/11. Third, we find a large, statistically significant gender difference in the effect of
9/11 on intermarriage outcomes. Finally, we conduct indirect tests of whether there
exist economic returns from intermarriage and of whether discrimination indeed
exists; and our results imply that they do.
Admittedly, the mechanisms discussed above are not the only ones consistent
with the results found here. In this paper, we thus afford alternative explanations for
our results and discuss potential ways to distinguish these explanations should data
be available. We want to emphasize that the purpose of this paper is to document the
empirical effect of 9/11 on intermarriage outcomes. The alternative explanations
afforded here are not intended to be exhaustive, and it is our hope that these results
could stimulate further research on this topic and on marriage markets as an area of
potential discrimination.
2 Conceptual framework
To guide our empirical analysis, we utilize the theory of competitive marriage
markets (Becker 1993; Grossbard-Shechtman 1993). This theory analyzes marriage
outcomes within a competitive marriage market with aggregate demand and
aggregate supply, with men on one side and women on the other and a price (or
quasi-wage termed by Grossbard-Shechtman and Fu 2002) that measures how well a
person is treated in the marriage. The theory is useful and important for analyzing
all forms of marriage outcomes since marriage is always voluntary and men and
women compete for suitable mates. Existing literature has applied such a framework
to analyzing a broad variety of marriage issues such as religious intermarriage
(Grossbard 1983; Grossbard-Shechtman 1993), household chores (Grossbard et al.
2010), and marriage migration (Celikaksoy et al. 2006). As we shall see, this theory
would help us better understand the potential effect of 9/11 on immigrants’
intermarriage outcomes. More important, it could also help us interpret the
empirical estimates of the effect of 9/11 and distinguish potential mechanisms
behind this effect.
Adopting this framework, we analyze the native-immigrant intermarriage market.
For either women or men, we assume immigrants are on the demand side and
natives are on the supply side. The observed intermarriage rate is the equilibrium
outcome where the aggregate demand equals the aggregate supply. Factors such as
changes in preferences and population size may potentially shift the demand and/or
the supply and thus affect the intermarriage rate.
174 C. Wang, L. Wang
123
In light of our discussions above, the worsened economic outcomes among
immigrants and stringent immigration policies after 9/11 increased the values
associated with intermarriage for Hispanic immigrants. As a result, 9/11 may have
changed immigrants’ preferences toward intermarriage by increasing their willing-
ness to marry natives and to actively seek out native partners. Not only may more
immigrants enter the intermarriage market, thereby increasing the population with
demands for intermarriage; the existing participants on the demand side may also be
willing to pay a higher ‘‘price’’ to marry natives than before. Both types of changes
serve as demand shifters, increasing the equilibrium intermarriage rate, given
supply. We denote the positive effect of 9/11 demand-side effect.
Despite the potentially increased willingness of immigrants to marry natives due
to 9/11, the effect of 9/11 on the equilibrium intermarriage outcome may not
necessarily be positive. In addition to its impact on immigrants’ preferences and
attitudes, 9/11 could change natives’ attitudes toward intermarriage as well. Given
the increased hostile sentiments toward illegal immigrants after 9/11, one would
expect a decrease in the willingness of natives to marry these immigrants. Similarly,
this preference change would lead to a decrease in the number of natives available
in the intermarriage market. Moreover, natives who stay in the market may ask for a
higher ‘‘price’’ to marry an immigrant. As a result, the supply of natives in the
intermarriage market would decrease and shift to the left, and the tragic event would
exhibit a negative impact on the intermarriage rate for these immigrants, holding
everything else unchanged; we denote the negative effect of 9/11 supply-side effect.
To summarize, there could be potentially two opposing effects—both demand
and supply side effects—of 9/11 on the equilibrium intermarriage outcome. On the
one hand, if gains from intermarriage for immigrants became more pronounced after
9/11, this would increase demand and thus the number of intermarriages. On the
other hand, if the discrimination against immigrants increased and the anti-
immigration sentiment became stronger after 9/11, this would decrease supply and
the intermarriage rate. Such a decrease, if indeed exists, would imply an increase in
discrimination against immigrants after 9/11. An empirical analysis of the change in
the equilibrium intermarriage rate among Hispanic immigrants after 9/11 allows us
to assess which force prevailed.
3 Empirical methodology
To estimate the effect of 9/11 on immigrants’ intermarriage outcomes, we use the
difference-in-differences estimation method. By comparing the treatment group and
the control group over time, we are able to control for factors or changes that affect
both groups in a similar way. For example, the overall trend of inter-ethnic
integration, the cost of intermarriage such as the easiness of applying for legal status
for foreign spouses, changes in the intenseness of investigation on immigrant
marriage fraud, or the overall trend of marriage and cohabitation decisions may all
have changed over time and would have affected intermarriage between immigrants
and natives even in the absence of 9/11. By choosing an appropriate control group,
we will be able to control for these confounding effects without having to measure
The effects of 9/11 on intermarriage 175
123
them and control for these variables explicitly in estimations. We use European and
other immigrants as the main control group. This control group is appropriate
because European and these immigrants were generally not the target of increased
discrimination and strengthened immigration policies after 9/11, and thus the
overall trend of their intermarriage outcomes is less likely affected by the event.
Following the literature (e.g. Orrenius and Zavodny 2009), we conduct a
falsification test and use an alternative control group to assess the validity of the
difference-in-differences estimation below.
The difference-in-differences (DD) estimator of the 9/11 effect can be expressed
as
b ¼ ðYTpost�9=11 � YT
pre�9=11Þ � ðYCpost�9=11 � YC
pre�9=11Þ
where Y measures the intermarriage outcome. T refers to the treatment group (i.e.
Hispanic immigrants) and C the control group (i.e. European and other immigrants).
The term in the first parentheses measures the difference in the intermarriage out-
comes of the treatment group, before and after 9/11. The term in the second
parentheses measures the difference in the intermarriage outcomes of the control
group, before and after 9/11. These two differences are then differenced to single
out the impact of 9/11 on the intermarriage decisions of Hispanic immigrants. The
following regression model is used to obtain the difference-in-differences estimator:
Yist ¼ aþ b1Ti þ b2Post911t þ bTi � Post911t þ XistBþ at þ bs þ uist
where Yist is a binary variable that is equal to 1 if immigrant i in state s at time t is
married to a native, and 0 otherwise. Ti is a dummy variable equal to 1 if immigrant
i is in the treatment group, and 0 if immigrant i is in the control group. Post911t is a
dummy variable equal to 1 if the outcome is observed after 2001, and 0 otherwise.
