Ethnic Diversity in Residential Neighbourhood
Table 2 shows the results from a series of proportional hazards models indicating the probability that native Finnish women and men will start first union with immigrant partner.
Table 2 Hazard ratios for the transition to exogamous and endogamous first partnerships by ethnic diversity in residential neighbourhood, Finland, native population, birth cohorts 1981–1995. The initial model (M1.1) includes the ethnic diversity in the residential neighbourhood. The estimates from this model reveal a strong positive association between the share of immigrants in the neighbourhood and the propensity of native Finns to form partnerships with them. Among women, living in areas where the proportion of immigrants ranges from five to nine per cent relates to 77% increase in the likelihood of inter-ethnic union, compared to the reference category (the share of immigrants below five per cent). For men, the effect appears closely similar (+ 82%). Larger proportions of immigrants further increase the chance of forming an ethnically mixed partnership. In the areas of highest concentration of immigrants, the likelihood that native women and men would start a mixed union exceed the reference category 2.8 and 2.7 times, respectively.Footnote 2
Further, a series of control variables were added to the model. This markedly reduced the hazard ratio for ethnic diversity in the neighbourhood (M1.2). For areas with the highest concentration of immigrants, the hazard ratio decreased from 2.88 to 1.45 among women and from 2.66 to 1.37 among men, respectively. However, despite a reduction relative to the initial model, living in an ethnically diverse neighbourhood significantly increases the likelihood of native Finns’ initiating a mixed partnership. A statistically significant increase also persists in areas with a relatively moderate (five to nine per cent) proportion of immigrants. This applies to women and men alike, with no marked gender difference in the effect.
A stepwise procedure of adding variables to the model (not shown in Table 2) revealed the controls that made a more sizeable contribution to the change in the effect of the neighbourhood variable. The inclusion of the size of the residential neighbourhood produced a moderate reduction in the hazard ratios associated with the proportion of immigrants in the area while the inclusion of the characteristics of individuals made only little difference in the relationship between ethnic diversity and the likelihood of forming a mixed partnership. By contrast, the inclusion of the region markedly reduced the hazard ratio for ethnic diversity in the neighbourhood. In our view, this indicates that local marriage markets are not confined to immediate neighbourhoods, but cover much wider geographical areas.
The results for endogamous partnerships corroborate the findings reported above. Unlike for mixed unions, the models for endogamous partnerships reveal a negative gradient for the main independent variable. Among native Finns, the ethnic diversity in residential neighbourhood is associated with significant decrease in the propensity to form endogamous unions. The comparison of results across models (M1.1 and M1.2) shows that the statistically significant negative effect does not fade away after the inclusion of controls for the size of residential area, individual characteristics and the region in the models. The opposite gradient of our main independent variable for endogamous and exogamous partnerships suggests that the increase in ethnic diversity in residential neighbourhoods has a potential of advancing the ethnic intermarriage and replacing endogamous unions in part with ethnically mixed partnerships.
Ethnic Diversity at Workplace
In order to get an account of the effects on partnership formation of ethnic diversity at work, another set of event history models was estimated (Table 3). In these models, the main independent variable is the share of co-workers of immigrant background at the person’s place of work.Footnote 3
Table 3 Hazard ratios for the transition to exogamous and endogamous first partnerships by ethnic diversity at workplace, Finland, native population, birth cohorts 1981–1995. The initial model (M2.1) includes the ethnic diversity at workplace. The estimates from the initial model indicate a significant positive relationship between the share of immigrant co-workers and the propensity of native Finns to enter ethnically mixed unions. For women, having a workplace in which the proportion of immigrants ranges from five to nine per cent of employees is associated with 51% increase in the hazard ratio relative to the reference category (workplaces with less than five per cent of immigrants). Among men, the relationship is closely similar (46% increase in the hazard ratio). Larger proportions of co-workers with immigrant background further add to the likelihood of forming an ethnically mixed partnership in more or less a linear fashion. In case the share exceeds 15%, native women and men in Finland exhibit a 2.5- and 2.3-fold increase in the chances of partnering with a person of non-native background, respectively.
The addition of the control variables to the model markedly reduced the hazard ratio for ethnic diversity at workplace (Model M2.1). For workplaces with the highest proportion of immigrants, the hazard ratio decreased from 2.53 to 1.77 among women and from 2.27 to 1.57 among men, respectively. However, notwithstanding the change related to the inclusion of the control variables, Model M2.1 clearly indicates that ethnically diverse workplaces are associated with an elevated likelihood of native Finns’ initiating an exogamous union. A significant increase in the likelihood of inter-ethnic unions also persists for workplaces with a moderate (five to nine per cent) proportion of immigrants.
