Many studies have documented that partners affect each other’s health [1-8]. These say that partnered individuals, and married people in particular, are generally healthier and live longer than others, and the positive association between partnership and health is especially true for men. Health benefits may accumulate due to protective effects that enable income pooling and increase economic resources , which in turn provide emotional support, social networks and social control [10-11].
Mutual associations of this kind may be both health beneficial and detrimental. Partners can influence the take-up of health-related benefits and improve quality of life, but also affect health behaviors in terms of smoking, drinking, diet, and exercise [12-15]. Among the elderly, hospitalization or death have been found to increase the mortality and morbidity risks of the spouse [16-17]. Mutual relations of this kind confirm that one of the many functions of partnership is health production, and this implies that the connectedness of individuals should be considered when studying health outcomes [18-20].
The literature on health dependencies among partners typically ignores diversity when it comes to partnership characteristics, which is problematic because it obscures effect heterogeneity . Different combinations of partnership characteristics, be they demographic or social, have consequences on many facets of life and may thus be important for the health of both parties. One salient example is the ethnic or racial composition of the couple. To the extent that partnerships across ethnic or racial borders have less resources, are not fully accepted socially, or are subject to discrimination, such partnerships may adversely impact health due to stress . Interethnic and interracial partnerships have increased considerably in number in both the U.S. and Europe, with extant studies focusing on health, but few studies have investigated associations between partnership composition and health. Most of the literature has been concerned with how one partner’s ethnicity or race affects and dominates health within the couple [23-24], while the empirical evidence on the role of the ethnic or racial composition itself has been less studied and is inconclusive .
In this study, we extend prior work on partnerships and health by investigating how married and cohabiting partners mutually influence each other’s receipt of health-related benefits. We focus on how any such correlations vary by the couple’s ethnic composition. Using longitudinal population register data from Finland, we study a person’s first-time receipt of sickness allowance or disability pension as a function of the partner’s receipt of the same. Both benefits are related to reduced working capacity before statutory retirement age, and they are by far the two most common health-related benefits available . These two measures could reflect not only health dependencies between partners but also rational decisions concerning mutual leisure time versus economic loss within the partnership.
Finland provides an interesting case of study that contrasts the dominating U.S. literature on interracial and interethnic partnerships and health. The country can be characterized as a comprehensive welfare state with universal coverage of social and medical services, high take-up of sickness allowance and disability pension, modest social disparities and income inequality, a well-projected ageing process, and high levels of gender equality. Finland also features two ethnic groups, Finnish and Swedish speakers, which are distinguished in the population register by their unique mother tongue. Finnish speakers amount to 90 percent of the total population, and Swedish speakers make up five percent. Both groups are native and have a similar socioeconomic position, but Swedish speakers are notably healthier than Finnish speakers on objective health measures, including sickness allowance, disability pension and life expectancy . Although intermarriage has been high during the past decades, the two groups can still be culturally distinct with regard to practices to family life and stability of the nuclear family . Divorce and separation rates are lower in Swedish than in Finnish unions, and higher in exogamous than endogamous ones . In the cohorts studied, intermarriage across other ethnic lines than the Swedish and Finnish has been very rare, and that also lies beyond the scope of this paper, as Finland has experienced foreign-born immigration only very recently. By focusing on the ethnolinguistic composition of the native couples, we contribute to the small but growing literature that addresses heterogeneity in the health impacts of partnerships, in a context that is much different from that in the United States, a subject explored in greater depth.
Partnership and health
Among the explanations for why partnered individuals generally have better health than others is a strand of literature emphasizing the protective importance of social structures in which individuals live their lives [10, 30]. A greater sense of connectedness with other people, not least a partner, may foster a healthy lifestyle [7, 31]. Moreover, marriage and other close social relationships offer support as a buffer against various stressors that may be damaging to mental and physical health . The marital resource model further proposes that marriage itself can improve individuals’ health through access to health-protective resources, such as income pooling [7, 9, 31]. Another explanation for better health among partnered individuals is selection, in that healthy people are more likely to enter partnership and marriage and less likely to divorce [3, 32].
