Acta geographica Slovenica, 62-2, 2022, 7–22 THE CHANGE IN EDUCATIONAL ASSORTATIVE MATING IN SERBIA AND SLOVENIA, 1970–2020 Damjan Bakić, Vera Gligorijević Rectorate of the University of Belgrade, Serbia. V E R A G L IG O R IJ E V IĆ 62-2_acta49-1.qxd 13.12.2022 12:07 Page 7 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 DOI: https://doi.org/10.3986/AGS.10496 UDC: 316.36-057.8(497.11:497.4) COBISS: 1.01 Damjan Bakić1, Vera Gligorijević1 The change in educational assortative mating in Serbia and Slovenia, 1970–2020 ABSTRACT: This paper explores marital matching patterns from the perspective of the partners’ educa- tional attainment, focusing on the link between gender asymmetry in education and educational hypergamy. In order to assess to what extent the tendency for women to marry men of higher educational status is related to the educational gender gap in Serbia and Slovenia, we calculate an index for women’s educational advantage, and an index for the prevalence of educational hypergamy. Our results confirm the following: the growth of education is associated with an increase in female educational advantage; the relationship between female educational advantage and educational hypergamy is strongly negative; and there are no significant differences in assortative mating patterns between Serbia and Slovenia. KEYWORDS: educational assortative mating, education, marriage, Serbia, Slovenia Sprememba v sklepanju partnerskih zvez z vidika izobrazbe v Srbiji in Sloveniji med leti 1970 in 2020 Ta članek raziskuje vzorce sklepanja partnerskih zvez vidika izobrazbe partnerjev. Osredotoča se na poveza- vo med spolno asimetrijo v izobraževanju in izobraževalno hipergamijo. Da bi ocenili, v kolikšni meri je težnja žensk po poroki z moškimi z višjo izobrazbo povezana z izobrazbeno razliko med spoloma v Srbiji in Sloveniji, smo izračunali indeks izobrazbene prednosti žensk in indeks razširjenosti izobraževalne hipergamije. Naši rezultati potrjujejo, da je rast izobrazbe povezana s povečanjem izobrazbene prednos- ti žensk, da je razmerje med izobrazbeno prednostjo žensk in izobrazbeno hipergamijo močno negativno in da med Srbijo in Slovenijo ni bistvenih razlik v selektivnih vzorcih sklepanja partnerskih zvez. KLJUČNE BESEDE: izobrazbeno pogojeno sklepanje partnerskih zvez, izobrazba, poroka, Srbija, Slovenija The article was submitted for publication on December 9th, 2021. Uredništvo je prejelo prispevek 9. decembra 2021. 8 1 University of Belgrade, Faculty of Geography, Belgrade, Serbia damjanbakic93@gmail.com (https://orcid.org/0000-0003-0732-1942), vera.gligorijevic@gef.bg.ac.rs (https://orcid.org/0000-0003-1658-674X) 62-2_acta49-1.qxd 13.12.2022 12:07 Page 8 1 Introduction Even though Serbia and Slovenia differ in socio-economic aspects, both countries are facing similar demo- graphical issues, and just like many other Eastern European countries they are experiencing a postponement of marriage and childbearing to higher ages and decline of fertility towards very low levels (Josipovič 2003; Penev 2010; Nikitović 2015; Josipovič 2016; Graovac Matassi and Talan 2021). Marriage rates, which are still considered to be an important demographical reproduction factors (Devedžić 2004; Lerch 2017; Rašević and Vasić 2017), are greatly influenced by the composition of the marriage market. This composition is heavily misbalanced in both countries due to the insufficient supply of the partners with desired socio- economic characteristics. One of the mostly desired characteristics is partner’s education, which determines social status. An »ideal« partner should be equally well or better educated, which is a requirement that cannot be fulfilled that easily today, due to the unequal educational coverage between men and women. Partner’s education is a key both for individual economic behaviour and for the behaviour of the unions, because the marriage stability depends on the suitable partner combination. The marriage stability fur- ther influences marriage outcomes, such as parenthood or divorce (Esping-Andersen and Billari 2015; Goldscheider, Bernhardt and Lappegård 2015). The stability is more pronounced if the partners share com- mon values which are today formed mostly through education, given that the influence of other institutions (e.g. family or religion) has declined over the time. This is the reason why the research of »educational assor- tative mating« can contribute to a better understanding of the effect the social changes have on the marital level. Moreover, understanding marital level leads to a better understanding of the changes in the fertili- ty level. Since the emergence of surpluses of educated women (Vincent-Lancrin 2008; KC et al. 2010), more marriages where the woman is more educated than the man have been observed (Esteve, García-Román and Permanyer 2012; Van Bavel 2012). Therefore, this work focuses firstly on the educational misbalance between the genders, and afterwards tries to answer the question if this misbalance can be connected to the pattern changes in educational assortative meeting. In the end, the question whether differences in educational assortative mating between Serbia and Slovenia exist is tackled, and to what extent those dif- ferences can be explained by (post) Yugoslavian socio-economic context. 