Ti * Post911t is the interaction term of the treatment group dummy and the after-9/
11 dummy. The coefficient of this interaction term b is the difference-in-differences
estimator, which measures the effect of 9/11 on the intermarriage outcomes of
Hispanic immigrants.
Since individual characteristics can also affect intermarriage outcomes, we
include a number of control variables to improve efficiency. Xist is a vector of
individual socio-economic characteristics including gender, age, age squared/100,
years of schooling, and years in the U.S.5 Furthermore, we include a year fixed
effect at and a state fixed effect bs to control for other unobserved state- and year-
specific heterogeneities. Following the literature (Bertrand et al. 2004), 9/11
affected people from different immigrant backgrounds, and thus we report robust
standard errors clustered at immigrant groups (i.e. Hispanic vs non-Hispanic
immigrant groups).
A number of social and economic studies have examined various determinants of
intermarriage outcomes (Fossett and Kiecolt 1991; Wood 1995; Lichter et al. 1992;
Kalmijn 1998; Anderson and Saenz 1994; Chiswick and Houseworth 2011). As
summarized by Chiswick and Houseworth (2011), these determinants can be loosely
5 The literature also suggests that language ability is an important determinant of intermarriage.
Unfortunately, such information is not available in the data.
176 C. Wang, L. Wang
123
classified into two main categories: ‘‘preference’’ and ‘‘opportunity for contact’’.
Factors such as language skills, human capital, discrimination, and economic
conditions fit into the ‘‘preference’’ category as they affect the tastes and preferences
of different ethnic groups toward intermarriage; factors such as the size and sex ratio
of different ethnic groups fall in the category of ‘‘opportunity for contact’’, as they
affect the opportunity of different ethnic groups interacting with each other.6
To simplify the interpretation of our results, we focus on the reduced-form effect
of 9/11 in this paper. That is, the estimates capture the total effect of 9/11 on
intermarriage through these two different sets of factors. Of the two categories, the
9/11 tragic event would impact intermarriage outcomes mainly by changing
people’s preferences, if it has any effect at all. While opportunity for contact is an
important type of determinants of intermarriage, it seems irrelevant in the 9/11
context, as the size and sex ratio of different ethnic groups are unlikely to have
changed substantially as a result of 9/11.7 Below, the summary statistics also
confirm that the effect of 9/11 on the sex ratio is not significant. Thus we focus our
discussion and estimation on capturing the preference changes after 9/11.
4 Data
The data are obtained from the March supplement of Current Population Survey
(CPS) over the period 1995 and 2010 (available at http://cps.ipums.org) (King et al.
2010). We restrict our sample to individuals aged between 20 and 40. The outcome
of interest is the intermarriage outcome, which is constructed using data on spousal
country of birth. It is defined as a binary variable, equal to 1 if the spouse was born
in the U.S. and 0 otherwise.
We define two categories of immigrants: Hispanic and European, Canadian and
other immigrants. Hispanic countries includes Mexico, Central American and
Caribbean countries, and South American countries. We exclude Muslims,
Arabians, Indians, and what is defined as other African countries in this analysis.
The remaining immigrant population forms our comparison group.8
In our analysis, we include gender, years of schooling, age, age squared, and
years in the U.S. as additional controls. Years of schooling is not directly available
in our data. The CPS asked individuals about the highest grade or diploma
completed instead of years of schooling. Moreover, grades completed below 9th
6 Chiswick and Houseworth (2011) also categorize a third group as factors that affect both preference and
opportunity for contact. Strictly speaking, all factors can more or less affect both, thus we discuss two
categories instead of three.7 Qian and Lichter (2007) and Lichter et al. (2007) find that the increased inflow of Hispanic and Asian
immigrants results in more endogamous marriages over time; this nation-wide trend is captured by year
fixed effects in the estimations.8 Following Kaushal et al. (2007), we define Muslim and Arabian countries as those countries on the
special registration list of the Department of Justice. These countries include Afghanistan, Algeria,
Bangladesh, Cyprus, Egypt/United Arab Republic, Indonesia, Iran, Iraq, Jordan, Lebanon, Morocco,
Nepal,Pakistan, Saudi Arabia, Syria, and Yemen. We also include Malaysia and Turkey, countries with
predominantly Muslim populations, as in Kaushal et al. (2007).
The effects of 9/11 on intermarriage 177
123
grade are reported in three-year intervals. Following the literature (e.g. Lemieux and
Card 2001; Henderson et al. 2010), we replace these intervals with midpoints.9
Years in the U.S. is defined as the difference between survey year and year of
immigration.10
Table 1 reports summary statistics of all variables used in the analysis separately
by different immigrant groups and post-911 status. Intermarriage rates across groups
are similar to the ones reported in Kalmijn and Van Tubergen (2010). The
percentage of immigrants married to natives declined after 9/11 for Hispanic
immigrants. The decrease in intermarriages over time suggests that discrimination
may exist in the marriage market and dominate the positive effect of 9/11. Notice
that the control group also experiences a slight decrease in intermarriages over time.
This could be caused by changes in the overall trend of inter-ethnic marriages in the
absence of 9/11. To identify the discrimination and incentive effects of 9/11 on
intermarriage outcomes among Hispanic immigrants, it is important to control for
these changes and thus calls for a formal difference-in-differences analysis. We now
turn to our formal analysis.
