Cohabitation among Chinese minorities in Western countries: an alternative family formation strategy for disadvantaged groups?

Using data of Chinese immigrants extracted from the Belgian National Register registering a cohabitation or marriage between 2005 and 2015 (N = 3945), this study aims to advance the understanding of cohabitation among disadvantaged groups in local marriage markets. From the difference in prevalence of cohabitation between Chinese men and women, we conclude that Chinese immigrants could use cohabitation rather as a practical living arrangement in the context of recent migration rather than as an alternative to marriage, as the receiving society does. The practical living strategy of ethnic minorities enhances our understanding of cohabitation among disadvantaged groups, for whom the choice for cohabiting would be largely influenced by the home familial culture.


Introduction
Although cohabitation has become prevalent globally in recent decades, the significance and meanings of cohabitation in the family system vary substantially across regions and ethnic groups (Yu & Xie, 2015). In Western countries, cohabitation along with family transitions has been given much attention in research over decades (Kalmijn, 2007;Soons & Kalmijn, 2009), such as decreased rates of marriage or remarriage, delayed age of first marriage, and low fertility and mortality. These family transitions have been described as a new stage of demographic development, or 1 3 the Second Demographic Transition (SDT) ( Van de Kaa, 1987) (Kaa, 2002). As one of the signature elements of SDT, cohabitation has been increasingly regarded as a full alternative to marriage for many Europeans (Heuveline & Timberlake, 2004). Though the focus has gradually shifted to other groups in Europe, such as ethnic minorities or immigrants, relevant information about these groups is quite limited.
As the disadvantaged groups in the receiving countries' marriage markets, immigrants received relatively insufficient focus on their family formation strategy, especially their cohabitations (Carol et al., 2014;Kogan & Weißmann, 2019). Previous cross-national studies about cohabitation mostly compare citizens from different nationalities (Heuveline & Timberlake, 2004) and focus on socioeconomic levels, institutional support or specific characteristics of each country. Little attention has been paid to minority groups' cohabitation in migration, though their cohabitation prevalence relates to the assimilation issue.
Moreover, existing studies mainly focus on the cohabitation of disadvantaged groups from the socioeconomic perspectives (Lesthaeghe, 2020;Ramm & Salinas, 2019), which shows that the low incoming or low education could be the significant reasons leading to people's cohabitation. However, the disadvantaged people include deprived groups in a socioeconomic sense and marginalised groups in the sense of culture and race, such as ethnic minorities or immigrants (Bumpass & Lu, 2000;Ramm & Salinas, 2019). In that light, cultural differences might be another factor that results in people's disadvantaged status. Therefore, this study fills the gap of less information on disadvantaged groups' cohabitation by focusing on the cohabiting behavioural changes of a particular group of individuals, specifically, Chinese immigrants in Belgium.
Although recent studies discussed the increasing prevalence and broader acceptance of cohabitation in China, the data remain limited on cohabitation (Yu & Xie, 2015). The lower popularity of cohabitation in China compared with the Western countries can be explained by the traditional Confucian culture that still shapes the conservative attitude to cohabitation in China (Zhang, 2017). Even in Japan, another Confucian non-western country characterized by other typical SDT features (such as the lowest-low fertility and very late age of marriage), available data show that nonmarital cohabitation remains fairly at a low level (Raymo et al., 2009).
Studying cohabitation among disadvantaged groups, such as immigrants, presents valuable opportunities for examining the significance and meaning of cohabitation in subgroups regarding ethnicity and disadvantaged groups. This study enriches the diversity of cohabitation dynamics and attitudes to family formation changes by race and gender. Moreover, with the limited comparison between gender in the previous studies, our study provides an important opportunity to evaluate the generality of observed subpopulation differences in the significance of cohabitation (Forste, 2001). In specific, concentrating on cohabitation status among Chinese immigrants in Belgium presents a valuable opportunity to examine the prevalence of cohabitation for individuals with a Confucian cultural background but living in western society. It is helpful for understanding the family formation or origin cultural frameworks changes for immigrants in the receiving society.
This study aims to advance our understanding of premarital cohabitation for immigrants and Chinese people by firstly describing and explaining the prevalence of cohabitation among Chinese immigrants in Belgium. We use data extracted from the Belgian National Register and select Chinese origin residents (based on their birth nationality) who married or started cohabiting between 2005 and 2015. By describing cohabitation trends and comparing them with relevant trends in China and among Belgian natives in Belgium, we expect that the Chinese immigrants' cohabitation rate will be higher than the rate in China and closer to the native Belgians' rate. Subsequently, we apply binomial logistic regression analyses to examine the determinants of cohabitation.