Within partnerships like marriage, spouses mutually influence each other’s mental and physical health. A partner’s health is often like that of the other, and these tend to converge over time . This can be understood on the basis of ideas that integrate theories of marriage markets and health capital formation . They argue that inter-spousal correlation in health status follows on assortative mating, common health behaviors, shared environmental risk factors, and the direct effects of the health of one spouse on that of the other. Assortative mating means that partners sort on certain characteristics and tend to be similar when it comes to demographic characteristics, preferences, and health-related behavior [14, 33]. Heterogamy, on the other hand, implies a difference between partners, which may lead to less health concordance compared to homogamous partnerships. According to marital resource theory, inter-spousal correlation in health may be a function of shared life events and resources, for good or bad. Thus, the presence of a partner is not necessarily protective, and partners may influence each other negatively in terms of health behavior and health  and thus contribute to mutual health benefit receipt.
Change in one partner’s health or health-related behavior may therefore induce change in those of the spouse. Among the elderly, it is commonly found that the hospitalization or death of one spouse affects the other through increased mortality or morbidity risks, and these associations are both immediate and long-term [16-17, 35]. Among aging adults, each partner influences the other’s quality of life  and health behaviors such as smoking, drinking, diet, and exercise . Likewise, each partner also influences the disease risks faced by the other regarding, for example, diabetes, metabolic syndrome, hypertension, cancer, and depression [33, 36-38]. A previous study has also documented strong and long-term interrelations among partners in receipt of sickness allowance and disability pension . Such interrelations could be explained not only by collateral health but also by strategic decisions concerning mutual leisure time and informal care needs versus income loss between partners.
Coordinated behaviors related to retirement and sickness absence may consequently also be understood from the perspective of how economists model labor supply decisions, where individuals strike an optimal balance between the cost of foregone leisure and the benefits of increased income through paid employment. Correlations in behaviors related to retirement and sickness absence may therefore not be exclusively explained by collateral health between spouses. If married couples place greater value on leisure time spent together, withdrawals from the labor market will increase when one spouse is not working. In this case, individuals consider their own leisure time and that of the spouse as complements [39-40]. Depending on their characteristics, couples may pool resources and influence each other in their decisions to retire or to leave the labor force . This would mean that the feasibility of leaving the labor force derives not only from each spouse’s own resources and life-course experiences but also from those of the spouse. The quality of the relationship will presumably have a strong impact on this mutual decision. The expectation of spending more time together following sickness absence or early retirement is most likely prevalent among couples who have already established some joint activities prior to retirement and who enjoy a high-quality relationship. Thus, partners in a stable and high-quality relationship are likely to look forward to opportunities for more mutual leisure time in their retirement years and may therefore decide on early retirement. Partners whose relationships are already strained and unstable prior to retirement, on the other hand, may view continued participation in the labor force as a welcome escape from marital tensions and may be more disposed toward delaying retirement. Accordingly, relationship satisfaction seems to rank second to standard economic reasons as the strongest predictor of retirement timing .
Ethnic composition of couples, health and mutual health benefit receipt
Financial and psychosocial resources that improve health and protect well-being differ across population subgroups [9, 43-44], and such differentials may impact health-related decisions concerning the receipt of health benefits. The extended kin networks and social integration that a partner brings to a relationship may improve the couple’s health and well-being [4, 7, 31], though this may be more limited among heterogamous couples than those that are homogamous . Economic resources and wealth may also influence decisions related to retirement or receipt of sickness benefits. Incentives to leave the labor force derive not only from each spouse’s own economic resources and life-course experiences but also from those of their spouse. The level of economic resources within ethnic groups may therefore contribute to any differential in this association across exogamous and endogamous couples.
The limited research on how the ethnic or racial composition of couples affects health has mainly hypothesized that exogamous partnerships will be less health-protective than endogamous ones because of less efficient communication and coordination within the couple and fewer social resources [11, 46]. Exogamy may also be less health-protective because of stress experiences that could result in psychological and physical health problems if such a partnership is not fully socially accepted or is subject to discrimination [30, 47]. Empirical evidence primarily from the U.S. is mixed, however [22-25]. Partner’s race appears to dominate the mere composition in this regard, and typically distinguishes between Whites and Non-Whites. Contrary to the arguments of stress process theory, having a White spouse is associated with better self-rated health for minorities, while intermarried Whites experience worse health . This would suggest that the greater resources brought by a majority White partner to the relationship may go so far as to benefit their minority partner’s health. Partners of different race or ethnicity may thus bring different economic and psychosocial resources to the partnership [11, 44, 48].