Table 1 Summary statistics
Hispanic immigrants Other immigrants
(1) (2) (3) (4)
Pre-911 Post-911 Pre-911 Post-911
Intermarriage 0.173 0.171 0.391 0.371
(0.378) (0.377) (0.488) (0.483)
Male 0.481 0.475 0.426 0.410
(0.500) (0.499) (0.495) (0.492)
Years of schooling 10.339 10.704 14.659 15.317
(4.127) (4.015) (3.752) (3.634)
Age 31.708 32.242 33.201 33.657
(5.235) (5.181) (4.792) (4.654)
Age squared/100 10.328 10.664 11.253 11.545
(3.264) (3.259) (3.078) (3.018)
Years in the U.S. 12.125 12.612 13.521 14.044
(7.643) (8.040) (10.35) (10.473)
No. of obs. 18,684 33,926 8,446 16,502
Standard deviations in brackets. The data are from IPUMS (http://cps.ipums.org/cps/). The paper
examines the effect of 9/11 on intermarriages for two ethnic groups: Hispanicsand Muslims. The sum-
mary statistics are reported for each group separately. See the text for the base category of the variable
‘‘ethnic group’’
9 The coding scheme to convert the education variable to years of schooling is as follows. No schooling
completed (0); Nursery school (0); Kindergarten (0); 1st to 4th grade (2.5); 5th grade to 8th grade (6.5);
9th grade (9); 10th grade (10); 11th grade (11); 12th grade, no diploma (12); High school graduate (12);
Some college (13); Associate degree (14); Bachelor’s degree (16); Master’s degree (18); Professional
degree (20); Doctorate degree (20).10 Year of immigration is also reported in intervals in the data. To be conservative, we use the upper
bound of year of immigration interval.
178 C. Wang, L. Wang
123
5 Results
5.1 Baseline result
Table 2 first presents the baseline result. The reported coefficient is obtained
from the linear probability model, and the robust standard error clustered at
immigrant group is in parentheses. The models include additional variables (state
and time fixed effects, gender, years of schooling, age, age squared, and years in
the U.S.).
We start by discussing the result in column (1). The coefficient of interest is the
DD estimate, capturing the effect of 9/11 on the probability of being married to a
native for Hispanic immigrants. We find a positive effect of 9/11 on intermarriage
probability for Hispanic immigrants. This effect is statistically significant at
p B 0.05 level. However, as noted in Cameron et al. (2008), the conventional
cluster-robust error reported here may cause over-rejection of the null hypothesis
when the number of clusters is small (for example, 2 clusters in our case). Cameron
et al. (2008) advocate use of the wild bootstrap that imposes the null hypothesis and
uses ?1, -1 weights to obtain p value for the test of statistical significance.
Following the procedure detailed in Cameron et al. (2008), we obtain the wild
Table 2 Difference-in-differences (DD) estimates of the effect of 9/11 on intermarriage among Hispanic
immigrants and specification tests of DD design
Baseline Specification tests
A. 1998 as the break
date & years before 2001
B. Alternative
control group
(1) (2) (3)
Hispanic -0.115*** -0.124*** -0.129***
(0.007) (0.010) (0.007)
Post-911 -0.065*** -0.016 -0.060***
(0.011) (0.015) (0.012)
DD estimates 0.022** -0.004 0.023***
(0.008) (0.015) (0.008)
Time fixed effects Yes Yes Yes
Control variables Yes Yes Yes
State fixed effects Yes Yes Yes
No. of obs. 77,558 18,102 73,534
Control variables in the models include gender, years of schooling, age, age squared and years in the
states
Column (1) reports the baseline results. Columns (2) and (3) report the specification tests of our dif-
ference-in-differences design. Column (2) reports the results using the sample prior to 2001 and 1998 as
an artificial break date. Column (3) reports the results using a refined control group including only
Canadian, Europeans (excluding Eastern Europeans), and Asians
Robust standard errors (clustered at immigrant group) in brackets. *** p \ 0.01; ** p \ 0.05; * p \ 0.1
The effects of 9/11 on intermarriage 179
123
bootstrapped p value less than 0.0001 based on 599 replications, and thus our
conclusion is even strengthened.11
The coefficient implies that 9/11 increases the probability of being married to a
native by roughly 2.2 percentage points. As discussed above, two opposing
effects—the demand-side and supply-side effects—of 9/11 on intermarriage may be
at work. There are two explanations consistent with the observed positive effect.
First, discrimination may not have increased in the marriage market at all after 9/11.
Second and more plausibly, discrimination increased after 9/11, but the effect of
increased discrimination on marriage outcomes is small relative to the positive
effect associated with potential economic gains. Both explanations suggest that
there exists a positive effect; that is, after 9/11, the deteriorated labor market
conditions and stringent immigration policies targeting illegal immigrants may have
led to increased incentives for Hispanic immigrants to marry natives.
A number of additional aspects of the baseline result in Table 2 are also worth
mentioning. First, we notice that the coefficient for the Hispanic group dummy is
negative and statistically significant. That is, compared to other immigrants,
Hispanic immigrants are less likely to marry natives, consistent with the literature
(see, e.g. Kalmijn and Van Tubergen 2010). Given the large gains from
intermarriage, our result implies that Hispanic immigrants may assimilate into the
local society at a slower pace than other immigrants. Second, the estimated
coefficient of post-911 dummy is also negative and statistically significant. This
result implies that the overall trend of intermarriage is decreasing after the tragic
event of 9/11, and the magnitude of the overall decline is 6.5 percentage points.12
5.2 Validity of the difference-in-differences method
Thus far, we have assumed the difference-in-differences method works in our
context. But the validity of the difference-in-differences method hinges on the
appropriateness of the control group used. We perform three exercises to assess this
issue.
First, following the literature that examines the effect of 9/11 on immigrants’
economic outcomes (e.g. Kaushal et al. 2007; Orrenius and Zavodny 2009), we
conduct a falsification test by choosing an artificial intervention event prior to 2001
for our analysis. If the control group is indeed comparable and the difference-in-
differences is valid, we would expect that the coefficient of the difference-in-
differences is not statistically distinguishable from zero. The results are presented
in Table 2, column (2). Following Orrenius and Zavodny (2009), we use 1998 as the
11 Please note that the Monte Carlo simulation results presented in Cameron et al. (2008) are still limited.
The smallest number of clusters considered in their paper is 5. Further studies of the applicability of the
wild bootstrap technique in the cases of even smaller clusters are undoubtedly still warranted. And the
robustness of our result should be further examined when other techniques are available. This is, however,
beyond the scope of this paper and left for future research.12 In addition, we find that all the additional variables are also important determinants of intermarriage.
For Hispanic immigrants, education increases the probability of being married to a native, consistent with
the existing literature (e.g. Furtado, in press). Staying in the U.S. longer could also increase the
intermarriage rates. In the interest of space, these results are omitted but available upon request.
180 C. Wang, L. Wang
123
artificial break date and restrict the sample to the years prior to 2001. We find that
the magnitude of the DD estimate is much reduced, and that the DD estimate is no
longer statistically significant at any conventional level.
Second, we also re-estimate all the analyses using an alternative control group
that includes only Canadian, Europeans (excluding Eastern Europeans), and Asians.