Theoretical frameworks of cohabitation and immigration:
the second demographic transition (SDT) and the pattern of disadvantage (POD)

The second demographic transition (SDT)
The prevalence of cohabitation among immigrants is closely related to that in the host societies. However, globally, the prevalence of cohabitation varies across regions or depends on different development stages in regions. Since World War II, cohabitation has been increasingly accepted in Western countries (Heuveline & Timberlake, 2004). The United States holds a cohabitation prevalence estimated from 50% (2000) to over 60% (2004) (Stanley et al., 2006). In Europe, France has a high prevalence (80%), as well as Sweden, and they consider cohabitation to be an alternative to marriage. In more traditional European countries, the prevalence of cohabitation is lower, such as Italy (10%) and Spain. As for Belgium, cohabitation was regarded as a prelude to marriage for a long time and increasingly became accepted as a full alternative to marriage (Heuveline & Timberlake, 2004). A government report (2017) showed that, on an individual level, 58% of adults had been cohabiting with a partner since 2000. Meanwhile, on a household level, the percentage of cohabitating households was around 54% in Belgium between 2000 and 2016 (Corijn, 2017). These regional differences in cohabitation between countries stem from different demographic development stages in countries, in line with the second demographic transition (SDT) (Van de Kaa 1987), in which cohabitation is central. The SDT proposes that demographic patterns in modern societies evolve towards below-replacement levels of fertility, delayed age at first marriage, rising divorce rates, tolerance towards non-marital fertility, and increasing prevalence of cohabitation (Lesthaeghe, 2010). Additionally, along with the shift in values towards individual self-realization, relations between man and woman are increasingly seen as the means of reciprocal emotional enrichment, regardless of marrying or cohabiting (Kaa, 2002).
The transitions mentioned above led to a decreased population size in Europe, and this new demographic disequilibrium needed to be complemented by immigration (Kaa, 2002). Western European countries resolved the labour shortage by recruiting guest workers from countries like Turkey and Morocco (Dupont et al., 2017). Immigration has become a determinant of population growth and helped stabilize population size in Europe (Kaa, 2002). However, it brought other challenges, such as integrating immigrants into the host society (Lesthaeghe, 2014).
To understand adaptation processes in Asian immigrant groups, we studied cohabitation among Chinese immigrants in Belgium. We firstly assess whether Chinese immigrants have comparable cohabitation rates as those in the mainstream population. As SDT claims, immigrant groups' demographic characteristics should ultimately become similar to those in the host country (Lesthaeghe, 2014). Secondly, we examine whether Chinese immigrants show openness towards interethnic partnerships. If this is the case, this could indicate assimilation towards family formation dynamics among Chinese in the host country. As Gordon stated in his classical assimilation model, marital assimilation is a crucial stage for ethnic minorities for overall acceptance into the mainstream society because assimilation in other dimensions will follow naturally (Gordon, 1961). The consequences then are the erosion of social and cultural boundaries between ethnic groups, the disappearance of an ethnic group as a separate entity, and the evaporation of its distinctive values (Gordon, 1961). However, critics argued that adapting does not necessarily imply the loss of ethnic group identity. In contrast, their racial awareness might be heightened in the process of direct contact with the mainstream culture (Song, 2009). Additionally, the theory of multiculturalism promotes the idea that group identity and unchanged ethnocultural relations remain and can be passed on from generation to generation (Pianko, 2008). Our research then has the ambition to gain insight into the integration of Chinese immigrants into the host by studying interethnic partnerships and providing in-depth information on the motives why Chinese immigrants could choose cohabitation rather than marriage.

Pattern of disadvantage (POD)
Besides an outcome of cultural revolution or as being driven by higher order needs as stated by the SDT, cohabitation can also start from a POD basis among the deprived groups (Lesthaeghe, 2020). The cohabitation studies in the US have supported this view, where the cohabitation are more common among the poorer classes (Perelli-Harris & Gerber, 2011). According to the POD approach, the rise of cohabitation relates to the socioeconomic stratification that a worsening of the living conditions of the poorer segment, and the cohabitation can be an arrangement of economic necessity (Perelli-Harris & Gerber, 2011;Ramm & Salinas, 2019).
Linking patterns of family formation to socioeconomic status has a long tradition among scholars on kinship and families (Ramm & Salinas, 2019). As Lesthaeghe (2020) suggested, "on a global scale, the rise of cohabitation is strongly conditioned by the contrasting historical patterns of kinship organisation, including the positions of women" (Lesthaeghe, 2020). Therefore, our study focusses on the cohabitation among Chinese immigrants in Belgium and contrast their cohabiting pattern with the native Chinese.
