PATTERNS BEHIND UNMARRIED COHABITATION TRENDS IN EUROPE

During recent decades patterns of family formation have undergone a transformation in Europe, with cohabitation without marriage emerging as a key characteristic. Considerable research in Western Europe, the USA and Canada has analysed the trends and patterns of cohabitating unions (Bumpass and Sweet 1989; Kuijsten 1996; Seltzer 2000, 2004; Smock 2000; Ermisch and Francesconi 2002; Kiernan 2002; Le Goff 2002), and relatively recently the research area has extended to Central and Eastern European countries (Heuvelin and Timberlake 2004; Sobotka and Toulemon 2008; Hoem et al. 2009). Until recently, research on cohabiting unions has been mostly country specific (see Mynarska and Bernardi 2007; Matysiak 2009 for Poland; Sobotka et al. 2003 for the Czech Republic; Aassve et al. 2006 and Spéder 2005 for Hungary; Hoem and Kostova 2008 for Bulgaria, Hungary, Romania and Russia; Katus et al.2007 for Estonia), or cohabitation has been applied as a background variable in the analysis of familial or societal activities such as housework (Batalova and Cohen 2004), work-family balance (Kasearu 2009), and partners’ educational homogamy (Hamplova 2009). Comparative research by Eastern European scholars is only now beginning to emerge.

The changes and variations in cohabitation trends and patterns in Europe over time have inspired researchers to look for the typical developmental stages of family formation. Most studies related to Western European countries make comparisons of cohabitation trends using country specific demographic characteristics such as the proportion of cohabiting couples in different age groups, the presence of children, the average age at the first marriage and the divorce rate (e.g., Prinz 1995; Kiernan 2002). Looking for the typical developmental stages that the family formation goes through inspires researchers to predict to some extent the future trends in cohabitation patterns from country to country and offers new insights into the process of possible convergence of family patterns in an enlarged Europe. The spread of cohabitation in Eastern Europe is seen as a universal trend and has been explained in the framework of developmental idealism (Thornton and Philipov 2009). On the other hand, especially if the Eastern European countries are included in the comparative analysis, some studies have indicated that developments in cohabitation patterns tend to be in flux (Daly 2005). Work by several authors suggests that unmarried cohabitation is really one of the most widely differentiated family patterns across Europe (Sobotka and Toulemon 2008; Perelli-Harris et al.2009) and that diversity among countries is even greater in the Eastern part of Europe compared to the West (Fokkema and Liefbroer 2008). The latter conclusion feeds the ongoing debate about the divergence or convergence of family patterns in Europe (e.g., Fokkema and Liefbroer 2008; Thornton and Philipov 2009).

The aim of the current paper is to contribute to the debate about divergence and convergence of family patterns in Europe by adding several Eastern European countries to the analysis alongside the Western ones, and also by adding several additional background indicators in order to reconsider the hypothesis of typical developmental stages of cohabitating unions in the case of each country under the observation. The European Social Survey (ESS) datasets from the second, third and fourth round (see Jowell et al. 2005, 2007, 2009 respectively) form the empirical basis of the study and determine the selection of the countries under observation.

Cohabitating unions started to emerge in the 1950–1960s in Sweden and Denmark. By the end of the 1970s they had spread across Central and North-Western Europe (Trost 1979; Therborn 2004) and since the 1980s also into Eastern Europe (Spéder 2005; Hoem et al.2009). Initially in the 1960s and 1970s couples practised unmarried cohabitation for rational reasons; they were predominantly middle-aged or older, and had dissolved or ended a first marriage (Kiernan 2001). Nazio and Blossfeld (2003) provide an example of the rationale in West Germany where remarriage meant losing either alimony or a widow's pension, so many couples preferred to continue cohabiting without marrying.

Several authors (Prinz 1995; Kiernan 2002; Heuvelin and Timberlake 2004; Perelli-Harris et al. 2009) share the idea that societies follow a standard path and the universal spread of cohabitation is observable as part of a society's typical developmental stages. For example, Sobotka and Toulemon (2008), relying on various studies of cohabitation over time and space, bring out three major stages: diffusion, permanency and cohabitation as a family arrangement. Diffusion is seen as the gradual increase in the number of young people entering a consensual union as a trial period before getting legally married (Bumpass et al.1991; Katus et al.2007; for reviews, see Seltzer 2000; Smock 2000). Nazio (2008: 162) argues that ‘cohabitation is contagious’. However, it varies across countries depending on the institutional context and the stage along the diffusion process.