The results are presented in Table 2, column (3). As we can see, the results are
similar to the baseline results. Overall, these results indicate that our differences-in-
differences estimates are valid in this context. Ideally, we may want to further refine
the control group by including either only Canadian and Europeans or only Asians.
However, when using either sub-sample, the estimates are close to zero, and none of
the results are statistically significant. These discrepancies between pooled sample
results and subsample results indicate a small sample problem. Any further
reduction in the sample size may prevent us from distinguishing truly insignificant
results from insignificant results due to small sample sizes.
Finally, we check whether the treatment and control groups were comparable
prior to 9/11. We show this by examining whether there existed differential trends in
the intermarriage rates between the treatment and control groups prior to 9/11. We
thus present the mean intermarriage rates over time for both groups in Fig. 1. The
vertical line is year 2001. Remember that the March CPS was collected in every
March, and thus the years before and including 2001 are classified as the pre-9/11
period. As we can see, the intermarriage trends of the two groups prior to 9/11 were
more or less the same, fluctuating around their own natural levels. During the first
four years, the cycles of the trends between these two groups almost match each
other perfectly. More interestingly, there was a dramatic increase in the
0.1
.2.3
.4
Inte
rmar
riage
Rat
e am
ong
Imm
igra
nts
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Year
Hispanic Non−hispanic
Fig. 1 Regression-adjusted means of intermarriage rates by immigrant group (1995–2010). Solid line ismean intermarriage rates for Hispanic immigrants; dashed line is mean intermarriage rates for non-Hispanic imigrants. Means have been regression-adjusted for gender, years of schooling, age, agesquared, and years in the USA
The effects of 9/11 on intermarriage 181
123
intermarriage rates among Hispanic immigrants after 9/11, as expected. Consistent
with the discussion of the data above, the control group experienced a slight
decrease in intermarriages in later years. This change could reflect the changes in
the overall trend of intermarriage that potentially continues to decline over time, and
thus it highlights the importance of including both the post 9/11 dummy and the
time fixed effects in our estimations.
5.3 Robustness checks
In this section we further check the robustness of our results to different subgroups
and to particular measures of intermarriage. This practice may also help us shed
some light on potential mechanisms of the 9/11 effect on intermarriage.
First, as suggested in the literature (e.g. Orrenius and Zavodny 2009), the
recession in 2001 may impact immigrants adversely, especially for those who are
illegal immigrants and those in low-paid jobs. This could change immigrants’
intermarriage decisions in a similar way as 9/11 did by increasing gains from
intermarriage. As such, the recession in 2001 could also lead to an increase in
intermarriages, confounding the 9/11 effect. Recall, again, that the year 2001 is
classified as the pre-9/11 period (all data were surveyed in every March), one may
thus be concerned that inclusion of year 2001 may underestimate the impact of 9/11
on intermarriage. To this end, we exclude data from 2001 and re-estimate our
models. The results are reported in column (1), Table 3. Confirming our hypothesis,
our results imply a slightly larger positive effect of 9/11 on intermarriage. In what
follows, we present the results using the sample including the data in 2001.
However, we do estimate all the specifications below using the data excluding year
2001. The results are strengthened, as expected, and our conclusions remain
unchanged. The results are thus omitted but available from the authors upon request.
Second, we use alternative measures of intermarriage as the dependent variable.
Intermarriage is previously defined as whether an immigrant is married to a native.
This definition includes marriages with second-generation (or multiple-generation)
immigrants or other natives whose antecedents are immigrants from the same
country. Since there is a large percentage of native-born Hispanics in the U.S., we
may want to exclude these individuals, as discrimination, if it exists, does not
necessarily reflect itself in the marriages with native-born offsprings of immigrants.
That is, inclusion of these marriages may underestimate the extent of discrimination,
and the resulting estimates tend to be larger and more positive. We utilize the
information on Hispanic origin in the CPS and re-estimate the model for the
Hispanic sample focusing on intermarriage with non-Hispanic natives. The effect of
9/11 on intermarriage, reported in column (2) of Table 3, is indeed smaller;
although the coefficient remains positive and statistically significant, the magnitude
of the effect drops by about 36%. This result is consistent with our discussion above
and suggests possible existence of discrimination.
Third, as discussed above, marriage with a native not only helps immigrants’ job
market outcomes through better knowledge of the labor market or more social
network in the states, but also helps illegal immigrants obtain legal status that could
eventually qualify them for a broader variety of jobs. While both mechanisms imply
182 C. Wang, L. Wang
123
an increasing return to intermarriage for Hispanic immigrants after 9/11, the latter
factor is more important for undocumented Hispanic immigrants than for authorized
immigrants; this implies a larger total effect for undocumented immigrants. We
want to assess whether or not our result is simply driven by the undocumented
immigrants and whether the positive incentive effect solely comes from obtaining
legal status. Measuring undocumented immigrants is not a trivial task, since the CPS
data do not ask about legal status in the country. However, Furtado and
Theodoropoulos (2010) and Passel (2006) note that undocumented immigrants are
more likely to (1) come from Mexico and Central America (2) live in California,
Texas, Florida and New York, and (3) have less than high school education.
Following their approach, we define those who meet these three criteria as a proxy
for whether they are potentially undocumented immigrants, and re-estimate our
model using only undocumented Hispanic immigrants and the original control
Table 3 Robustness checks of difference-in-differences (DD) estimates of the effect of 9/11 on inter-
marriage among hispanic immigrants
A:
Excluding
data from
2001
B:
Intermarriage
with non-
hispanic
native only
C:
Undocumented
immigrants
only
D.1
Immigrants
who Arrive
under 22
D.2
Drop couples
who arrive
in the same
year
E.