Chinese immigrants to Belgium mainly rely on kinship relationships to find their way in Belgian society (Pang, 1998), which stimulates the maintenance of strong bonds with their origin nations' social networks and cultural backgrounds. The first group of Chinese to arrive in Belgium consisted mainly of single seamen sailors or workers who jumped from ships to seek a better life (Federaal Migratiecentrum, 2015). The majority of them originated from southeastern China, mainly from Guangdong, Zhejiang, and Fujian provinces. Afterwards, many seamen paved the way for their families or associates of families to migrate to Belgium (Pang, 1998). Accordingly, the Chinese, who value kinship highly, took kinship as the organizing principle to institutionalize their lives and communities abroad (Baker, 1979). This can also be seen in the predominance of a few surnames in overseas Chinese communities (Baker, 1979). The strong kinship connections in Chinese communities and the traditional rigid values on cohabitation behaviour may still influence Chinese immigrants in Belgium. Since the specific cultural backgrounds could influence immigrants' cohabitation, we then discuss it in the next section.

Cohabitation in China
Historically, premarital sex was prohibited due to the rigid sexually suppressive doctrine from dominant Confucianism in traditional China (Ng & Lau, 1990). This social norm had remained unchanged for thousands of years until the Economic Reform and Opening-up in 1978 when booming economics led to dramatic social transformations (Xiao et al., 2011). Cohabitation increasingly became acceptable due to three social reforms: economic expansion, institutional transformation, and ideological change (Yu & Xie, 2015;Zhang, 2017).
First, economic development accelerates the dissolution of traditional family systems and increases opportunities for cohabitation. On the one hand, with China's urbanization process and the unbalanced economic developments between regions, young people increasingly tended to leave poor hometowns for modern cities in pursuit of a better life. This trend impeded the traditionally significant and direct influence of parents on their children's family formation. Without parents' direct intervention, there was more freedom in the choice of partner and living arrangement for young people. Combined with higher divorce rates, premarital cohabitation has become a more socially accepted form of living arrangement for young couples before formal marriage in China (Xie, 2013a). On the other hand, housing prices sharply increased along with the economy booming. High housing prices made it more difficult for people to marry since the groom is expected to provide housing for the new family in China (Xie, 2013a). High housing also prices led to a huge financial burden for young couples planning to get married. Therefore, to avoid excessive debts from marriage, young couples increasingly preferred cohabitation.
Apart from economic considerations, institutional, ideological, and educational factors further stimulated cohabitation in China. In 1950, the Marriage Law formally legalized the free choice of partners and promoted equal gender rights and interests between partners (Xie, 2013a), which improved women's position in the household. Besides, attaining universal primary education provided an equal chance for men and women to learn the skills necessary for the labour market, creating more opportunities for women's economic independence. Moreover, Chinese people were increasingly exposed to information from Western societies, emphasizing self-fulfilment, individualization, choice of a romantic partner, and cohabitation as a widely accepted alternative to marriage, resulting in increased acceptance and spreading of cohabitation in China (Burnett, 2010;Ma & Rizzi, 2017).
Research on the topic of cohabitation in China is emerging but is hampered by the lack of data. Yu and Xie's (2015) study on cohabitation provided the first national estimates of cohabitation rates in China by analyzing recently released data from the 2010 and 2012 waves of the China Family Panel Studies (CFPS). This nationally representative longitudinal survey was launched in 2010 and provided information on 42,590 individuals. The follow-up survey in 2012 included the respondents from the first wave and their new family members (44,693). The 2012 survey indicated that the general cohabitation rates in China were 8.1% for males and 6.2% for females, while the rate of cohabitation was much higher in regions with a developed economy and modern culture than in less-developed regions with more traditional ideologies. Although they claim that the prevalence of cohabitation in China is rising, the rate is still lower than in most industrialized countries. Zhang (2017) explained why the cohabitation rate in contemporary China was lower than in Western societies by underlining the significance of traditional Confucian culture. Such culture involves strong social pressure that forces cohabitators into marriage. In other words, because premarital cohabitation and divorce are not widely accepted due to Confucian cultural norms, cohabitators are pushed into marriage eventually (Zhang, 2017). In western societies, however, this is not the case.

Correlates of cohabitation
This section discusses the main determinants of cohabitation, more specifically, gender, education, generation, employment status, and partner type. We especially pay attention to these elements that can help explain the differences in cohabitation rates between Chinese immigrants and the majority population on the one hand and differences in gender, on the other hand. By comparison, we try to understand the changes and influence of cohabitation for immigrants and ethnic minorities.
The increasing prevalence of cohabitation is a global trend. As shown in a demographic survey in the United States in 2019, a majority (61%) of adults aged 18 to 44 have cohabitation experience (Horowitz et al., 2019). Similar increases are also documented in Europe, where 80% of men in Northern Europe cohabited with a partner before marriage (Kalmijn & Vermunt, 2007). In Asian countries, like China and Japan, cohabitation has become increasingly popular (Raymo et al., 2009;Zhang, 2017). In Belgium, the cohabitation rates steadily increased after the Supreme Court confirmed the legal status of cohabitation in 1990 (Swennen & Mortelmans, 2015) (see Fig. 1).