When cohabitating unions reach a normative status among the population, they enter the permanency stage – they become an alternative form of living together with no clear intention of getting married. To confirm this, for example, Lichter and Qian (2008) brought out the instances of serial cohabitations and post-marital cohabitations. In its third developmental stage cohabitation adopts childbearing and rearing as the main function of the family institution and couples maintain family arrangements similar to those of marriage-based families.

Unmarried cohabitation is a trend which is spreading across Europe; it varies in its typical developmental stages from country to country and may eventually reach a common stage (van de Kaa 1987; Lesthaeghe 1995; Kuijsten 1996; Thornton and Philipov 2009). This conclusion supports the idea of second demographic transition which, according to Kuijsten (1996), ends in the convergence of living arrangements across Europe. Calling this explanation a developmental idealism, Thornton and Philipov (2009) emphasise that it leads to a belief ‘…that modern family is good and attainable, that a modern family helps to produce a modern society and a modern society helps to produce modern families, and freedom and equality are fundamental human rights’ (2009: 140). The spread of cohabitation in Central and Eastern Europe according to Thornton and Philipov could mean that these countries are taking over the family patterns of Western Europe as ones that fit in to a modern society. However, recent empirical findings on unmarried cohabitations give more support to the divergence rather than the convergence (Fokkema and Liefbroer 2008; Perelli-Harris et al.2009) of cohabitation patterns. Next we concentrate on the main aspects of determining the spread of cohabitation: the age of cohabiting couples, the legal marital status and childbearing within cohabitation relying on previous studies.

In most European countries, cohabitation is still age-dependent and is practised mostly by couples in their twenties. However, Kiernan (2002): 11) does show that cohabitation in older age groups is not exceptional in the Nordic countries. Thus the diffusion of cohabitation into different age groups allows for a construction of developmental stages of cohabitation as a social phenomenon. Prinz (1995) divided cohabitation patterns according to the age of the participants into four stages. In the first stage, cohabitation is rare in a society and the mean ages of cohabiting and married populations are close to each other. In the second stage, cohabitation receives social acceptance as a prelude to marriage. In this case, the frequency of cohabitation is high in younger age groups but declines suddenly at the optimal childrearing ages. At the third stage, cohabitation is an alternative form to marriage-based unions. Although the proportion of consensual unions is still higher among the younger age groups and declines with age, the decline is more gradual. Finally, in the fourth stage, cohabitation is an alternative to marriage, i.e., it is distributed equally between the age groups and it is indistinguishable from marriage in its major characteristics.

The dynamics between age and cohabitation follow some regularities and it takes some time before the stage of being an alternative to the marriage-based unions is reached. The realisation of this stage may lie in two demographic developments. First, the younger generations who start cohabiting may remain in this union over their lifetime. In this case, the proportion of cohabiting people among older generations will also increase over time, but more slowly. Secondly, there are individuals who, after the death of a spouse, a divorce or separation, start a consensual relationship and thus add to the increasing number of consensual unions. However, both developments are country specific. For instance, in Britain, whereas post-marital cohabitation has been the most frequent type of cohabitation and continues to increase over time, the number of cohabitations is also increasing among the never-married young population groups (Haskey 2001). Following the idea of developmental idealism, it is particularly interesting to see how cohabitation has been spreading in different age groups in Central and Eastern European countries.

According to Kiernan (2002) and Lesthaege and Moors (2000), the diffusion of cohabitation has reached the final developmental stage when having children in cohabiting unions is customary and an acceptable alternative to childbearing within marriage. To assess the spread of cohabitation over time, the extramarital birth rate can be considered a suitable indicator. Over the last four decades the proportion of children born outside marriage has increased enormously (Figure 1). The process started in the 1970s in Northern Europe. Since the mid-1980s the proportion of extramarital births has also increased in France, the United Kingdom and in some Eastern European countries. Especially rapid was the growth of childbearing outside marriage in Estonia in the beginning of the 1990s. Whilst the extramarital birth rate is a good indicator of the spread of cohabitation across different countries, the recent study of Perelli-Harris et al. (2009) suggested that in the case of childbearing within consensual unions, cohabitation in Europe is not overwhelmingly an alternative to marriage. Non-marital childbearing is very diverse across Europe and does not follow simple geographical boundaries or old geo-political lines. In particular, the countries of Central and Eastern Europe remain very diverse in this respect (Lesthaeghe and Moors 2000; Sobotka and Toulemon 2008).