2 Theoretical frameworks Assortative mating is based on the partner search theory which assumes that individuals have preferences for partners with certain socio-economic characteristics and that there exists a market where those pref- erences can be realized (Lewis and Oppenheimer 2000). Socio-economic attributes such as income, wealth, occupation, and education are the most cited ones, because they influence the status of an individual and of the couple (Kalmijn 1994), also from of all socio-economic features, education is an attribute being explored pretty often, because it represents one of the main components of the social stratification (Petrović 2011; Bobić 2017) and it is an important part of an individual’s lifecycle (Mare 1991). Partner search theory accounts for differences in the preferences of men and women, and those dif- ferences were addressed for the first time in the model of household specialization and division of labour introduced by Becker (1973). According to this model, when forming households, couples will exploit the gains from trade by having one spouse specialized in market work while the other specialized in house- hold work. Household specialisation enhanced the emergence of educational hypergamy, i.e. marriages where a husband would be more educated than a wife (Blossfeld 2009; Schwartz and Han 2014). Women’s ability to earn caused a change in men’s preferences, who now began favouring the economic character- istics of their partners (Oppenheimer 1994; Lichter et al. 1995; Torr 2011). This resulted in an ever-increasing number of marriages with equally educated partners (Sweeney 2002; Schwartz and Mare 2005). Once edu- cation became gender-neutral, and once educated women outnumbered educated men, a new pattern was born – husbands marrying up. This pattern emerged as an answer to, i.e. acknowledgment of, the new demo- graphic reality among younger cohorts, especially of the highly educated populations of both genders (Grow and Van Bavel 2015). This composition of marriage market is influenced by social macrostructures (Blau 1977), i.e. by the relative group size entering the marriage market (De Hauw, Grow and Van Bavel 2017; Chiappori 2020). Acta geographica Slovenica, 62-2, 2022 9 62-2_acta49-1.qxd 13.12.2022 12:07 Page 9 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 It also explains the process of so-called »squeezing out« of less competitive individuals from the marriage market (marriage squeeze), which occurs in case of a significant misbalance towards a certain gender or socio-economic attribute (Schoen 1983; Van Bavel 2012). Individuals experiencing difficulties in finding a suitable partner (equally or better educated) due to a lack of candidates with their preferred attributes generally resort to one of two strategies. The first is the strategy of extended search, where individuals con- tinue searching for a partner with their preferred attributes for longer than they normally would, at the risk of weakening their own attributes and increased costs that can arise over time. In the second case, they resort to the criterion of selection. Falling back on the first strategy leads to an increased number of single people among the young population, whereas adopting the second strategy increases the number of marriages which are not aligned in terms of individual preferences (Schoen 1983; Oppenheimer 1988). Two groups are specifically prone to the education-specific mating squeeze: highly educated women and non-educated men. From the perspective of highly educated women, the supply of suitable, highly educated men are limited by the bounded number of university graduates. From the perspective of une- ducated men, the fact that there are more female graduates at every level of education implies that they are facing a decreasing number of potential partners, as women tend to choose a partner that is at least as educated as they are. However, a more educated wife is not an obstacle to marriage, as demonstrated by Esteve et al.’s (2016) results from 120 countries with differing socio-economic contexts – the marriage market accepts an increas- ing number of highly educated women thanks to suitably flexible gender norms, though the second phase is taking place very slowly (England 2010; Esping-Andersen and Billari 2015). While Second demographic transition has a negative narrative regarding the changes within the families, gender revolution foresees positive changes in the families (greater number of marriages, greater marriage stability, smaller number of divorces and higher fertility). Those positive changes shall happen due to the changes in the currently existing structures and the prevalence of the structural factors over the ideological ones (Goldscheider, Bernhardt, and Lappegård 2015). Other authors agree that the dominance of the gender roles reflects the chances from the social environment and that they change under the influences of structural misbalances (West and Zimmerman 1987; Blossfeld 2009; Šobot 2012; Bobić 2017). Speaking of the big structural changes, such as globalization or very strong increase of education level, gender attitudes formed under the old struc- tures can survive »if people are not well prepared by education or experience to function comfortably and successfully in the new gender structure« (Goldscheider, Bernhardt and Lappegård 2015, 218). Taking edu- cational assortative mating as an example, studies have shown that in a  western-oriented context, non-traditional educational hypogamy couple have a lower divorce risk compared to educational homogamy ones, given that they are represented with an abundancy greater than 50%, i.e. given that the norms are accept- ed within a broader local community (Theunis et al. 2018). Starting with theoretical background and previous results, we postulate the following hypotheses: 1. Development of the education leads to the creation of a gender-based misbalance in favour of edu- cated women. 1a. This misbalance is mostly pronounced when talking about highly educated people. 2. The prevalence of women within the population with tertiary and secondary education leads to a decline in educational hypergamy and a rise in educational hypogamy. 2a. Highly educated women are the ones that are mostly susceptible to hypogamous marriages. 3. Patterns in educational assortative mating are different in Serbia and Slovenia. 3a. Educational hypergamy declines more rapidly in Slovenia. Even though the education is highly relevant to fertility and mortality (Rašević and Vasić 2017; Marinković 2018; Mirić 2019), not many papers investigating the consequences on partner selection patterns have been published. To our best knowledge, this is the first paper of that kind concerning Serbia and Slovenia and it will make its contribution to the knowledge fond in the area of educational assortative mating, where- as the results specific to Serbia and Slovenia will be joined to the ones (De Hauw, Grow and Van Bavel 2017) referring to the other European countries. 