From an assimilation perspective, it could be expected that the prevalence of cohabitation would increase towards prevalence among the majority population. Furthermore, we could also expect that cohabitation prevalence will be higher for ethnically mixed couples than for homogamous Chinese partnerships. As Gordon stated, overcoming structural barriers is necessary for assimilation (Gordon, 1961) and intermarriage is a decisive way to overcome cultural, social, and physical boundaries (Lieberson & Waters, 1988). Cohabitation, furthermore, has the advantage over marriage for mixed partnerships that it can avoid social resistance towards interethnic marriages by keeping interethnic cohabitation partnerships (Qian & Lichter, 2007). In line with this, we expect that the prevalence of cohabitation among Chinese immigrants in Belgium is increasing.
Although gender equality has improved notably in recent decades, being unmarried is a status that is less consistent with females than males in some cultures (Stavrova et al., 2012). Men are more likely to choose cohabitation than women. Prior research provides evidence that men consider cohabitation more as a preliminary relationship test than women (Stavrova et al., 2012). This claim is also supported by Yu and Xie's (2015) research. They reveal that the cohabitation rate of Chinese men (8.1%) is higher than that of women (6.2%). Thus, we expect that Chinese male immigrants choose cohabitation more often than female immigrants (Hypothesis 1).
Another recent Chinese study noted that education level is positively associated with cohabitation (Xie, 2013b). After being restricted by traditional Confucian female norms for thousands of years, women's social position in recent years fundamentally changed along with improvements in their education and their financial independence (Sassler & Lichter, 2020). Cohabitation, then, has increased among the general public, which provides more freedom for choosing an alternative partnership to marriage. These ideational changes can be ascribed to increased educational attainment and more egalitarian gender status in society (Qian, 2012). We thus expect that people who have higher education levels are more likely to choose cohabitation than lower educated individuals (Hypothesis 2).
Financial concerns, moreover, are additional elements in traditional family systems, as marriage is only considered a viable option when partners have sufficient financial means (Smock et al., 2005). According to the pattern of disadvantage (POD), it links cohabitation and non-marital childbearing to socioeconomic stratification and reveals that cohabitation is more common among deprived groups in the US (Ramm & Salinas, 2019), because cohabitation serves as an adaptive family formation strategy by allowing union formation despite economic uncertainty (Brown et al., 2008). Similar findings from other studies claim that economic security is a necessary condition to enter a marriage (Kalmijn, 2007). The financial status of a couple is considered to be an "economic package" that includes financial stability, the absence of debts and often also the male partner's ability to support a family (Smock et al., 2005). Therefore, unemployment could increase the likelihood of cohabitation in Europe and East Asia, as cohabitation implies less financial burden than marriage (Sassler & Lichter, 2020). And it is reasonable to believe that cohabitation may be an attractive alternative to marriage for couples with fewer financial resources and guarantees, owing to fewer financial burdens and obligations, as confirmed by Raymo et al. (Raymo et al., 2009). Accordingly, income could be an additional element in favour of unmarried cohabitation. However, because we lack information on income in our data (see further), we use employment status as a proxy for financial stability. We hypothesis that those who do not have a job or are in an unstable financial situation are more likely to cohabit than people with stable incomes (Hypothesis 3).
Previous studies have proved that cohabitation levels increase with each generation for all racial and immigrant groups, in line with the assimilation theory (Brown et al., 2008). For example, Alba and Nee (2003) stated that race or ethnicity is a socially constructed boundary that can be "crossed," "blurred," or "shifted" over time or across generations (Alba & Nee, 2003). Based on this, we expect that generational succession may be distinctive among Asians because the first generation was socialized in a traditional cultural context with very low cohabitation rates. In contrast, the second generation was socialized in Belgium, where premarital cohabiting is highly accepted. Furthermore, other studies suggest that the second generation could delay family formation in pursuit of selfachievement and upward social mobility, also resulting in a lower marriage rate among the second generation than the first generation (Brown et al., 2008). Given all that, we expect the second generation to have a higher cohabitation rate than the first generation (Hypothesis 4).
Another strand of research, moreover, indicates that cohabitation could be more prevalent among heterogeneous ethnic couples than among ethnic homogamous couples (Kasearu & Kutsar, 2011;Landale, 2008). Especially among ethnicities with a large social resistance to interethnic relationships, as is documented among Muslim immigrants (Lievens, 1998), interethnic couples may choose cohabitation rather than marriage. With a slowly increasing acceptance of premarital relationships by Muslim immigrants (Van Kerckem et al., 2013), cohabitation could then be regarded as an alternative to marriage for mixed-relationship couples to avoid religious resistance. As for Chinese immigrants, although they do not have religious restrictions on intermarriage, social resistance against intermarriage could be present in the majority population (Tolsma et al., 2008). Additionally, it has been shown that interethnic marriages have higher divorce rates than homogamous marriages. Hence, cohabitation might be considered an alternative to intermarriage since the cohabitation's dissolution procedure is much easier than divorce from marriage. For all these reasons, we expect that Chinese immigrants with mixed partnerships are more likely to choose cohabitation (Hypothesis 5).