Immigrants
who came
before 9/11
(1) (2) (3) (4) (5) (6)
Ethnic
group
-0.123*** -0.198*** -0.150*** -0.122*** -0.148*** -0.111***
(0.007) (0.006) (0.009) (0.009) (0.009) (0.007)
Post-911 -0.070*** -0.072*** -0.089*** -0.091*** -0.078*** -0.097***
(0.011) (0.010) (0.013) (0.015) (0.015) (0.013)
DD
Estimates
0.031*** 0.014* 0.026*** 0.022** 0.034*** 0.016**
(0.008) (0.007) (0.008) (0.010) (0.010) (0.008)
Time fixed
effects
Yes Yes Yes Yes Yes Yes
State fixed
effects
Yes Yes Yes Yes Yes Yes
Other
controls
Yes Yes Yes Yes Yes Yes
No. of obs. 72,170 72,285 38,642 47,867 52,805 64,490
Control variables in the models include gender, years of schooling, age, age squared and years in the
states
Column (1) reports the results using the full sample but excluding year 2001. Columns (2) reports the
results using a refined definition of intermarriages: intermarriages with non-Hispanic natives only. Col-
umn (3) reports the results restricting the sample to non-Hispanic and Hispanic undocumented immi-
grants. Undocumented immigrants are those (1) whose birth places are Mexico and Central America; (2)
who live in California, Texas, Florida, and New York; (3) and who have less than high school education
(see, e.g. Furtado and Theodoropoulos (2010) for more discussion). Columns (4) and (5) restrict to the
sample of who are more likely to be married after arriving in the states: immigrants who arrive under age
22 (Column 4) and immigrant-immigrant couples that arrive in the same year (Column 5). Column (6)
keeps only the immigrants who came before 9/11
Robust standard errors (clustered at immigrant group) in brackets. *** p \ 0.01; ** p \ 0.05; * p \ 0.1
The effects of 9/11 on intermarriage 183
123
group. The result is reported in column (3) of Table 3. As expected, the coefficient
is larger than the baseline result by 0.4 percentage points.13 This is consistent with
our discussion. However, since the result is still close to the baseline result, this
implies that the incentive to marry a native does not solely come from legal status,
but from other sources as well. Interestingly, the result also suggests that obtaining
legal status is an important mechanism through which 9/11 affected intermarriage
outcomes.
Fourth, one may be concerned that some of the immigrants arrived in the U.S.
already married, and that their marriage decisions were not affected by 9/11 per se,
thus contaminating the results. We assess the robustness of our results against this
concern in two ways. As noted in Georgarakos and Tatsiramos (2009), it is more
likely that immigrants who arrived at a later age were married prior to their arrival.
Thus, we first exclude those who arrived after age 22. Marriage rates are usually low
among students, and individuals aged 22 and above are more likely to have
completed their education. The result is displayed in column (4) of Table 3. As we
can see, the result is almost identical to the baseline result. An alternative test is to
exclude those immigrant-immigrant couples who arrived in the same year. Our
result is even strengthened, as shown in column (5) of Table 3. Overall, it appears
that our results are not driven by the immigrants who were married before their
arrival.
Finally, even though the size and sex ratio of immigrants remained stable after
9/11, as mentioned above, it is still possible that other dimensions of immigrants’
characteristics have changed. In that case, our results cannot indicate if immigrants
changed their behavior after 9/11 or if 9/11 changed the composition of immigrants.
We thus re-estimate our model excluding those immigrants who arrived after 9/11.
Although the coefficient now becomes slightly smaller, we still find a statistically
significant and positive effect of 9/11 on intermarriage [Table 3, column (6)] among
Hispanic immigrants. The result suggests that the majority of the positive effect of
9/11 comes from a change in the behavior of existing immigrants and not from a
change in composition of the immigrants.
5.4 Heterogeneity results
To this point, we have been assuming that the effect of 9/11 on intermarriage is
homogeneous across gender and over time. The gains from intermarriage as well as
the extent of discrimination may however vary across gender and over time. To shed
further light on these issues, we investigate whether there exist any time-varying
effects of 9/11 as well as heterogenous effects across gender.
First, whether the tragic event has varying effects on marriage outcomes over
time depends on whether the event has different impacts on the factors we discussed
above. On the one hand, if a surge in the hostile attitude was temporary after 9/11
and faded away gradually, the negative effect of 9/11 on intermarriage related to
discrimination would become smaller over time. On the other hand, stricter policies
13 The result using only documented Hispanic immigrants and the original control group indicates 9/11
increased the intermarriage rates among them by 1.8 percentage points, and it is statistically significant.
184 C. Wang, L. Wang
123
and laws are likely to be permanent or in effect for a long period; also, due to
lengthy legislation process, there would be a time lag for any changes. The gains
from intermarriage are likely to be persistent, and the resulting positive impact may
remain roughly the same. Altogether, we may expect the positive demand-side
effect of 9/11 on intermarriage to become even larger over time. We thus
re-estimate our models by adding interactions between dummy variables (indicating
different time periods) and the treatment dummy. The results are reported in
Table 4. We observe that the magnitude of the effect gets larger over time; this is
consistent with our discussion that the positive economic gains from intermarriage
remain constant over time, while the negative effect of 9/11 related to discrimi-
nation gets smaller over time.
Second, our baseline results indicate there exists a mean gender difference in
intermarriage outcomes. On average, males are 0.6 percentage points more likely to
intermarry than females, and it is significant at 0.1 level. In the U.S. there also exists
evidence of gender asymmetry in racial preferences for interracial marriages (e.g.
Fisman et al. 2008). This leads us to suspect there could be a gender difference in
the impact of 9/11 on immigrants. In order to assess the existence of the gender
difference in the impact of 9/11, we re-estimate our models by adding an interaction
between gender, the treatment dummy, and the post-9/11 dummy. The coefficient of
this interaction measures the gender difference in the effect of 9/11. The results are
reported in column (2) of Table 4. We notice a striking gender difference in the
Table 4 Difference-in-differences (DD) estimates of the effect of 9/11 on intermarriage among Hispanic
immigrants by time and gender
Time-varying effects Gender-varying effects
(1) (2)
Hispanic -0.115*** -0.113***
(0.007) (0.006)
Post-911 -0.063*** -0.065***
(0.011) (0.011)
DD estimates
(2002–2005)
0.007
(0.009)
DD estimates
(2006–2010)
0.019**
(0.009)
DD estimates 0.051***
(0.008)
Female x DD Estimates -0.055***
(0.006)
Time fixed effects Yes Yes
Control variables Yes Yes
State fixed effects Yes Yes
No. of obs. 77,558 77,558
Control variables in the models include gender, years of schooling, age, age squared and years in the
states
Robust standard errors (clustered at immigrant group) in brackets. *** p \ 0.01; ** p \ 0.05; * p \ 0.1
The effects of 9/11 on intermarriage 185
123
effect of 9/11. For Hispanic male immigrants, the positive effect is very large—an
increase of 5.1 percentage points. By contrast, for Hispanic female immigrants, the
DD effect of 9/11 is 5.5 percentage points less than that for male. In other words, the
effect of 9/11 on Hispanic female immigrants is a small negative effect
(5.1–5.5 = -0.4 percentage points).