The stepwise addition of the control variables to the model (not shown in Table 3) identified those that made a more important contribution to the reduction in the effect of the workplace. In accord with the findings for residential neighbourhood, the inclusion of the region significantly reduced the hazard ratios for ethnic diversity in the workplace. This suggests that the hazard ratios for ethnic diversity in the workplace may partially reflect exposure to immigrants beyond the workplace, which is more common in cities, especially the capital region.
The hazard ratios for ethnic diversity at workplace appear somewhat higher for women. Although the contrast is not large, it cuts across all models and levels of the independent variable. Perhaps, this can be regarded as yet another sign of the advanced equity of men and women achieved in the Nordic societies, with work being an equally important part of people’s lives, irrespective of gender.
For endogamous unions, the association with ethnic diversity at work runs in the opposite direction: the increase in proportion of immigrants at workplace leads to lower propensity to start endogamous partnerships. According to the final model (M2.2), native men exhibit up to 11% decrease in the rate of entry into endogamous partnerships, associated with the increase in immigrants at workplace. Among women, the reduction in the hazard ratio is only slightly smaller (− 9%). Across models, the hazard ratios for endogamous unions are more stable than those observed for exogamous partnerships. This lends support to the notion that the control variables employed in the analysis are more extensively modulating the formation of the latter kind of partnerships. The opposing gradients found for exogamous and endogamous partnerships suggest that the increase in ethnic diversity at workplaces tends to promote intermarriage between natives and immigrants.
Joint Effects of the Ethnic Diversity in Neighbourhoods and Workplaces
In the two previous sections, the role of ethnic diversity in residential neighbourhoods and workplaces was investigated separately. However, the contacts and integration between the natives and immigrants in different life domains do not develop in isolation. To provide insight into how the inter-ethnic encounters occurring in neighbourhoods and workplaces come together, we estimated additional models with ethnic diversity in both domains considered (Table 4).
Table 4 Hazard ratios for the transition to exogamous and endogamous first partnerships by ethnic diversity in neighbourhood and workplace, Finland, native population, birth cohorts 1981–1995. The estimates obtained from the joint model (M3) show that the effects of ethnic diversity in residential neighbourhood and workplace are only slightly altered compared to the separate models (M1.2 and M2.2 in Tables 2 and 3). For exogamous partnerships, the largest reduction in the hazard ratio (from 45 to 40%) can be observed among women residing in the neighbourhoods with high share of immigrants. Otherwise, the typical change in the hazard ratios is limited to 2–3 percentage points; for endogamous unions, the change is even smaller. A tentative conclusion that can be drawn is that the influence of ethnic diversity in residential neighbourhoods and workplaces is to a large extent independent from each other.
The results obtained from the joint models also suggest that the increase in ethnic diversity at workplaces exerts somewhat stronger influence on the propensity to start exogamous partnerships than diversity at residential neighbourhoods. For both women and men, the hazard ratios for the share of immigrant co-workers in most cases exceed those for the proportion of immigrants in the neighbourhood.Footnote 4
To further elaborate the interplay of ethnic diversity in different domains, we employed an interaction between our two main independent variables (Table 5). In the interaction model, we distinguished between residential neighbourhoods and workplaces with higher (10% or above) and lower (below 10%) share of immigrants. To facilitate the interpretation of the results, we have derived two additional sub-tables (the middle and lower panels) from the primary interaction (the upper panel).
Table 5 Interaction of ethnic diversity in neighbourhood and workplace for the transition to exogamous first partnerships, Finland, native population, birth cohorts 1981–1995 The middle panel of the table shows how the chance of starting a mixed partnership is modulated by the ethnic diversity in the neighbourhood for native women and men who are employed in establishments with lower and higher share of immigrants, respectively. The high ethnic diversity in neighbourhood makes a larger contribution among people who encounter less diversity at work; both women and men belonging to the latter group exhibit a 20% increase in the likelihood of exogamous union. By contrast, among those who are more exposed to inter-ethnic contacts at workplace, the role of ethnic diversity in the neighbourhood appears more limited. For women, it adds seven per cent to the likelihood of partnering with immigrants. For men more exposed to inter-ethnic contacts at workplace, living in ethnically diverse neighbourhood fails to make any additional contribution to mixed partnerships.