2.1 Former Yugoslavia context Even though a very intense rise in education could be noticed throughout the whole Yugoslavia after WW2 (Breznik 1991), this process commenced a little bit earlier in Slovenia, it was faster compared to central 10 62-2_acta49-1.qxd 13.12.2022 12:07 Page 10 Serbia, but also mainly focused on mid-school education. This can be corroborated by the fact that the differences regarding the mid-school education levels between those two countries in the second half of 20th century were very much noticeable (in 1981. 34.5% of the Slovenian population had finished mid- school, whereas this number was only 25% in central Serbia), in contrast to the dynamics of higher education (higher education percentages from 1981: Serbia 6.5%, Slovenia 6.2%). However, Slovenia has gained a sig- nificant advantage over the last decades: in 2020, Slovenia had 47% of population with higher education, compared to Serbia which had 33%. During the last three decades, the variations in the population growth rates and therewith connected structures showed an uneven speed of modernization between those two countries, as well as cultural differences and particulars of the socio-economic development and politi- cal system (Sardon 2001; Kuhar and Reiter 2010; Penev and Predojević-Despić 2019; Istenič, Ograjenšek and Sambt 2017). Those particulars are reflected today through differences in the gender structure of the educated people. Expansion of the higher education, and the ever-increasing number of women attend- ing universities (a trend that started at the beginning of the 1980’s in Slovenia and in the 1990’s in Serbia) (Šircelj 2007; Jovanović-Gavrilovic and Radivojević 2017), lead to an increase in percentage of women with- in highly educated population. This misbalanced was observed earlier in Slovenia, probably due to the Second Demographic Transition (Sobotka 2008; Bobić and Vukelić 2011) and the proximity of the European labour market (Kuhar and Reiter 2010; Josipovič 2018; Dobrotić and Stropnik 2020). A comparative analysis of the changes in educational assortative mating in Serbia and Slovenia is jus- tified by several arguments: looking at the similarities in partner selection patterns in two countries that used to be on opposing sides throughout the history when it came to the socioeconomic and demographical development is rather interesting. Furthermore, would be intriguing to see if Serbian society changes as a consequence of a dynamic development of the education, regardless of the of the slow economic devel- opment and low life standard (this change can be measured through the acceptance level of atypical marriage forms, where women are more educated than men). Lastly, it is compelling to understand the extent to which Serbia is lagging behind Slovenia, where economic prerequisites for social changes are fulfilled to a greater degree. 3 Data and methods The main source of data regarding the gender-educational structure on the marriage market were cen- suses of the Socialist Federal Republic of Yugoslavia, censuses of the Republic of Serbia and censuses of the Republic of Slovenia, as well as the Labour Force Survey data that are available on the website of Eurostat. They were used to show the change in gender and educational structure of the marriage market. Educational attainment is divided into three categories: primary (low) education (ISCED 1-2), sec- ondary (medium) education (ISCED 3), and tertiary (highly) education (ISCED 5-8). Our main data source for investigation educational assortative mating is the »Demographic Statistics« publication which was published on a yearly basis by the Federal Office for Statistics until 1990, and since 1990 by the Office for Statistics of Republic of Serbia. The publication provides data about gender distri- butions and educational attainments of a bride and a groom, which we used to classify all marriages from a given year into three groups: a) marriages where partners possess the same level of education (educational homogamy); b) marriages where the husband is more educated than the wife (educational hypergamy); c) marriages where the wife is more educated than the husband (educational hypogamy). This data is used to create the 50-year time series for Serbia, together with the data from the European Social Survey for the years 2010 and 2018 (only for Slovenia). Further on, based on those data we were able to collect information about which marriage types are the most common ones for women, for men, for mid-school educated population, for highly educated peo- ple and for population without education. The data of educational assortative mating are limited only to legally registered marriages, as there is no data regarding the educational structure of the people living in cohabitations. It is known that the cohab- itations make up to 5.5% of the total number of families in Serbia, and that approximately 6% of the population older than 15 live in cohabitations. The frequency of cohabitations is the same for all types of education- al profiles (5.5% when it comes to the highly educated population and 7.5% among the people with primary school only). Acta geographica Slovenica, 62-2, 2022 11 62-2_acta49-1.qxd 13.12.2022 12:07 Page 11 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 According to the results of the previous papers, there are no clear evidences which could prove that the patterns of the marriage section differ than the cohabitation patterns. Only in couple of cases, a slight- ly lower level of educational homogamy within the cohabitations has been recorded (Blackwel and Lichter 2000; Hamplova 2008; Schwartz 2010; Esteve, McCaa and 2013). Analytical strategy was as follows: comparing two countries, we have shown the dynamics of the edu- cation development of the population by showing the increase in the share of highly educated people, and we have shown the differences in number of women and men studing at the universities. The prevalence of women at all educational levels in Serbia and Slovenia was analysed using the Index of female educa- tional advantage (F-index), which is defined by