Our results imply that the positive effect we found in our base model is mainly
driven by males. We think that this gender difference may be driven by the different
roles of men and women in traditional households. Hispanics tend to be traditional
when it comes to gender roles (Roehling et al. 2005), implying that men tend to be
breadwinners and women tend to be more active in household production than
men.14 By lowering job opportunities and occupational stability 9/11 made it more
difficult for men to satisfactorily fulfill the expectations of a breadwinner, which led
more men to seek spouses with the stability that native status offers. This caused the
upward shift in demand discussed earlier. This shift is expected to be smaller in the
case of women, who were already seeking high earnings and job stability in a
spouse before 9/11, as part of the prescribed roles of traditional wives. Absent a
significant change in demand among women, the negative effect of 9/11 prevails,
implying the existence of discrimination. Moreover, if females are perceived as less
of a threat than males, the discrimination against female immigrants would be
smaller, relative to their male counterparts. Therefore, the negative effect observed
for females could be seen as a lower bound of the discrimination effect for males.
6 Discussion
6.1 Are there any economic gains from intermarriage?
Our results consistently imply that there exists a large positive effect of 9/11 on
intermarriage among Hispanic immigrants. In light of our discussions above, one of
the main sources of the positive effect comes from the increased demand for
intermarriage that may potentially improve the labor market outcomes among
immigrants. In this section we are interested in whether 9/11 affected Hispanic
immigrants’ intermarriage decision via this particular channel. Ideally, we want to
measure ex ante returns to intermarriage (and in turn incentives to marry a native).
This is a more direct test of the mechanism. However, such information is generally
not available and difficult to estimate. Pre-marriage labor market outcomes
(employment status and earnings)—which may be used as proxies for ex ante
returns to intermarriage—are also not available in the cross-sectional CPS data.
Instead we test the mechanism by examining ex post economic gains from
intermarriage. Assuming information about marriage choices is nearly perfect and
uncertainty plays a minor role, the ex post economic returns to intermarriage should
well approximate the ex ante returns. To estimate the ex post economic returns to
14 This is even more so among undocumented immigrants. For example, Passel and Cohn (2009) find that
only 58% of female undocumented immigrants participate in the labor force, which is much lower than
the participation rate for native women or legal female immigrants.
186 C. Wang, L. Wang
123
intermarriage, we examine the effects of intermarriage on both employment status
(whether or not one is employed) and log earnings among immigrants. Our results
are reported in Table 5. We find that intermarriage is indeed associated with better
labor market outcomes. Being married to a native increases employment rates by
1.2% and earnings by 21.8%. This is consistent with the literature (e.g. Furtado and
Theodoropoulos 2010). Moreover, we also find that Hispanic immigrants’ job
market prospects deteriorated after 9/11, as shown by the negative DD estimates.
Conditioning on intermarriage, this effect captures the direct effect of 9/11 on
immigrants’ labor market outcomes. Altogether, the positive effect of 9/11 on
intermarriage (our baseline result) and the positive effect of intermarriage on labor
market outcomes imply that part of the detrimental labor market effects of 9/11 can
be offset by intermarriage. The offsetting effect may not be particularly large, given
the estimated effect of 9/11 on intermarriage being about 2%. Nevertheless, this
result highlights the possible importance of the labor market incentives that are
associated with the increased rates of intermarriage among Hispanics.
6.2 Does discrimination exist?
One of our goals is to examine whether discrimination actually existed and
increased after 9/11 in marriage markets. However, our baseline result for Hispanic
immigrants suggests either that discrimination may not have increased in marriage
markets at all after 9/11, or that the discrimination effect existed after 9/11, but the
effects of increased discrimination were small relative to the positive effects
Table 5 Difference-in-differences (DD) estimates of the effect of 9/11 on employment status and
earnings
Employment Log earnings
(1) (2)
Ethnic group -0.007 -0.134**
(0.007) (0.061)
Post-911 -0.013 0.264**
(0.012) (0.107)
DD estimates -0.010 -0.120*
(0.008) (0.071)
Intermmariage 0.012*** 0.218***
(0.004) (0.040)
Time fixed effects Yes Yes
Control variables Yes Yes
State fixed effects Yes Yes
No. of obs. 77,558 77,558
Columns (1)–(2) report the effect of 9/11 on employment (whether one is currently employed) among
Hispanic immigrants, while Columns (3)–(4) report the effect of 9/11 on log earnings; Control variables
in the models include gender, years of schooling, age, age squared and years in the states. Both columns
(2) and (4) also control for immigrants’ intermarriage outcome (whether one is married to a native)
Robust standard errors (clustered at immigrant group) in brackets. *** p \ 0.01; ** p \ 0.05; * p \ 0.1
The effects of 9/11 on intermarriage 187
123
associated with potential economic gains. The gender difference in the effects of
9/11 on intermarriage and the results using intermarriage with only non-Hispanic
natives seem to imply the existence of discrimination in marriage markets.
To further examine the existence of discrimination in the marriage market, we
look at the effect of 9/11 on intermarriage among Muslims immigrants. Changes in
attitudes toward immigrants after the 9/11 tragedy may be targeted more at Muslims
than at Hispanics. As such, one may expect that the negative supply-side effect of
9/11 may be stronger for Muslims, and that the overall effect of 9/11 on Muslim
immigrants may be much smaller than that for Hispanic immigrants or even negative.