3
Cohabitation among Chinese minorities in Western countries:…

Data
In this study, we analyze data from an extraction 1 of the Belgian National Register, including all individuals residing in Belgium who married or started cohabiting between 2005 and 2015 and met the following conditions: (1) being resident in Belgium at least one year before marriage or start of cohabitation, and (2) having a third-country nationality at birth. From these data, we selected residents in Belgium of Chinese origin based on their birth nationality (N = 3945). This dataset contains both marriages and legally registered cohabitations. A drawback of using this data is that it does not include information on unregistered cohabitations and persons who are not recorded in the Belgian National Register, such as refugees, asylum seekers, and people residing illegally in Belgium. The major advantage is that it provides information on the full population who married or began cohabitation within that time frame.

Method
For understanding the role and meaning of cohabitation for immigrants, this study explores (1) the prevalence of cohabitation among the Chinese minority in Belgium, and (2) the factors that affect the choice for cohabitation over marriage. To answer the two questions, we start with a descriptive analysis by focusing on the trends in the cohabitation rate between 2005 and 2015. As we expect differences between genders, the trends are shown separately for men and women. Subsequently, we perform a descriptive bivariate analysis of each determinant of cohabitation considered here.
Finally, we assess the effects on cohabitation by estimating binomial logistic regression models in which the log-odds of choosing cohabitation (vs marriage) are the dependent variable. Educational Attainment is operationalized as a four-category variable: No diploma or only having a diploma of lower education (generally went to school until the age of 12 years), lower and upper secondary education (until the age of 18 years), higher education level (tertiary education diploma), and a separate category for missing information on educational attainment (these are primarily first-generation immigrants who acquired their diploma abroad). As a measure for socioeconomic status, we use the variable of Employment Status to imply financial stability. We distinguish between students, those who do not have a job, those who work, and a separate category for those with missing information. Migrant Generation is operationalized as a three-category variable: second generation (born in Belgium or came to Belgium before the age of 6), middle generation (came to Belgium between the ages of 6 and 14), and first-generation (came to Belgium after the age of 14). Partnership type is a three-category variable: mixed partnership, local coethnic partnership, and transnational partnership. A mixed partnership is a relationship of someone of Chinese descent living in Belgium with a partner of another birth nationality (mostly Belgians). The local co-ethnic partnership is a relationship between two Chinese descent partners who start cohabitation after residing in Belgium. A transnational partnership is a relationship of someone of Chinese descent living in Belgium with a partner who lived in China before the official start of the relationship (who migrated to Belgium after or through cohabitation or marriage). Finally, we use the year of cohabitation or marriage (from 2005 to 2015) as covariant to control for possible period effects.

Results
This section firstly presents descriptive findings of the prevalence and determinants of premarital cohabitation among Chinese immigrants in Belgium. We then show results from binomial logistic regression, examining the effects of the determinants of choice for cohabitation (vs marriage).
We observe a steadily increasing prevalence of cohabitation, both for men and women, between 2005 and 2015 (Fig. 2). Correspondingly, the marriage rate gradually declines during this period, although marriage remains the primary relationship type. The lowest marriage percentage is in 2015 and lies around 70%, which is over two times more than the highest percentage of cohabitation (30%) in 2015. Although the rate of marriage is higher than the rate of cohabitation (Table 1), we notice that cohabitation seems to be more popular among Chinese women than Chinese men. Figure 2 shows that the percentage of cohabitation among women increased from about 12% in 2005 to nearly 40% in 2015, which is higher than the average cohabiting rate and the male cohabitation rate. This is also visible in Table 1, where the general percentage of cohabitation among women is 25.9%, which is more than that of men (16.9%). Furthermore, Table 1 compares three types of partnerships. As can be seen in the table, more than half (52.9%) of those in mixed partnerships choose cohabitation, followed by the local co-ethnic partnerships (41.4%). Only very few of those in transnational partnerships choose (5.7%) cohabitation.