Esteve, García-Román and Permanyer (2012) as follows: (1) where pf and pm stand for the share of women and men within a certain educational category, and the indices 1,2 and 3 denote different educational categories. F-index shows the probability that when ran- domly choosing a woman and a man (with different degrees of education), the woman would be the more educated one. This index takes all education levels into consideration, from the primary to the tertiary education, and lies between 0 and 1. If F = 0, there is no woman whose educational attainment is higher than or equal to that of any man, and if F = 1, the reverse is true. If the gender distribution is symmetric, the F-index equals 0.5. If F-index is between 0.5 and 1, it follows that the women are more educated with- in the considered group (Esteve, Garcia-Roman and Permanyer 2012). Based on the same source, we were able to calculate the prevalence index of the educational hyper- gamy (H-index), in order to understand, which type of heterogamous marriages is more common. The equation is as follows: (2) where A and B are the numbers of educational hypergamic and educational hypogamic couples, respec- tively, and the operator is the natural logarithm. If the H-index equals 0, it follows that the number of educational hypergamy and educational hypogamy couples is the same (A = B). If the H-index is higher than 0, the number of hypergamous marriages prevails (A > B). In the case of educational hypogamy preva- lence, the H-index is negative. In the end, the Pearsons’s correlation coefficient between the F- and H- indices has been calculated, in order to show the nature of the dependency between the prevalence of women in education and prevalence of the educational hypergamy. 4 Results The results are presented in three sections. The first one shows the development of the education and the greater reach of education within the female population, compared to the male population. In the second part we examine which patterns of educational assortative mating are prevalent in Serbia and Slovenia, and we track its change over time. The third part describes the correlation between the female majority in education and changes in the educational assortative mating. 4.1 The development of the education and female education advantage Over the last 20 years, education in Serbia and Slovenia expanded significantly. The extent of this expan- sion can be seen from Table 1. The dynamics of this process was largely similar in Serbia and Slovenia, even though Serbia shows a slower trend, which is a direct consequence of Slovenia’s faster economic growth shortly after joining the EU. The expansion of higher education as measured by the share of highly educated individuals within the age group 25–34, occurred in the 2000s, as this parameter increased by 14 percentage points in Slovenia between 2002 and 2011. In Serbia, this phenomenon was much slower and is visibly lagging behind Slovenia, but the share of highly educated individuals within the age group 25–34 showed a reasonable increase dur- ing the last decade. 12 62-2_acta49-1.qxd 13.12.2022 12:07 Page 12 Acta geographica Slovenica, 62-2, 2022 13 The increase in share of the highly educated among the 25–34 population was accompanied by an ever- increasing number of female students. Female majority in higher education occurred in 1980 in Slovenia, and in 1990 in Serbia. The biggest surplus of women within the student population was recorded in 2010 in Slovenia (63% of students were women), whereas the peak in Serbia occurred in 2005 (Figure 1). In the last decade, the share of women in higher education has been nearly constant around 60%. This implies that, in recent years, there are more highly educated women reaching the optimal reproductive age than there are men. Apart from the share of women in the higher education, there are some indicators of the changes in the educational composition, which could provide a more complete picture of a given country. One such indicator is the measure of female educational advantage (F-index). Figure 2 shows the F-index trend, i.e. the increase of F-index during the time period 1970–2019. The increasing tendency of the F-index corre- sponds to the increase in the level of higher education, especially in Slovenia. In that context, one can claim Table 1: Share of individuals with primary, secondary and tertiary education in 25–34 population. Country Year Education Primary Secondary Tertiary Serbia 2002 22.8 63.3 13.9 2011 14.1 62.7 23.2 2019 10.5 56.4 33.1 Slovenia 2002 15.1 65.5 19.4 2011 5.4 61.2 33.4 2019 4.1 48.9 47.0 Source: Statistical office of the republic of Serbia – The population and households of Serbia according to the 2002 census; Eurostat databases – Population by sex, age and educational attainment level 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 2019 10 20 30 40 50 60 70 SerbiaSlovenia % Figure 1: Female students enrolled in tertiary education as the share of all those in tertiary education, 1970–2019. 62-2_acta49-1.qxd 13.12.2022 12:07 Page 13 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 14 that the changes in the higher education are the main factor that influenced the educational structure of the population (and not changes in other levels of education). Even though the higher education was ris- ing rapidly in Serbia as well, its effect was not completely the same in Serbia and Slovenia. Figure 2 clearly shows that the probability of female advantage in education is lower in Serbia, as the overall educational structure is still under the influence of middle-school education, where female dom- ination is lacking. In the recent times, the probability that a randomly chose woman compared is more educated to that of a randomly chosen man, is approximately the same in Serbia (0.5:0.5), whereas this number is around 0.57 in Slovenia. This implies that women gained an educational advantage over men even when taking all educational levels into consideration. 