Table 6 displays the results. Compared to the results for Hispanics, a very different
picture emerges here. The estimated effect of 9/11 on intermarriage among Muslim
immigrants is close to zero and is not statistically significant at any conventional
level.15 In light of our discussion above, the estimates suggest that even though 9/11
may have increased the incentives for Muslim immigrants to marry natives, this
positive effect on intermarriage outcomes is however completely offset by the
negative effect due to discrimination. One may argue that this could also imply there
exists neither positive nor negative effects, instead of the effects offsetting each
other. To further assess this issue, we also restrict our sample to those states that were
most directly impacted by 9/11 and thus potentially more discriminating against
immigrants, i.e. New York, DC, and Massachusetts. The effect of 9/11 turns
negative, and the magnitude (column (2), Table 6) is much larger, although
statistically insignificant due to dramatic declines in sample size. In particular, the
Table 6 Difference-in-differences (DD) estimates of the effect of 9/11 on intermarriage among Muslims
(1) (2)
Full sample States with stronger discrimination
Muslim -0.119*** -0.058**
(0.014) -0.027
Post-911 -0.113*** -0.114***
(0.016) -0.038
DD estimates 0.002 -0.038
(0.017) -0.035
Time fixed effects Yes Yes
Control variables Yes Yes
State fixed effects Yes Yes
No. of obs. 28,211 3,637
Control variables in the models include gender, years of schooling, age, age squared and years in the
states
Column (1) reports the baseline result for the full sample of Muslims. Column (2) restricts the sample to
the states with potentially stronger discrimination. These states are New York, DC, and Massachusetts
Robust standard errors (clustered at immigrant group) in brackets. *** p \ 0.01; ** p \ 0.05; * p \ 0.1
15 The imprecisely estimated effect could be due to the problem of small portion of intermarriages among
Muslim immigrants. The standard errors are indeed larger for the sample of Muslim immigrants.
However, even assuming the same standard errors as for Hispanic immigrants, the estimates do not obtain
the same level of statistical significance as with the estimates for Hispanic immigrants.
188 C. Wang, L. Wang
123
effect is now -3.8 percentage points! This result is consistent with our hypothesis
that discrimination against Muslim immigrants indeed exists in the intermarriage
market, and that it is even larger in New York, DC, and Massachusetts.
While all these discussions do not provide a direct proof of the existence of
discrimination in marriage markets, the evidence is suggestive of the fact that
discrimination is an important phenomenon in marriage markets.
6.3 Alternative explanations
In our discussions above, we focus on discrimination as the only source of the
negative impact of 9/11 on intermarriage (i.e.natives are less willing to marry
immigrants). However, it is possible that 9/11 increased tensions between natives
and immigrants for both sides. As a result, immigrants may also be less willing to
marry natives, thereby decreasing the number of immigrants participating in the
intermarriage market; as with discrimination, this effect could also offset the
incentives to marry a native, giving rise to the observed patterns above. Although
the conclusions – that economic gains from intermarriage outweigh or offset the
negative effect of 9/11—remain unchanged, the extent of discrimination implied by
our results may be exaggerated. For example, we do not observe any statistically
significant impact of 9/11 on intermarriage for Muslim immigrants. Given
potentially large gains from intermarriage for immigrants, this result implies that
the positive effect of 9/11 due to increased demand for intermarriage is completely
offset by the negative impact. Without considering other sources of negative impact,
we may attribute this result only to discrimination.
One way to distinguish these two mechanisms may be to test the effect of 9/11 on
the treatments (or the price) received by natives in the marriages. Note that although
both mechanisms have similar impacts on the number of intermarriages, they have
different implications for the gains received by native spouses in the marriages,
holding everything else constant. On the one hand, reduced willingness to
participate in the intermarriage market among immigrants (i.e. reduced demand
for intermarriage) would decrease the gains from intermarriage for natives. On the
other hand, reduced willingness to participate in the intermarriage market among
natives (i.e. reduced supply for intermarriage) would increase the gains from
intermarriage relative to alternatives for natives. As a measure of gains from
intermarriage for natives, the treatments received by native spouses within the
households (e.g. domestic violence) may allow us to test the relative importance of
alternative mechanisms. Grossbard et al. (2010) use an innovative measure—time
devoted to household chores—to measure racial discrimination within a marriage.
Such information is unfortunately not available in our data. Grossbard-Shechtman
and Fu (2002) propose that women’s labor supply can be used as an indicator of
how well women are treated in a marriage, since reduced gains received within
marriages may decrease reservation wages for women and thus increase women’s
labor supply. However, this measure is not applicable in the 9/11 context, since, as
discussed above, the existing studies have found that 9/11 has affected the labor
market outcomes of immigrants. It would be impossible to disentangle the changes
in labor supply caused by 9/11 from those caused by intermarriage.
The effects of 9/11 on intermarriage 189
123
Some alternative ways may be to look at the effect of 9/11 on certain spouse
characteristics that people generally value, e.g. spouse’s education and income. For
example, if people prefer more educated spouses, then spouse education might also
be considered as a type of ‘‘price’’ of marriage. Everything else being equal, when
immigrants have a larger incentive to marry a native than an immigrant, they may
pay a higher ‘‘price’’ to a native spouse (e.g. lower level of spouse education). We
believe that a systematic and comprehensive examination of alternative mechanisms
is important in its own right, and we thus leave this for future research.
7 Conclusions
In this paper, we attempt to assess how the 9/11 tragic event affected immigrants’
probability of marrying a native. We find a positive effect of 9/11 on intermarriage
among Hispanic immigrants. Hispanic immigrants are about 2.2 percentage points
more likely to marry natives after 9/11, using European, Canadian and other
immigrants as the control group. The magnitude of the effect is much smaller in the
first two years after 9/11 and becomes larger after that. There also exists a large,
statistically significant gender difference in intermarriage outcomes. In fact, the
large positive effect of 9/11 on Hispanic immigrants’ intermarriage rates seems to
be driven solely by males. Our results suggest that 9/11 changed the attitudes and
preferences toward intermarriage among both immigrants and natives. Further
robustness checks imply that our results are not driven by undocumented
immigrants alone, Hispanic immigrants marrying Hispanic natives, immigrants
who were married before arrival, or new immigrants. Finally, we conduct indirect
tests of whether there are economic gains from intermarriage and of whether
discrimination indeed exists. Our results suggest that they do.
While these findings are of interest, they are reduced-form estimates that are at
best suggestive of underlying mechanisms of how 9/11 affected the intermarriage
outcomes and immigrants’ behavior in general. Future research to further
investigate these mechanisms is warranted for a deeper understanding of
immigrants’ intermarriage decisions. Moreover, should panel data be available, it
would also be of interest for future studies to examine how 9/11 affected the number
of new intermarriages and the divorce rate as a function of intermarriage.
Acknowledgments The authors would like to thank Shoshana Grossbard (the editor), a knowledgeable
and helpful referee, John Hurdelbrink, and session participants at the SEA conference in Atlanta, GA for
helpful comments, suggestions, and discussions of this research.
References
Anderson, R. N., & Saenz, R. (1994). Structural determinants of Mexican American intermarriage,
1975–1980. Social Science Quarterly, 75, 414–430.