The lowermost panel of Table 5 illuminates the ways how the forming of exogamous partnerships is shaped by the ethnic diversity at workplace among natives who are living in the neighbourhoods with varying proportion of immigrants. In our view, two main conclusions can be drawn from the panel. First, the effect of ethnic diversity at workplace appears more pronounced among persons who live in neighbourhoods with lower share of immigrants. This corroborates an observation made in the previous paragraph for the area of residence. Viewed together, these findings tell us that the effect of inter-ethnic encounters on partnership formation may not be (fully) multiplicative across life domains. The increased exposure to ethnic diversity in one domain seems to reduce the relative contribution of additional inter-ethnic contacts that occur in other domains. Second, the results lend further support to the notion derived from the main effects models that the share of immigrants at workplaces has a more significant bearing on the formation of native-immigrant partnerships than the ethnic diversity in residential neighbourhoods. We think that the latter result plausibly arises from the greater intensity of interpersonal encounters at workplaces relative to those that occur in residential neighbourhoods.
Finally, the results from interaction models suggest that the effect of ethnic diversity is somewhat more pronounced among women. This can be observed in the lowermost panel of Table 5, which presents the effect of ethnic diversity at workplace. Notably, this finding runs counter the traditionalist expectation according to which work should occupy a more central role in the lives of men.
Heterogeneity Associated with the Partner’s Origins
The purpose of extending the analysis in this direction was to ascertain whether the effects of ethnic diversity in residential neighbourhoods and workplaces reported earlier in the article are universal or driven by immigrants who are culturally close to the host country. In order to answer this question, additional competing risk models were estimated for exogamous unions between native Finns and immigrants from Western and non-Western countries.Footnote 5 In these models, we reconfigured the main independent variables so that they reflect the group-specific proportions of immigrants of Western and non-Western origin in residential neighbourhoods and workplaces rather than the overall proportion of immigrants.
The upper part of Table 6 shows strong associations between the proportions of immigrants of different origins in the residential neighbourhood and the propensity of native Finns to partner with them. However, the associations seem to be group-specific. A higher proportion of immigrants of Western origin in the neighbourhood is related to a greater likelihood of forming partnerships with them. For areas with the highest proportion of Western immigrants, the group-specific hazard ratios are 1.85 for Finnish women and 2.32 for men. An elevated proportion of immigrants of non-Western origin in the neighbourhood is also associated with a higher propensity to partner with them, but the effect is somewhat less pronounced than that observed for immigrants from Western countries. In areas with the highest concentration of non-Western immigrants, the hazard ratios are 1.59 for native women and 1.61 for men.
Table 6 Hazard ratios for the transition to exogamous first partnerships with Western and non-Western immigrants, by ethnic diversity in the neighbourhood and at workplace, Finland, native population, birth cohorts 1981–1995. The lower part of Table 6 reveals significant associations with regard to the proportion of immigrants of Western and non-Western origin in the workplace. For both groups, a higher concentration of immigrants tends to increase the likelihood of partnering with native Finns. However, unlike residential neighbourhood, the group-specific effect of ethnic diversity in the workplace appears somewhat stronger for immigrants from non-Western countries. When the proportion of non-Western immigrants exceeds 15% in the workplace, the chance of their partnering with native Finnish women and men increases by 2.4 and 2.0, respectively.
The hazard ratios for immigrant groups that are beyond the focus of the particular group-specific model show a negative or neutral gradient with regard to our main independent variables.Footnote 6 A higher proportion of non-Western immigrants in the neighbourhood is associated with a decrease in the propensity of native Finns, both women and men, to partner with those of Western origin. Likewise, a high proportion of Western immigrants tends to reduce the likelihood of Finnish men’s partnering with women of non-Western origin. The pattern among Finnish women is less clear. These results support the notion that different immigrant groups may to some extent be competitors in the local marriage market. However, this pattern does not extend to the workplace, where an elevated proportion of potential partners from another immigrant group does not in most cases have a statistically significant effect.
In summary, our results are at odds with the expectation that the effects of ethnic diversity in residential neighbourhoods and workplaces reported in the previous sections are primarily driven by immigrants who are closer to Finns. The findings suggest that encounters in neighbourhoods and workplaces may increase the likelihood of exogamous partnerships both for groups who are more similar to the host society and for those who are more distant.