Moreover, countries with similar levels of prevalence of non-marital fertility have different patterns with respect to the propensity of unmarried couples to get married after the birth of a child (Perelli-Harris et al.2009). It can be concluded from their work that a higher number of extramarital births alone cannot characterise the level of diffusion of cohabitation into a particular country; rather one should concentrate as well on the duration of childrearing within cohabiting unions.

There is a good deal of diversity across European countries in respect of the former marital status of cohabiting partners. Lesthaeghe (1995) showed that the countries with higher levels of divorces tend to have higher levels of cohabitations. Although cohabitation prior to marriage tends to increase the divorce risk (see, e.g., DeMaris and Rao 1992; Berrington and Diamond 1999), on the other hand it could lead to the overall decrease in the divorce rate. If premarital cohabitation is a common practice, then those who get married tend to have more stable unions, while the others split up before marriage (on transitions to marriage or separation, see Moors and Bernhardt 2009). On the contrary, if cohabitation is rare in a country, then the majority of union dissolutions are divorces and the divorce rate tends to be higher. For instance, the divorce rate is higher in some post-communist Eastern European countries such as Hungary and the Czech Republic (Council of Europe 2004: 20), whereas their general cohabitation rate is quite low (Heuveline and Timberlake 2004). As Liefbroer and Dourleijn (2006) showed, the union dissolution expectancy depends on the incidence of unmarried cohabitations in a society. According to Liefbroer and Dourleijn, in European countries with 50 percent of the population in cohabitation the union dissolution risk for married women with and without cohabitation experience does not differ significantly. Thus, we can conclude that diversity still exists across Europe in terms of the extent of cohabitation in a particular country and that the level of its diffusion partly shapes the behavioural outcomes of an individual.

The objective of our further analysis is to reconsider the cohabitation patterns in Europe with a particular focus on the Eastern European countries and using the most recent data. As a general question, we ask how much the diffusion of cohabitating unions into the Eastern European countries matches the typical developmental stages of spreading cohabitation in Western Europe referred to above as a developmental idealism. We also wish to explore which of the cohabitation indicators can play a key role here, and if the Eastern European countries contribute to the convergence or the divergence of family formation patterns in Europe.

The availability of internationally harmonised datasets for more detailed comparative studies on cohabitation has been extremely limited. The exceptions are the studies based on the data from the Family and Fertility Surveys, for instance Heuveline and Timberlake (2004) who applied the data of 17 countries dating back to the mid-1990s. However, since 2002, the European Social Survey (ESS) has offered a free download of survey data with new datasets every second year, providing a powerful source for international comparisons of more than 20 European countries (Jowell et al. 2005, 2007, 2009). We focus on 15 countries across Europe, including one post-communist and three ex-Soviet bloc countries of East and Central Europe and we expand the set of indicators applied in previous studies to describe the phenomenon of cohabitation in a more detailed way.

Because the number of cohabiting individuals in the ESS national samples is relatively small, we combined the second (2004), third (2006) and fourth round (2008) data of the ESS. After merging the files, we calculated the proportion of cohabiting individuals (the adults who were legally married but living with another partner or lived together with a partner without marriage) and other relevant indicators for each country. The total sample for the 15 countries is 85,894 respondents including 9,012 cohabiting individuals and ranging from 132 people in Poland to 1,155 in Sweden.

To measure the spread of cohabitation in a society, we distinguish between the measures of prevalence and duration of cohabiting unions. Regarding prevalence, cohabitation can be widespread in a society in the sense that the majority of younger population lives together before marriage, but the duration of such unions can be short. As for duration, the spread of cohabitation can be seen in its longevity. Based on the current union and marital status, we classify individuals as married or cohabiting. To capture the spread of cohabitation in different birth cohorts, we divide the total sample into four age groups: 18–25, 26–35, 36–45 and 46 years and older. Further, we calculate the proportion of cohabiting unions from all the unions for the age group 26–45. Next we estimate the spread of cohabitation relying on childrearing in cohabiting unions and we calculate the proportion of cohabiting people who have a child or children of 12 years or younger and form this indicator for the 26–35-year age group of the respondents. To test the type of cohabitation, we construct an indicator showing how common post-marital cohabitation is across the countries. We calculate the proportion of the cohabiting individuals who have been previously married relative to the total proportion of cohabiting individuals for each country.

To estimate diffusion, the acceptance and the type of cohabitation, we compute the proportion of individuals who over their lifetime had, at least once, lived together with a partner outside marriage, excluding those currently cohabiting. Using the ESS second round data, we create an additional indicator to measure the duration of currently cohabiting unions (how long the currently cohabiting couples have been living together).