Baker, M., & Benjamin, D. (1997). The role of the family in immigrants’ labor-market activity: an
evaluation of alternative explanations. American Economic Review, 87, 705–727.
Becker, G. (1993). A treatise on the family, enlarged edition. Cambridge, MA: Harvard University Press.
Becker, G. S. (1974). A theory of marriage: Part II. Journal of Political Economy, 82(2), s11–s26.
190 C. Wang, L. Wang
123
Bertrand, M., Duflo, E., & Mullaninathan, S. (2004). How much should we trust differences-in-
differences estimates? Quarterly Journal of Economics, 119, 249–275.
Cameron, A. C., Gelbach, J. B., & Miller, D. L. (2008). Bootstrap-based improvements for inference with
clustered errors. Review of Economics and Statistics, 90, 414–427.
Celikaksoy, A., Nielsen, H. S., & Verner, M. (2006). Marriage migration, just another case of positive
assortative matching? Review of Economics of the Household, 4, 253–275.
Chiswick, B. R., & Houseworth, C. A. (2011). Ethnic intermarriage among immigrants: Human capital
and assortative mating. Review of Economics of the Household, 9, 149–180.
Fisman, R., Iyengar, S. S., Kamenica, E., & Simonson, I. (2008). Racial preferences in dating. Review ofEconomic Studies, 75, 117–132.
Fossett, M. A., & Kiecolt, K. J. (1991). A methodological review of the sex ratio: Alternatives for
comparative research. Journal of Marriage and the Family, 53, 941–957.
Furtado, D. (in press). Human capital and interethnic marriage decisions. Economic Inquiry. doi:
10.1111/j.1465-7295.2010.00345.x.
Furtado, D., & Theodoropoulos, N. (2009). I’ll marry you if you get me a job: Marital assimilation and
immigrant employment rates. International Journal of Manpower, 30, 116–126.
Furtado, D., & Theodoropoulos, N. (2010). Why does intermarriage increase immigrant employment?
The role of networks. The B.E. Journal of Economic Analysis & Policy (Topics), 10, 1–31.
Georgarakos, D., & Tatsiramos, K. (2009). Immigrant self-employment: Does intermarriage matter?
Research in Labor Economics, 29, 253–271.
Grossbard, S. A. (1983). A market approach to intermarriage. Technical report, Avraham Harman
Institute of Contemporary Jewry.
Grossbard, S., Gimenez-Nadal, J. I., & Molina, J. A. (2010). Racial discrimination and household chores.
IZA discussion papers, 5345.
Grossbard-Shechtman, S. A. (1993). On the economics of marriage. Boulder: Westview Press.
Grossbard-Shechtman, S. A., & Fu, X. (2002). Women’s labor force participation and status exchange in
intermarriage: A model and evidence for Hawaii. Journal of Bio-economics, 4, 241–268.
Henderson, D. J., Polachek, S. W., & Wang, L. (2010). Heterogeneity in schooling rates of return.
Unpublished Manuscript.
Kalmijn, M. (1998). Intermarriage and homogamy: Causes, patterns, trends. Annual Review of Sociology,24, 395–421.
Kalmijn, M., & Van Tubergen, F. (2010). A comparative perspective on intermarriage: Explaining
differences among national-origin groups in the United States. Demography, 47, 459–479.
Kantarevic, J. (2004). Interethnic marriages and economic assimilation of immigrants. IZA Discussion
Papers, 1142.
Kaushal, N., Kaestner, R., & Reimers, C. (2007). Labor market effects of September 11th on Arab and
Muslim residents of the United States. Journal of Human Resources, 42, 275–308.
King, M., Ruggles, S., Trent Alexander, J., Flood, S., Genadek, K., Schroeder, M. B., Trampe, B., &
Vick, R. (2010). Integrated public use microdata series, current population survey: Version 3.0.[Machine-readable database]. Minneapolis: University of Minnesota.
Lemieux, T., & Card, D. (2001). Education, earnings, and the ‘Canadian G.I. Bill’. Canadian Journal ofEconomics, 34, 313–344.
Lichter, D. T., Brown, J. B., Qian, Z.-C., & Carmalt, J. (2007). Marital assimilation among hispanics:
evidence of declining cultural and economic assimilation? Social Science Quarterly, 88, 745–765.
Lichter, D. T., Mclaughlin, D. K., Kephart, G., & Landry, D. J. (1992). Race and the retreat from
marriage: A shortage of marriageable men? American Sociological Review, 57, 781–799.
Mehta, C., Nik, T., & Hincapie, M. (2003). Social Security Administration’s no-match letter program:
Implications for immigration enforcement and workers’ rights. Technical report, Center for Urban
Economic Development, University of Illinois at Chicago.
Meng, X., & Gregory, R. G. (2005). Intermarriage and the economic assimilation of immigrants. Journalof Labor Economics, 23, 135–176.
Meng, X., & Meurs, D. (2006). Intermarriage, language, and economic assimilation process: A casestudy of France. IZA Discussion Paper, 2461.
Orrenius, P. M., & Zavodny, M. (2009). The effects of tougher enforcement on the job prospects of recent
latin American immigrants. Journal of Policy Analysis and Management, 28, 239–257.
Passel, J. (2006). The size and characteristics of the unauthorized migrant population in the U.S. Pew
Hispanic Center Research Report.
The effects of 9/11 on intermarriage 191
123
Passel, J., & Cohn, D. (2009). A portrait of unauthorized immigrants in the United States. Pew Hispanic
Center Research Report.
Qian, Z.-C., & Lichter, D. T. (2007). Social boundaries and marital assimilation: Evaluating trends in
racial and ethnic intermarriage. American Sociological Review, 72, 68–94.
Roehling, P. V., Jarivis, L. H., & Swope, H. E. (2005). Variations in negative work-family spillover
among white, black, and hispanic American men and women. Journal of Family Issues, 26,
840–865.
Sheridan, M. B. (2002). Records checks displace workers: Social security letters cost immigrants jobs.
Washington Post, August 6.
Vaughan, J., & Edwards, J. R. (2009). The 287(g) program: Protecting home towns and homeland.
Technical report, center for immigration studies.
Wood, R. G. (1995). Marriage rates and marriageable men: A test of the Wilson hypothesis. Journal ofHuman Resources, 30, 163–193.
192 C. Wang, L. Wang
123
top related