Effects Associated with Control Variables
The estimates for control variables presented in Table 7 are obtained from models that jointly consider the effect of ethnic diversity in residential neighbourhoods and workplaces for the formation of exogamous unions (M3 in Table 4, and M5 in Table 6). In addition to the overall pattern, the differences between unions with Western and non-Western migrants are also addressed.
Table 7 Hazard ratios of control variables for the transition to exogamous first partnerships with Western and non-Western immigrants, by control variables, Finland, native population, birth cohorts 1981–1995. The size of residential neighbourhood features a similar gradient across models. The increase in the number of residents in the area evidently reflects an expanding pool of potential partners available for single women and men. The effects are largely similar for partnerships formed with immigrants of Western and non-Western origin. The size of workplace shows less systematic pattern. For women, a larger number of co-workers moderately reduces the likelihood of partnering with immigrants of Western origin. However, for partnerships with immigrants of non-Western origin, the size of the workplace exerts no significant influence. For men, the effects are insignificant irrespective of the partner’s origins.
Region of residence has a strong bearing on the formation of exogamous unions. Not surprisingly, the likelihood of partnering with immigrants is highest in Helsinki, followed by Turku and Tampere, the two largest urban centres besides the capital region. The pattern is similar for unions with Western and non-Western partners. Since immigration on a larger scale is a relatively recent phenomenon in Finland, younger women and men have had more opportunities to meet a non-native partner. Despite that, our analysis covers a relatively narrow range of cohorts, women and men born in late 1980s and early 1990s show a significantly higher likelihood of forming inter-ethnic partnerships than their counterparts in generations born in early 1980s. The increase can be observed for partnerships between native Finns and immigrants of Western as well as non-Western origin.
Being born in Finland to foreign-born parents markedly increases the chance of forming an exogamous union. For these women and men, the hazard ratios for mixed partnerships amount to 2.6 and 2.4 times over the reference category, respectively. The effect appears more pronounced for unions with non-Western partners. This probably reflects a lesser likelihood of partnering with native Finns among the descendants of non-Western immigrants in Finland, compared with the offspring of Western immigrants. It is also interesting to note that the second-generation immigrants feature a much stronger inclination towards exogamous partnerships than return migrants of Finnish origin who were born abroad to Finnish parent(s). In our view, this finding once again underlines the role of individual’s family origin for the partner choice. The effect appears stronger for unions with non-Western partners.
Further, the experience of living abroad for some period makes a remarkably strong contribution to the likelihood of exogamous unions. For men, this experience implies a 2.6-fold increase in the likelihood of mixed partnering, and for women the hazard ratio appears slightly smaller. Unlike family background, living abroad exerts a stronger influence on the propensity to form exogamous unions with partners of Western origin. We assume that this is related to the fact that native Finns migrate predominantly to Western countries. However, we are not able to establish the exact causality here. For instance, migration to a foreign country may be itself driven by partnership formation. In addition, experience of living abroad may also involve selection as international migrants do not constitute a random subgroup of the sending population (Ng and Nault 1997; Frank and Heuveline 2005). If they are selected for greater openness to inter-ethnic contacts, this can reinforce the observed relationship beyond causality. Mother tongue other than Finnish also relates to significantly elevated chance of starting an exogamous union and reduced likelihood of having a native partner. The contrast with the reference group (persons speaking Finnish as mother tongue) is more pronounced for those non-Finnish speakers who do not belong to the Swedish-speaking population. This result is not surprising because the majority of Swedish speakers come from an historical minority which has lived in Finland for many centuries. The increased likelihood of Swedish speakers’ forming partnerships with those of Western origin plausibly reflects their higher propensity to partner with Swedes.
Finally, the relationship between the ethnically mixed unions and educational attainment follows a U-shaped pattern. Judging from hazard ratios, the inclination towards partnerships with immigrants is lowest among women and men with medium level of education. By contrast, the association of schooling with endogamous unions follows a linear pattern. The results from the group-specific models, estimated for inter-ethnic unions with partners of Western and non-Western origin, reveal that a U shape is produced by amalgamating two different patterns. Finnish women with low educational attainment are more likely than their better educated peers to partner with immigrants from non-Western countries. By contrast, highly educated Finnish women exhibit a significantly greater propensity to form unions with men from Western countries. However, this pattern is not characteristic of Finnish men. For the latter, an elevated risk of inter-ethnic union is associated with high educational attainment, regardless of the partner’s origins.