Using the ESS third round data, we calculate the indicator of the duration of premarital cohabitation as a country average (how long on average the couple has been living together before marriage in that country). Because of the relatively short history of cohabitation in some countries, we expand the age range of the respondents and form the indicator based on the age group of 18–45-year-old respondents.

Due to right-skewed distribution of duration of cohabitations, we applied the median instead of the average mean.

In our analyses we face several shortcomings. First, the indicators that we apply in the analysis (duration of cohabitation, the percentage of ever cohabited persons and the proportion of premarital cohabitation) are functionally related to each other and therefore the results should be treated with care. Moreover, from the point of view of the event history analyses, the duration of currently cohabiting unions may exaggerate the expected duration for those starting cohabitation.

The incidence of cohabitation is captured by two measures: firstly the proportion of individuals in different age groups who are currently cohabiting, and secondly the proportion individuals in the age group of 18–45 years (the currently cohabiting excluded) who have ever cohabited. Our data support the results of previous studies that the experience of cohabitation is more widespread in Northern European countries than it is in Western, Central and Eastern Europe. The prevalence of cohabitation is well reflected by the proportion of individuals who have ever cohabited, which is highest in Sweden and Denmark (Table 1). However, if we consider the prevalence of cohabitation among individuals in the age group 18–45, we find that in addition to Sweden and Denmark, Estonia also has more than 60 percent of people who have entered cohabitation. Cohabitation occurs most frequently in the 26–35-year group, e.g., in Sweden almost 43 percent of the age group 26–35 are cohabiting; in Estonia, Slovenia, Belgium, Spain and Portugal the proportion of 26–35 years old cohabiting individuals exceeds the total proportion of those who have ever cohabited. In these countries the proportion of cohabiting individuals has increased in younger age groups, indicating the tendency of further diffusion of cohabitating unions. In those countries where the proportion of cohabiting people among the population is only 5 percent or lower, the age pattern is not so distinct and there is no obvious trend. However, we can assume that cohabitation does appear as a prelude or a substitute to the first or second marriage.

Despite a relatively short time span, the data reveal the tendency of continuing diffusion of cohabitation across Europe (Table 1). The growth of cohabiting unions has been especially high in Eastern and Southern European countries. For example, during the period of 2004–2008, it doubled in Poland as a relatively new phenomenon. The rate of increase of cohabitation has slowed down in Northern and Western European countries, where it has been practised more widely since the 1960s and 1970s.

In those countries with a low prevalence of cohabitation, it is usually entered into following the end of a marriage. This trend is most characteristic of former Soviet-bloc countries such as Poland and Slovakia but it is also discernable in Portugal. By contrast, in the Nordic countries where cohabitation has been widespread for a longer time, the proportion of previously married people is lower, remaining under 25 percent of the total.

In Western European countries, as well as in post-communist Estonia, cohabitation is not as common as in the Nordic countries, but the proportion of people for whom the current cohabitation is the second experience of living as a couple is higher. Finally, in Spain and Slovenia, the cohabitation rate is low and those cohabiting represent more often the ‘never married’ population group. On the one hand this picture indicates the linkage between the cohabitation level in a society and the marital background of the cohabiting persons; on the other hand, it indicates some differences which can be traced to the socio-historical background of the family formation in different countries (see, e.g., Dittigen 1995; Kalmijn 2007). In other words, in the countries where cohabitating unions are not yet so widespread, people do not practise cohabitation as a prelude or alternative to a first marriage but as an alternative to a second marriage. However, there are two exceptions – Spain and Slovenia. Thus, according to marital background, cohabiting individuals across European countries are continuing to be a very diverse group, and there is no direct linkage between the divorce rate and the spread of cohabitation at the aggregated level. The mutual relatedness of cohabitation and divorce is rather indirect and has a long-term influence in the societies where the so-called second demographic transition started earlier; cohabitation starts to spread as an overall pattern and not only as a post-marital informal union.

In the Nordic countries where cohabitation is most extensive, the proportion of the cohabiting people who have been previously married is the smallest. We suggest that here the extent and the form of cohabitation are interrelated. According to Kiernan (2001): 2), based on British data, post-marital cohabitation was the most prevalent form of cohabitation in the 1950s and 1960s. This allows us to conclude that the proportion of post-marital cohabitating unions illustrates the duration of the phenomenon of cohabitation in a society. The countries with a high number of post-marital cohabitations are those where cohabitation is a new phenomenon. By contrast, the countries where the evidence of post-marital cohabitation is lower are those where the tradition of cohabiting without marrying is historically longer.