4.2 The surplus of highly educated women and educational assortative mating The female majority in education influenced the gender distribution of the educated population within the optimal marital age (25–34 years of age). Figure 3 illustrates this phenomenon. This is especially notice- able when it comes to the highly educated population, where the share of highly educated women tops the share of highly educated men by approximately 7–14 percentage points within this age group. Female majority in education induced changes in educational assortative mating patterns, as well. Those changes are reflected through an educational hypergamy decrease and an educational hypogamy increase, and a constant share of homogamous marriages compared to 1970 (Table 2). Educational homogamous marriages have always been dominant compared to all other types of marriages, both in Serbia and in Slovenia. In 1970, their share in Serbia was 72% and in Slovenia it was 66%. In Serbia, this value has not changed in 2019, but in Slovenia it increased slightly, to 73%, in 2018. With respect to educational heterogamy, the share of marriages where men are more highly educated than women have significantly declined (i.e. edu- cational hypergamy decreased), from 24% to 10% in Serbia, and from 26% to 8% in Slovenia. At the same time, the number of marriages where women are better educated than men went up (educational hypogamy) – from 3% to 16% in Serbia, and from 7% to 18% in Slovenia. The time frame when those changes took place differs though – the intense decrease in educational hypergamy in Slovenia was recorded during the 1990s, 1971 1991 2011 2019 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 Serbia BalanceSlovenia Figure 2: Index of female educational advantage (F index). 62-2_acta49-1.qxd 13.12.2022 12:07 Page 14 whereas in Serbia it happened with a delay of one decade, which corresponds to the delay in the higher education gender gap reversal relative to Slovenia. A powerful indicator of the changes in educational assortative mating is the prevalence of education- al hypergamy, which can be seen based on the behaviour of the H-index. Figure 4 shows a decline in H-index values in Serbia and Slovenia from 1970 to 2018/19. Declining values of educational hypergamy overlap (temporally) with the increasing trend of female majority in education. It follows that the tendency of women to marry an equally well or better educated partner is temporarily stalling. That tendency has been replaced by educational hypogamy. In Slovenia, educational hypogamy exceeded educational hypergamy more quick- ly than in Serbia, which is in line with the faster dynamics of female majority in education. Acta geographica Slovenica, 62-2, 2022 15 Table 2: Share of educational homogamous, hypergamous, and hypogamous marriages, 1970–2018/19. Year Serbia Slovenia Educational Educational Educational Educational Educational Educational homogamy hypergamy hypogamy homogamy hypergamy hypogamy 1970 72.5 24.1 3.4 66.6 26.2 7.2 1980 68 25 7 64.4 21.8 13.8 1990 65.8 24.7 9.5 65.9 18.5 15.6 2000 69 19.2 11.8 –* – – 2010 71.6 12.9 15.5 71.8 11.2 17.0 2018/19** 72.7 10.6 16.7 73.3 8.6 18.1 *missing data ** due to the limitations of different types of sources, data for Serbia from 2019, and for Slovenia from 2018 Sources: Federal Statistical Office of Yugoslavia- Yearbook of Demographic Statistics (1970, 1980 and 1990; Statistical office of the Republic of Serbia – Yearbook of Demographic Statistics (2000, 2010 and 2019); European Social Survey – databases for a Round 5 (2010) and 9 (2018) 2011 2012 2013 2014 2015 2016 2017 2018 2019 10 0 20 30 40 50 60 70 Serbia BalanceSlovenia Sh ar e o f w o m en i n h ig h l y e d u c a t e d in d i v id u a l s a g e 25 34 ( % ) – Figure 3: Surplus female in optimal marriage age (25–34). 62-2_acta49-1.qxd 13.12.2022 12:07 Page 15 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 16 –1 –0.5 0 0.5 1 1.5 2 2.5 1970 1980 1990 2000 2010 2018/19 Serbia BalanceSlovenia 1970 1980 1990 2000 2010 2018/19 –1 –0.75 –0.5 0 0.25 –0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 Serbia medium Slovenia medium Serbia low Slovenia low Serbia highly Slovenia highly Balance Figure 4: The prevalence of hypergamy (H index), by countries. Figure 5: The prevalence of hypergamy (H index), by education level and country. 62-2_acta49-1.qxd 13.12.2022 12:07 Page 16 Acta geographica Slovenica, 62-2, 2022 17 0 20 40 60 80 100 % SRB 1970 SRB 2019 Women, high education SLO 1970 SLO 2018 Homogamy Hypogamy 0 20 40 60 80 100 % SRB 1970 SRB 2019 Men, low education SLO 1970 SLO 2018 Homogamy Hypogamy 0.3 0.35 0.4 0.45 0.5 0.55 0.6 F I– ndex –1.5 –1 –0.5 0 0.5 1 1.5 2 H In d e x – slope = 9.5553 y intercept = 4.2546 R squared = 0.881 p value = 0.000553 – – – – Figures 6a and 6b: Education assortative meting for tertiary educated women and low educated men. Figure 7: Correlation between female educational advantage (F-index) and prevalence of hypergamy (H-index). a b 62-2_acta49-1.qxd 13.12.2022 12:07 Page 17 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 Figure 5 shows the prevalence of educational hypergamy for different categories of people (low, medi- um and high educated population). The H-index of those individual groups follows the trend seen with the total population: it declines. Educational hypogamy exceeded education hypergamy when we speak about highly and medium educated people. In the case of low educated population, this event has not occurred yet, i.e. the number of hypergamous marriages prevails. In Slovenia, the H-index became negative almost 20 years earlier that in Serbia, even in the category of highly educated people, which is plausible given the dynamics of female majority in education. Figure 6a and 6b reports change of the patterns of educational assortative mating for the high edu- cated women and for the low-educated men. Both groups show similar trends: educational homogamy decreases in both countries, while educational hypogamy increases, which corresponds to the surplus of women. Highly educated women and low educated men are the ones who are mostly affected by the new gender composition of the educated population. Highly educated men benefit from an abundance of women with tertiary educational attainment. For them, an excess of highly educated women is associated with stronger educational homogamy in Serbia and Slovenia (in 80% of cases). Educational hypergamy maintained its status as an educational assortative mating pattern only when it comes to women with primary education only. Today, approximately 30–35% of women with primary education marry more highly educated men. In Figure 7 the negative linear relationship between female educational advantage and the prevalence of educational hypergamy is clearly visible, as higher values of the F-index are associated with lower val- ues of the H-index. The correlation measure also confirms this (R-squared = 0.881, p = 0.0005), even with the 95% confidence intervals considered. This suggests that the two index are indeed negatively connected. 