To explore our second research question, we start with an exploratory bivariate analysis, investigating the degree to which the choice of cohabitation (vs marriage) is associated with the migrant generation, employment status, educational level, and partnership type. Since associations might differ between men and women, we report the results separately by gender. Generally, we found that migrant generation ( x 2 =10.985, p < 0.05), education ( x 2 =27.482, p < 0.05), and partnership type ( x 2 = 364.502, p < 0.05) are significantly associated with the choice to cohabitation. Table 2 provides the summary statistics for the determinants of cohabitation. Firstly, as we already discussed earlier, it is clear that cohabitation is more prevalent among women than among men in all three migration generations. We expected that the second generation would opt for cohabitation more than other generations, in line with the expectations derived from assimilation theory. The results mainly confirm our hypotheses. For male immigrants, around 25.9% of 1 3  Trends of the prevalence of marriage and cohabitation by sex second-generation men choose cohabitation, whereas only 15% of the middle generation male immigrants and 16.2% of the first-generation male immigrants choose cohabitation. Among women, the highest percentage of cohabitation is also found in the second generation (32.6%). Although this difference is not statistically significant, it is in line with our expectations. For middle and firstgeneration women, it is 27.4% and 25.2%, respectively. Although the difference between generations is not statistically significant for women, it is in line with our expectations.
We hypothesized that Chinese immigrants with higher education levels would choose cohabitation more than lower-educated immigrants. This expectation is only confirmed among women. The percentage of women with a high educational level who choose cohabitation is 30%, compared with 19.3% of highly educated men. By contrast, Chinese male immigrants with no or a low education level choose cohabitation more than women with the same education level (30.5 vs 25.9%). These findings lead to the conclusion that the higher the level of education, the more prevalent the choice of cohabitation for women, whereas the opposite is true for men.
Employment status is also associated with Chinese immigrants' choice to cohabit. As shown in Table 2, 35.4% of female students choose cohabitation, whereas this is only 10.1% for male students. Concerning partnership type, we notice that results are in line with our hypothesis (5). For mixed relationships, the percentage that chooses cohabitation is higher than for the homogamous relationship types, both for men (39.7%) and women (34.7%). For transnational partnerships, we observe the lowest percentages of cohabitation. Only 3.6% of men and 5.2% of women that are in a transnational relationship choose cohabitation. Table 3 reports the results of the binomial logistic regression analysis in which we estimate the effects of the year of cohabitation or marriage, migrant generation, education, employment status, and partnership type for each gender separately. The results largely confirm what we already observed in the bivariate exploratory analysis. In the regression analysis, we observe significant positive effects of the year of cohabitation or marriage for both men and women, indicating an increased rate of cohabitation in recent years among Chinese immigrants. For education, we observe a negative effect on the odds of cohabiting (vs being married). The lowest educated men have odds of cohabiting that are  (50) 100% (872) 3.15 times higher than those of highly educated men. Meanwhile, the lowest educated women have odds of cohabiting that are almost 1.6 times of those of the highest educated women. Concerning work status, we observe only one significant effect: Chinese female students have odds of cohabiting that are 1.2 times higher than those who work. No significant net effect is found on the migrant generation. This implies that the difference we observed in the bivariate analysis was due to the association of immigrant generation with the other correlates that we included in the multivariate analysis. As for partnership type, we observe a positive effect on the odds to cohabit (vs being married) for mixed partnerships for both genders. Chinese male migrants in mixed partnerships have odds of cohabiting that are almost 19 times higher than men in transnational partnerships and1.9 times higher than men in local co-ethnic partnerships. For women in mixed partnerships, the odds of cohabiting are nearly 9.7 times higher than for

Discussion and conclusion
The significance and meaning of cohabitation vary considerably across ethnic groups depending on specific contexts. It has been shown that cohabitation can be especially common among disadvantaged minority groups, such as Blacks and Hispanics in the US (Bumpass & Lu, 2000). To our knowledge, only a limited number of studies are available on cohabitation among ethnic groups in Europe (Carol et al., 2014;Kogan & Weißmann, 2019). These studies, moreover, mostly consider larger ethnic minorities, often with a Muslim background. Attention is lacking for small ethnic groups from other historical and cultural backgrounds (e.g., Asians). This Cohabitation among Chinese minorities in Western countries:… paper aims to fill the gap by investigating the prevalence and correlates of cohabitation among Chinese immigrants in Belgium using data extracted from the Belgian National Register, which provides information on the whole population of interest between 2005 and 2015.
Regarding the prevalence of cohabitation, we observe that the cohabitating trend among Chinese immigrants in Belgium gradually increases. Regression results of the year of cohabitation and marriage have shown a continually increased rate of cohabitation in recent years for both Chinese men and women. The increasing trend of cohabitation for Chinese immigrants is parallel to the increase observed among Belgians and in China, while it is not as steep as that among Belgians but is more apparent than in China. Our data show that around 33% of Chinese immigrants chose cohabitation in 2015. Such proportion is generally smaller compared to the Belgian population, of which half choose for cohabitation since 2000 (Corijn, 2017). By contrast, the proportion cohabiting among Chinese in Belgium is high compared to that of the Chinese in China, where only 8.1% of men and 6.2% of women in China choose cohabitation (Yu & Xie, 2015). Although we have to remark that the prevalence of cohabitation cannot simply be compared between Belgium and China, also because of different definitions used (e.g., registered cohabitation in the Belgian data vs. factual cohabitation in China), the large difference clearly indicates that cohabitation is more frequently chosen as an alternative to marriage among Chinese immigrants than among people living in China.