Gender may moderate the associations between relationship-related resources, stress and health. Overall, being married seems to matter more for men’s health than for women’s health. This is in line with gendered specialization of household labor , where women do more unpaid work, including caregiving, though the extent of specialization may differ across racial or ethnic groups . Thus, the health impact of partner’s race or ethnicity may matter more for men than for women. On the other hand, women generally experience more stress than men, which may explain why White women in interracial partnerships in the U.S. experience more psychological distress . In that case, the health impact of the partnership’s interracial or interethnic composition may matter more for women than for men.
The present study
We investigate partners’ mutual receipt of sickness allowance and disability pension in Finland in the period 1987-2011 among native-born couples with different ethnic characteristics. We distinguish between endogamous Finnish couples, endogamous Swedish couples, exogamous couples with a Finnish-speaking man and Swedish-speaking woman, and exogamous couples with a Swedish-speaking man and Finnish woman. The ethnic composition of the couple should be considered as a proxy for partnership context of relevance for health concordance rather than as a causal factor.
As in most other countries, endogamy is the prevailing norm for mate selection in Finland, though intermarriage has become more common, particularly in the case of Swedish speakers. In the 1950s, approximately 20 percent of the Swedish-speaking population married a Finnish speaker. This share rose until the 1980s when it leveled off, and today about 40 percent of Swedish speakers form a union with a Finnish speaker . Unlike intermarriage among many other ethnic or racial groups, there are few easily discernable differences between Finnish and Swedish speakers except for their mother tongue, which is a marker for ethnic affiliation in the population register. Discrimination against and social stigma associated with this form of exogamy is therefore very limited. There are nevertheless group differences of relevance, in that Swedish speakers have better health and live longer than Finnish speakers [27, 50]. These differences in health may be partly associated with Swedish speakers having better and more extensive social networks. There are also group differences in cultural practices related to family life and the stability of the nuclear family [28, 51-52]. Finnish speakers have almost twice as high a risk of divorce and separation from their cohabiting partner compared with Swedish speakers, which could reflect lower relationship quality. In addition, exogamous unions have an elevated, or approximately ten percent higher, risk of ending in divorce or separation compared to endogamous Finnish unions . As already noted, Finland has experienced foreign-born immigration only very recently. In the cohorts studied, intermarriage across other ethnic lines than the Swedish and Finnish has therefore been very rare, and lies beyond the scope of this paper.
Based on this context and the mechanisms involved, we make the following conjectures. First, considering that all the study individuals are part of the same institutional setting and benefit system, we expect to find evidence for partners’ mutual receipt of health benefits in both endogamous and exogamous couples. Second, irrespective of the couple’s ethnic composition, we expect that partner concordance in benefit receipt would be stronger for disability pension than for sickness allowance, because the former indicates a more severe state of poor health. It also reflects a permanent exit from the labor market and could thereby be influenced to a greater extent by joint retirement decisions between partners. Third, we expect stronger cross-spousal dependence in health and benefit receipt in endogamous Swedish couples than in endogamous Finnish couples. The strength of partner correlation regarding health benefit receipt may differ by ethnic composition of the couples. Different cultural practices regarding family dynamics, such as union stability and marital satisfaction, across the ethnic groups in our study, suggest more commitment and stronger family ties among Swedish speakers than among Finnish speakers. This may be associated with higher relationship quality and more efficient communication and coordination within the couple. This would mean that Finnish partners may influence each other less than Swedish partners when it comes to health and concordant decisions about disability pension/health benefit receipt. Fourth, because women in general are more likely to shoulder more caregiving responsibilities and suffer from stress reactions, we expect the woman to be more sensitive to the man’s benefit receipt than vice versa. Fifth, if the gender-specific sensitivity to partner’s benefit receipt dominates the behaviors typical for each ethnic group, we expect the woman in an exogamous couple to be less sensitive to the man’s benefit receipt than vice versa. For example, exogamous couples in which the man is Swedish and the woman is Finnish would then display stronger associations in cross-spousal benefit receipt than exogamous couples in which the man is Finnish and the women is Swedish.