5 Discussion In looking for an answer to the question »who marries whom« in Serbia and Slovenia, the main focus shift- ed to education, as increased participation in education, for both men and women, was the most important social trend of the 20th century. The dramatic expansion in educational opportunities in Serbia and Slovenia was accompanied by a significant gender gap in higher education, while in the last decade, the share of women in higher education has been nearly constant around 60%. Greater numbers of high- ly educated women have led to changes in the conditions governing the marriage market (Hypothesis 1). This implies that, in recent years, there are more highly educated women reaching the optimal reproductive age than there are men. Today, the ratio of highly educated women to highly educated men in the 25–34 age group, is 57:43 in Slovenia, and 59:41 in Serbia (Hypothesis 1a). Our analysis suggests that increases in women’s education are closely followed by increasing numbers of couples in which the woman’s educational attainment exceeds man’s (the increase of the F-index above 0.5 led to a decrease in H-index below 0) (Hypothesis 2) and this finding also corroborates the results of De Hauw, Grow and Van Bavel’s (2017) and (Erat 2021). Consequences of the gender gap in education differ between women and men with different educational attainment (Hypothesis 2a). In Serbia and Slovenia, highly educated women face a shrinking number of equally educated men, so a high female advantage in tertiary education is associated with lower educational homogamy among highly educated women. Across different contexts (e.g. Serbia and Slovenia), the norms governing marriage markets have proven flexible enough to accommodate the increasing numbers of highly educated women and, as consequence, the num- bers of women marrying down has grown steadily. This is demonstrated by the fact that, on average, up to 40% of highly educated women in Serbia and Slovenia accept to »marry down«, which is in accordance with the trends observed in other countries (Esteve et al. 2016; De Hauw, Grow and Van Bavel 2017; Erat 2021). The higher education also affected the educational assortative mating in the case of less educated women, because it is more difficult for a less educated woman to »marry up« given that she has to »compete« with highly educated women. While highly educated women are indeed ready to »marry down«, moderately edu- cated women are not. Moderately educated women rarely decide to marry a less educated partner: this has happened in only 2.5% of cases in Slovenia and 5% of cases in Serbia. Such results corroborate the theory of preferences where a desirable partner is one with a higher or equal socio-economic status, as »marrying down« is not a viable option for moderately educated women in lower-paying jobs. On the other hand, we have une- ducated men, whose position is the most complicated. For less educated men, the fact that women now often have more advanced degrees than men imply that they may face a shrinking number of potential mates, given 18 62-2_acta49-1.qxd 13.12.2022 12:07 Page 18 that women tend to prefer a partner with at least the same educational attainment. In Serbia and Slovenia, there is a general tendency among uneducated men to partner less often with uneducated women (the share of homogamous marriages declines and the share of hypogamous marriages increases, up to ten times in Serbia). This might be the result of a general decline in the number of uneducated women in the marriage market, but it might also be caused by an increase in men’s preferences for more highly educated women. Comparative analysis between Serbia and Slovenia has not shown any significant changes in the edu- cational assortative mating patterns. Even though both countries were developing within the same economic and political area, and belonged to the same country, certain differences in demographic, eco- nomic and social changes have started forming pretty early on. Upon the breakup of Yugoslavia, their development paths started completely diverging, with Slovenia joining the EU rather quickly, while Serbia has not finished the process of economic and social transition even after two decades. Social differences between Serbia and Slovenia, whose root cause is different historical and cultural influence of Austro-Hungarian and Ottoman Empire, persisted in the times of SFR Yugoslavia, even though the ideas about equality, equi- ty and uniform development were present. Socioeconomic differences became more noticeable upon the breakup of Yugoslavia and after Slovenia experienced faster Second Demographic Transition. Consequently, it was expected to see a great difference in assortative mating patterns in the two countries (Hypothesis 3). However, the comparative analysis has shown that the differences in the magnitudes of those changes are very low. The main cause – higher education – and the corresponding trends are nevertheless the same. It should be noted though, that the start of the changes in educational assortative mating differs in Serbia and in Slovenia, with Serbia lagging behind Slovenia by almost two decades (Hypothesis 3a). 