The observed upward trend in the prevalence of cohabitation among Chinese immigrants in Belgium could be understood from two alternative evolutions. First, it could refer to improvements in gender equality. Gender inequality in patriarchal culture has been criticised for a long time since it hierarchically places males' authority above females in the family (Zhang et al., 2005). With improved education levels and economic independence, women's social status in contemporary China is better than ever before. However, some traditional gender restrictions still exist in a patriarchal society. Immigrants could more freely enjoy premarital relationships by migrating to a society where gender equality is comparatively better. Secondly, being exposed to western media that introduce liberal views on premarital sex and cohabitation can also reinforce the growing sexual liberation of Chinese women living in a Western-European context. Moreover, at a general level, the observation that the rate of cohabitation among Chinese immigrants, lying between that of Belgians and that observed in China, is in line with what could be expected from the assimilation hypothesis. Chinese immigrants originate from a context with a low cohabitation rate and a generally low degree of acceptance of cohabitation. As they are confronted with a western view on family formation in the process of the Second Demographic Transition, Chinese minorities gradually adapt western attitudes and family-formation patterns to the majority group. Meanwhile, in China, the traditional concept of cohabitation is changing as well, especially with respect to the decreasing restrictions imposed on women's behaviour and the increasing tolerance of premarital sex. For Chinese immigrants in Western societies, this process then is intensified and accelerated.
However, another important finding was that there is a large difference in cohabitating between male and female Chinese immigrants. We observe that more women (25.9%) than men (16.9%) involve in cohabitation (Table 1). To understand this unexpected result, we advance several reasons pertaining to education level, the uncertainty of individual status, social resistance and traditional family pressure.
Firstly, we noticed that Chinese minority individuals who prefer cohabitation are often in "disadvantaged" positions in the marriage market. Our data showed two groups that have high rates of cohabitation: one is the highly educated women; the other is the low educated men. They are both disadvantaged in marriage markets due to the Chinese hypergamy traditions (Xie, 2013a;Xu et al., 2000). Status hypergamy is a tradition that implies that women marry men with higher social status (education levels, economic resources) (Xu et al., 2000). Based on this norm, men with low education levels find it difficult to marry because they lack economic resources. Women with high education levels have difficulties with marrying as well because of the scarcity of highly educated (or higher social status) Chinese partners in Belgium. In China, it is also the case that fewer men than women have high education levels, as college enrolment numbers among Chinese women have surpassed men in recent years (Ji & Yeung, 2014).
Similarly, the data of our study shows that, among the Chinese immigrants, the proportion of highly educated women (30%) is much larger than men (19%) ( Table 2). With more numbers of 'disadvantaged' highly educated women, these women cannot find proper marriage partners in an unbalanced marriage market. Consequently, cohabitation could be the second-best alternative to marriage for these women.
Partly contrary to the pattern of disadvantage (POD) thesis, stating that the rise of cohabitation was due to a worsening of the living conditions of the poorer segment (Lesthaeghe, 2020), our analyses reveal that Chinese women's high rates of cohabitation cannot fully be ascribed to the disadvantaged socioeconomic status but also relate to cultural restrictions on their marriage markets. Besides, based on the second demographic transition (SDT) theory, the rising trend in cohabitation follows an increase in education levels in most Europe countries, and the rise of cohabitation was due to a cultural revolution (Lesthaeghe, 2020). However, because of different cultural frameworks, the positive effect of education levels may not have the same intensity among ethnic minorities. Among Chinese women, moreover, high rates of cohabitation could also relate to practical issues in the migration process. Despite the POD explanation indeed revealing 'a divergence in family formation strategies based on the socioeconomic status' (Ramm & Salinas, 2019), our study indicates that the reasons for rising cohabitation are more complex than merely related to socioeconomic status. Secondly, cohabitation could be preferred due to uncertainty about the future, especially for female students. Our results show especially high odds of cohabitating among female Chinese students in Belgium, with 35.4% of those choosing for cohabitation versus only 10.1% among male students (Table 2). Students could choose cohabitation because of uncertainty about their future, such as how long they will stay in Belgium and whether they will not or will be able to stay permanently. Cohabitation then becomes an alternative to marriage when confronted with uncertain situations. Another reason for cohabitation is that cohabitation is much easier to dissolve than marriage. This could be especially important for someone who may only stay in Belgium for a short period. Although other studies (Brown et al., 2008;Kalmijn, 2007;Sassler & Lichter, 2020;Smock et al., 2005) found that couples could prefer cohabitation (rather than marriage) because of financial insecurity, we did not find confirmation for this effect in our analysis. In contrast, we found indications that the uncertainty about the possibilities for a long-term stay in the host country could lead to a preference for cohabitation, especially among Chinese females.