Having children in cohabitating unions has been considered the most important indicator of the acceptance of cohabitation in a society (see e.g., Wu 2000: 172; Raley 2001). In our study, we calculate the presence of children aged 12 years or younger in a cohabiting union, taking into account not only children of the union but children brought into the union. This criterion is necessary, especially in those countries where at least half of the cohabiting unions take place after the end of first marriages. Some studies have considered Sweden as the only society where cohabitation has adopted a childbearing function (Heuveline and Timberlake 2004) but the recent studies of Sobotka and Toulemon (2008) and Perelli-Harris et al. (2009) have shown that childbearing intensities are still higher in the marriage-based unions compared to cohabitations. Our results also confirm that in Sweden the proportion of cohabiting couples in the 26–35 age group who are raising children under 12 years old is less than 50 percent (see Table 1). In Northern European countries with high cohabitation levels in general, the proportion of couples who raise children in cohabiting unions is much lower compared to the proportion of the married couples with children in the same age group; although cohabitation is most widespread in these countries, the incidence of having/rearing children in cohabiting unions is not as widespread as for married couples. For example, in Denmark 40 percent of cohabiting couples have children compared to 91 percent of the married couples in the same age group. The fact that the level of cohabitation and extramarital birth rates is high but the evidence of children within cohabiting unions compared to married ones is considerably lower may relate to an inversion of the traditional process of family formation – i.e., that marriage is contracted after having children. This result supports Kiernan's (2001) findings that even when the proportion of non-marital births is quite high, there are many cases where parents will marry later. For instance, in Switzerland, Austria, Italy and Sweden around 70 percent or even more marry during the 5 years after the birth of a child (Kiernan 2001: 17).

Our analyses based on the more recent data reveal that the childbearing within cohabiting unions is most common in Estonia and France where more than 60 percent of cohabiting couples in the age group of 26–35 have children. In the last years, Estonia has almost the highest rate of extramarital births in Europe (58.1 percent of children were born out-of-marriage in 2007). Our data show that Estonia does not fit into the classification of Perelli-Harris et al. (2009) whereby in the Eastern European countries, cohabiting couples more often legalise their union before the birth of a child. In the case of Germany and Switzerland, the incident of childrearing among cohabiting couples is relatively low; cohabitation is seen as a prelude to marriage and children are planned to be born into a marriage (see Perelli-Harris et al. 2009) while cohabitation is seen mostly as a childless union (Nave-Herz 1999; Le Goff 2002).

We can also distinguish those countries where the proportion of cohabiting people is relatively low but, of those, quite a large proportion have children. This is the case of some post-communist countries and Portugal, and can be explained by family background factors: people start cohabiting after divorce, separation or death of a spouse, and children may come from former marriages.

Although the indicators of duration of cohabitation slightly overstate the real duration of the unions, we still can conclude that the duration of premarital cohabitation is strongly related to the overall prevalence of cohabitation in a society. In those countries where cohabitation is most widespread the average duration of premarital cohabitation is longer. However, the duration of current cohabitation is not directly linked to the diffusion of cohabitation in the country. Our analysis revealed four country groups.

In the first country group (Estonia, Slovenia, Slovakia and Belgium) premarital cohabitation is short (the median duration of premarital cohabitation is less than a year) and those who view their cohabitation union as an alternative to marriage have been living together significantly for longer (the median value is up to 6 years). In the case of Eastern European countries, cohabitation as an alternative to marriage is still a relatively new phenomenon and therefore the overall duration of premarital cohabitation is short for two reasons. Firstly, the older people who started their unions in the 1970s and 1980s married relatively quickly. The reason was often that marriage involved some extra rights or privileges, e.g., it led to better prospects of getting employment, permission to buy a car or publicly owned accommodation. Secondly, in the 1990s the duration of cohabitating unions increased, characterised in our study by a relatively long duration of current cohabitations. The mean age on first marriage in Eastern European countries has increased in recent years but it is still low compared to Western European countries.

In the second group of countries (the Nordic countries – Sweden, Finland and Denmark) both types of cohabitation are relatively long in duration and cohabitation has longer traditions compared to the other countries under study. The mean age at marriage in these countries is the highest in Europe.