6 Conclusion This paper analyses the changes in educational assortative mating patterns in Slovenia and Serbia. The changes in educational assortative mating patterns have been analysed in the context of the female advan- tages in higher education, which affected the marriage market in those two countries in a very similar manner. The results of our research on the changes in educational assortative mating patterns in Serbia and Slovenia show that young people have adapted to new demographic realities by increasingly forming unions in which women have the educational advantage, leading to substantial declines in historical hypergamic patterns. For demographers, the fact that educational hypergamy is declining has a twofold importance. Firstly, it can be linked with greater gender equality (McDonald 2000; Dorius and Firebaugh 2010; Low 2019), because greater gender equality has in some cases been linked to a recovery in the fertility rate in low fer- tility countries (Esping-Andersen and Billari 2015), and Serbia and Slovenia are indeed both low fertility countries. Secondly, declining traditional patterns of hypergamy suggest an increased prevalence of edu- cational hypogamy (assuming constant prevalence of homogamy), which may affect the survival of marital unions where the wife is more educated than the husband (Theunis et al. 2016; Međumurec and Čipin 2019), especially in environments where educational hypogamy is the second most prevalent type of partnership, as it is the case in Serbia and Slovenia. The observed trends and the extent of the changes in assortative mat- ing patterns, which emerged due to female dominance in the highly educated population, were nearly identical in Serbia and in Slovenia – the only difference was the timeframe when those changes occurred. The gen- eral impression is that these changes occurred with a certain delay in Serbia compared to Slovenia, but there seems to be convergence between the two countries in recent years. ACKNOWLEDGMENT: This paper was written as part of the 2021 Research Program of the University of Belgrade Faculty of Geography, with the support of the Ministry of Education, Science and Technological Development of the Republic of Serbia. 7 References Becker, G. 1973: A theory of marriage: Part I. The Journal of Political Economy 81-4. Blackwell, D. L., Lichter, D. T. 2000: Mate selection among married and cohabiting couples. Journal of Family Issues 21-3. DOI: https://doi.org/10.1177/019251300021003001 Blau, P. M. 1977. A macrosociological theory of social structure. The American Journal of Sociology 83-1. Acta geographica Slovenica, 62-2, 2022 19 62-2_acta49-1.qxd 13.12.2022 12:07 Page 19 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 Blossfeld, H.-P. 2009: Educational assortative marriage in comparative perspective. Annual Review of Sociology 35. DOI: https://doi.org/10.1146/annurev-soc-070308-115913 Bobić, M. V. 2017: From partnership to parenthood. From mutuality to divisions. Limesplus 14-2. Bobić, M., Vukelić, J. 2011: Second demographic transition de-blocked? Sociologija 53-2. DOI: https://doi.org/ 10.2298/SOC1102149B Breznik, D. 1991: The population of Yugoslavia. Beograd. Chiappori, P.-A. 2020: The theory and empirics of the marriage market. Annual Review of Economics 12. DOI: https://doi.org/10.1146/annurev-economics-012320-121610 De Hauw, Y., Grow, A., Van Bavel, J. 2017: The reversed gender gap in education and assortative mating in Europe. European Journal of Population 33. DOI: https://doi.org/10.1007%2Fs10680-016-9407-z Devedžić, M. 2004: Značaj bračnosti za nivo fertiliteta. Demografija 1. Dobrotić, I., Stropnik, N. 2020: Gender equality and parenting-related leaves in 21 former socialist countries. International Journal of Sociology and Social Policy 40-5,6. DOI: https://doi.org/10.1108/ IJSSP-04-2019-0065 Dorius, S. F., Firebaugh, G. 2010: Trend in global gender inequality. Social Forces 88-5. DOI: https://doi.org/ 10.1353/sof.2010.0040 England, P. 2010: The gender revolution: Uneven and stalled. Gender and Society 24-2. DOI: https://doi.org/ 10.1177/0891243210361475 Erat, D. 2021: Educational assortative mating and the decline of hypergamy in 27 European countries: An examination of trends through cohorts. Demographic Research 44-7. DOI: https://doi.org/10.4054/ DemRes.2021.44.7 Esping-Andersen, G., Billari, F. C., 2015: Re-theorizing family demographics. Population and Development Review 41-1. DOI: https://doi.org/10.1111/j.1728-4457.2015.00024.x Esteve, A., García-Román, J., Permanyer, I. 2012: The gender-gap reversal in education and its effect on union formation: The end of hypergamy? Population and Development Review 38-3. DOI: https://doi.org/ 10.1111/j.1728-4457.2012.00515.x Esteve, A., McCaa, R., López, L. A. 2013: The educational homogamy gap between married and cohabit- ing couples in Latin America. Population Research and Policy Review, 32-1. DOI: https://doi.org/10.1007/ s11113-012-9263-4 Esteve, A., Schwartz, C. R., van Bavel, J., Permanyer, I., Klesment, M., García-Román, J. 2016: The end of hypergamy: Global trends and implications. Population and Development Review 42-4. DOI: https://doi.org/ 10.1111/padr.12012 Goldscheider, F., Bernhardt, E., Lappegård, T. 2015: The gender revolution: A framework for understanding changing demographic behavior. Population and Development Review 41-2. DOI: https://doi.org/10.1111/ j.1728-4457.2015.00045.x Graovac Matassi, V., Talan A. 2021: Recent marriage and childbearing trends in Croatia and Slovenia: A comparative review. Acta geographica Slovenica 61-1. DOI: https://doi.org/10.3986/AGS.8596 Grow, A., Van Bavel, J. 2015: Assortative mating and the reversal of gender inequality in education in Europe: An agent-based model. PLoS ONE 10-6. DOI: https://doi.org/10.1371/journal.pone.0127806 Hamplova D. 2008: Educational homogamy among married and unmarried couples in Europe: Does con- text matter? Journal of Family Issues 30-1. DOI: https://doi.org/10.1177/0192513X08324576 Istenič, T., Ograjenšek, I., Sambt, J. 2018: The gender gap in economic dependency over the life cycle: Some theoretical and practical considerations. Economic Research-Ekonomska Istraživanja 31-1. DOI: https://doi.org/10.1080/1331677X.2018.1426479 Josipovič, D. 2003: Geographical factors of fertility. Acta geographica Slovenica 43-1. DOI: https://doi.org/ 10.3986/AGS43104 