Thirdly, choosing cohabitation can be a reaction to avoid social resistance from the majority population in the receiving countries towards inter-ethnic relationships for Chinese immigrants in Western societies. Our results show that for mixed partnerships cohabitation is the most popular choice for Chinese immigrants (52.9%) ( Table 1). Such a high interethnic cohabitation rate maybe because of the resistance of interethnic marriages from the mainstream society. Previous studies have shown that some unfavourable attitudes exist towards foreign immigrants, especially during periods of high unemployment (Coenders & Scheepers, 2008). Resisting inter-ethnic marriages can be one of the elements of a more general resistance toward the social integration of foreigners (Tolsma et al., 2008).
Last, cohabitation could be a solution to avoid pressure from family responsibilities for some young Chinese couples. Confucianism's central role in shaping the Chinese family's moral responsibilities can still impose a huge burden on married young couples on the obligation between parents and offspring, ancestor worship, filial piety, and a strong sense of familial honour and transmission (Meyer, 2007), even though traditional cultural influences have become attenuated because of modernization and individualism. These values can still put heavy moral burdens on married women to take responsibility for childbearing to continue the family line. With such pressure from family, it is then conceivable that some Chinese minority women prefer cohabitation instead of marriage to avoid social pressure concerning fertility or other expectations.
To summarise, we could advance the interpretation that Chinese immigrants mainly regard cohabitation as a practical strategy to solve issues caused by migration rather than being a positive choice for a specific type of partnership. Especially for female Chinese immigrants who are confronted with a lack of eligible marriage candidates, who feel uncertain about their future, who possibly suffer from discrimination from the majority group, and feel obliged to fulfil family responsibilities, cohabitation could be an attractive alternative to marriage. Since cohabitation might be considered as the second choice (compared to marriage) for Chinese immigrants, it has not been regarded as a full alternative as the receiving country does, but rather it is a practical living strategy to cope with the consequences of migration.
The gender difference in the prevalence of cohabitation could indicate that the adaption process and attitudes concerning cohabitation are different for Chinese men and women. The disparity is shown as that Chinese women more quickly than men accept the western attitudes concerning family (formation) to that in the majority population. More importantly, such gender difference in the tendency of cohabitation suggests an uneven gender equalisation process in family changes and ideological shifts. Although our study clearly could confirm the assimilation process among Chinese immigrants, it shows that home-culture still strongly influences the meaning and role of cohabitation for the ethnic minorities. This analysis is consistent with the study of Zhang (2017), who assumed that the Confucian culture still has a strong influence on the meaning of cohabitation in China (Zhang, 2017).
In conclusion, as for Chinese minorities, we acknowledge that cohabitation is not considered to be a full alternative to marriage as is the case in the receiving western society, nor is it the only choice for socioeconomically disadvantaged people. This complex process depends on the assimilation status of immigrants and their attitudes to cohabitation, which can be different by subgroups within the ethnic community. Accordingly, cohabitation is rather a practical living strategy to cope with the consequences of migration for ethnic minorities or immigrants. Meanwhile, the gender differences in Chinese minorities show that the transition of family formations and attitudes could differ within an ethnic minority The latter observation implies the importance, also for further research on this topic, that an ethnic minority group cannot be considered as a homogeneous group and that differences within the group, such as gender differences, are to be taken into account to reach more in-depth insights.
A major limitation of our study is that the data used from the Belgian National Register is limited to basic socio-demographic information. This implies that the set of explanatory variables that could be used is rather limited. Further research on this topic, possibly collecting new survey data, could include information on motivations for cohabitation (vs marriage) and other elements that we used to interpret research findings, such as gender role attitudes, plans for length of stay in Belgium, preferred characteristics in the partner, adherence to traditional family values, fertility status, experiences with previous relationships and so on. Another limitation of this study is that the data is limited to the period between 2005 and 2015, which is a relatively short time to reveal structural trends.
In spite of the exploratory nature of this research and its limitations, it provides possible explanations for cohabitation motivations among disadvantaged groups, adding to our understanding of the diverse reasons why people may choose cohabitation, by race and gender. The observed diversity within a single ethnic minority groups suggests the possibility of a different process of ideological shifts and transitions in familial attitudes between both genders within an ethnic group, which, moreover, cannot be understood without taking the cultural frameworks from the country of origin into account.