In the third country group (Switzerland, Germany, the United Kingdom, Norway and France) the duration of currently cohabiting unions is shorter than in the previous groups, but the duration of premarital cohabitation is longer than in Eastern European countries. Finally, in Eastern and Southern European countries (Poland, Spain, Portugal) cohabitation is only an emerging phenomenon, with the shortest durations of both types of cohabitation.

By analysing the spread of cohabitation in Europe in 2004–2008, this article attempts to contribute to the ongoing debate of divergence and convergence of family patterns across the continent. Our results showed the continuing diffusion of cohabitation across Europe, and in this respect they support the idea of convergence. Classifying countries into different groups by cohabitation level and related indicators confirms also the idea of developmental stages of cohabitations; this supports the universalistic views of Kiernan (2001) and Prinz (1995) that the trends of cohabitating unions follow similar paths in different countries. This is especially true if we take into account that the transformation of family forms (the second demographic transition) started in the Nordic countries earlier than in the rest of Europe. The described theoretical linkage between post-marital cohabitation and cohabitation level in a country supports this idea as well. However, some exceptional cases like Spain weaken this conclusion. Moreover, the differences in having children within cohabitations encourage us to review this standpoint. We partly agree also with Reher (1998): 221), according to whom there may be evidence for a convergence in the external indicators of family life. But this convergence will not undermine the deep disparities that have always characterised families in different regions and cultures across Europe (see also Kalmijn 2007).

Following the framework of developmental idealism proposed by Thornton and Philipov (2009) we may consider that the countries are continuing to move toward the Nordic family system but cultural traditions, religion, policy and socio-economic situations determine the speed and the extent of this movement. Even if the recent trends in family formation indicate that the post-communist countries are moving towards the Western European family formation patterns, we cannot totally disregard the so-called path dependency – the individual historical trajectories of societies, while different in each case, will contribute to the specific contours of the present and the future (Reher 1998: 221). This particularly applies to Slovenia and Estonia. These two countries are not following the universal Western European trend nor are they similar to the other Eastern European countries, although they might have crossed or passed through some similar stages. Estonia in particular shows a pattern where cohabitation is as common among younger age groups as in Northern European countries and the proportion of cohabiting couples with children even exceeds that of Sweden and other Nordic countries. However, it is not the case in other countries of Eastern Europe. Thus, when we concentrate more precisely on the type of cohabitation and its characteristics, then the idea of divergence brought out by Sobotka and Toulemon (2008) finds support.

But still, which out of the indicators captures best the spread of cohabitation – the incidence, duration and previous marital status of the cohabiting partners or childbearing within cohabitating unions? It seems that, in addition to incidence and prevalence, childbearing within cohabitation should be considered as the relevant factor for family changes, especially in those countries where cohabitation has spread very rapidly during the recent decades. Although these indicators are internally related, it is not always the case of linear associations, e.g., childbearing and the incidence of cohabitation, or the incidence of cohabitation and the duration of cohabiting unions. Moreover, we are aware that the cross-sectional data are not the best solution to the challenge of measuring the spread of cohabitation. As shown by Matysiak (2009), the cross-sectional data underestimate the incidence of cohabitation, especially in the relatively early phase of its diffusion. Knab and McLanahan (2007) showed that cohabitation rates are lower if the survey questions ask about the current status whilst in retrospective studies the number of people who admit their cohabitation is actually higher.

In conclusion, our study showed there is evidence of the developmental stages that cohabitation as a social phenomenon goes through by diffusing into a country. However, if individual indicators of cohabitating unions are added to the analysis, then dissimilarities become apparent. Consequently, when cohabitation is adopted in the country as a normative behavioural value, it begins to develop towards higher divergence. If the diffusion is a recent phenomenon in a country, the divergence of the cohabiting unions is higher and combines features from its typical developmental stages.

Kairi Kasearu is a PhD candidate in sociology at the University of Tartu. Her research interests are related to the field of family sociology: family formation, social networks and intergenerational relations. She has recently published on unmarried cohabitation in the following journals: Social Indicators Research and Trames.

Dagmar Kutsar is an associate professor at the Institute of Sociology and Social Policy at the University of Tartu. Her research interests are: family, childhood and welfare research and policies (including poverty and social exclusion/inclusion, childhood relative deprivation and poverty, changing family structures and family decision-making), social indicators and social reporting (problems of international comparability), science policy. Her most recent publication is Internationalisation of Social Sciences in Central and Eastern Europe. The ‘Catching Up’ – A Myth or a Strategy? (Routledge, 2010), co-edited with Ilona Pálné Kovács.