Josipovič, D. 2016: The post-Yugoslav space on a demographic crossway: 25 years after the collapse of Yugoslavia. Stanovništvo 54-1. DOI: https://doi.org/10.2298/STNV160415006J Josipovič, D. 2018: Mobility and highly educated workers: Brain drain from Slovenia: National and regional aspects of brain circulation. Labour Mobility in the EU. Ljubljana. Jovanović-Gavrilović, B., Radivojević, B. 2017: Education of population for the future and the future of education. Stanovništvo 55-1. DOI: https://doi.org/10.2298/STNV171106006J Kalmijn, M. 1994: Assortative mating by cultural and economic occupational status. American Journal of Sociology 100-2: DOI: https://doi:10.1086/230542 20 62-2_acta49-1.qxd 13.12.2022 12:07 Page 20 KC, S., Barakat, B., Goujon, A., Skirbekk, V., Sanderson, W. C., Lutz, W. 2010: Projection of populations by level of educational attainment, age, and sex for 120 countries for 2005-2050. Demographic Research 22-15. DOI: https://doi:10.4054/DemRes.2010.22.15 Kuhar, M., Reiter, H. 2010: Transformation and demographic change in the ex-Yugoslav countries – mate- rialist, idealist, and institutionalist perspectives on reproductive trends. Annales, Series Historia et Sociologia 20-1. Lerch, M. 2017: Fertility and union formation during crisis and societal consolidation in the Western Balkans. Population Studies 72-2. DOI: https://doi.org/10.1080/00324728.2017.1412492 Lewis, S. K., Oppenheimer, V. K. 2000: Educational assortative mating across marriage markets: Nonhispanic whites in the United States. Demography 37. DOI: https:// 10.2307/2648094 Lichter, D. T., Anderson, R. N., Hayward, M. D. 1995: Marriage markets and marital choice. Journal of Family Issues 16-4. DOI: https://doi.org/10.1177/019251395016004001 Low, C. 2019: A »reproductive capital« model of marriage market matching. Working paper. Philadelphia. Mare, R. D. 1991: Five decades of educational assortative mating. American Sociological Review 56-1. DOI: https://doi.org/10.2307/2095670 Marinković, I. 2018: Demographic analysis of the difference in male and female mortality. Demografija 15-15. DOI: https://doi.org/10.5937/demografja1815001M McDonald, P. 2000: Gender equity in theories of fertility transition. Population and Development Review 26-3. DOI: https://doi.org/10.1111/j.1728-4457.2000.00427.x Međumurec, P., Čipin, I. 2019: Who get divorced more often? Education and marital (in)stability in Croatia. Revija za socijalnu politiku 26-3. DOI: https://doi.org/10.3935/rsp.v26i3.1581 Mirić, N. 2019: Difference in fertility between women with and without tertiary education. Zbornik Matice srpske za drustvene nauke 170. DOI: https://doi.org/10.2298/ZMSDN1970245M Nikitović, V. 2016: Long-term effects of low fertility in the region of former Yugoslavia. Stanovništvo 54- 2. DOI: https://doi.org/10.2298/STNV161115009N Oppenheimer, V. K. 1988: A theory of marriage timing. American Journal of Sociology 94-3. DOI: https://doi.org/10.1086/229030 Oppenheimer, V. K. 1994: Women’s rising employment and the future of the family in industrial societies. Population and Development Review 20-2. DOI: https://doi.org/10.2307/2137521 Penev, G. D. 2010: Recent changes in the population dynamics of Serbia and the Balkan countries. Demografija 7. Penev, G. D., Predojević-Despić, J. R. 2019: The population change in Serbia in the post-Yugoslav period (1991-2017): Significant demographic aspects. Sociološki pregled 53-3. DOI: https://doi.org/10.5937/ socpreg53-21902 Petrović, M. 2011: Changes of marital behaviour and family patterns in post-socialist countries: Delayed, incomplete or specific second demographic transition? Stanovništvo 49-1. DOI: https://doi.org/10.2298/ STNV1101053P Rašević, M., Vasić, P. 2017: Education as a factor of fertility and population policy in Serbia. Annales, Series Historia et Sociologia 27-3. DOI: https://doi.org/10.19233/ASHS.2017.42 Sardon, J. P. 2001: Demographic change in the Balkans since the end of the 1980s. Population 13-2. Schoen, R. 1983: Measuring the tightness of a marriage squeeze. Demography 20-1. DOI: https://doi.org/ 10.2307/2060901 Schwartz, C. R., Han, H. 2014: The reversal of the gender gap in education and trends in marital dissolution. American Sociological Review 79-4. DOI: https://doi.org/10.1177/0003122414539682 Schwartz, C. R., Mare, R. D. 2005: Trends in educational assortative marriage from 1940 to 2003. Demography 42-4. DOI: https://doi.org/10.1353/dem.2005.0036 Sweeney, M. 2002: Two decades of family change: the shifting economic foundations of marriage. American Sociological Review 67-1. DOI: https://doi.org/10.2307/3088937 Šircelj, M. 2007: Fertility and educational attainment in Slovenia. Anthropological Notebooks 13-2. Šobot, A. 2012: Three demographic consequences of gender-specific behavior pattern: The case of Serbia. Stanovništvo 50-2. DOI: https://doi.org/10.2298/STNV1202085S Sobotka, T. 2008: Overview chapter 6: The diverse faces of the second demographic transition in Europe. Demographic Research 19. DOI: https://doi.org/10.4054/DemRes.2008.19.8 Acta geographica Slovenica, 62-2, 2022 21 62-2_acta49-1.qxd 13.12.2022 12:07 Page 21 Damjan Bakić, Vera Gligorijević, The change in educational assortative mating in Serbia and Slovenia, 1970–2020 Theunis, L., Schnor, C., Willaert, D., Van Bavel, J. 2018: His and her education and marital dissolution: Adding a contextual dimension. European Journal of Population 34. DOI: https://doi.org/10.1007/s10680- 017-9448-y Torr, B. M. 2011: The changing relationship between education and marriage in the United States, 1940-2000. Journal of Family History 36-4. DOI: https://doi.org/10.1177/0363199011416760 Van Bavel, J. 2012: The reversal of gender inequality in education, union formation, and fertility in Europe. Vienna Yearbook of Population Research 10. DOI: https://doi.org/10.1553/populationyearbook2012s127 Vincent-Lancrin, S. 2008: The reversal of gender inequalities in higher education: An ongoing trend. Higher Education to 2030, Volume 1, Demography. Paris. DOI: https://doi.org/10.1787/9789264040663-11-en West, C., Zimmerman, D. H. 1987: Doing gender. Gender and Society 1-2. 22 62-2_acta49-1.qxd 13.12.2022 12:07 Page 22