Is Inheritance Fair in South Korea? The Difference in Inheritance Allocation According to Birth Order, Gender, and Caregiving Contribution

Intersectionality: Birth Order, Gender, and Social Norms

In the socio-cultural context of Korea, the phenomenon of birth order and gender affecting inheritance allocation can be explained by the intersection theory emphasized in third-wave feminism. Intersectionality is a theory associated with this wave, and its main argument is that the complex operational mechanism of an oppression mechanism is not revealed when a group is divided solely by gender due to the overlap with or superimposition of other characteristics (e.g., race, hierarchy, or religion) (Cho, Crenshaw, & McCall, 2013). The main focus of this study is patriarchy and inheritance in South Korea. In this case, the main heir is the eldest son among direct family members in the paternal line. Therefore, the power is mostly inherited by one eldest son satisfying the specific condition. It is necessary to divide siblings into the eldest daughter, the eldest son, other sons, and other daughters to understand the oppression mechanism in the inheritance allocation. However, previous studies examining the effect of gender or focusing on the difference between the eldest son and other siblings have overlooked the difference between the eldest daughter and the sons other than the eldest son and that between the eldest daughter and the other daughters (Bae, 2008; Park, Whang, Jung, & Lee, 2014). Therefore, our study emphasizing intersectionality analyzes not only the gender difference between the eldest son and eldest daughter but also the difference in the inheritance allocation within daughters (the eldest daughter vs. the other daughters).

The relationship among inheritance, birth order, and gender vary depending on the socio-cultural context in which social norms work (Szydlik, 2012). In Australia and Sweden, bequests tend to favor over men (Erixson & Ohlsson, 2018; Kelly & Harding, 2006). In Sweden or the United States, a youngest daughter who mainly takes care of the parents can become the main heir (Erixson & Ohlsson, 2018; Light & McGarry, 2004). Unlike in these bilineal and gender-equal countries, however, traditional attitudes still operate in patrilineal familism-dominant Asian countries such as India and China (Jayachandran, 2015; Jiang, Li, & Feldman, 2015; Nauck, 2010). As such, Nauck (2010) reported that the gender of a child has until recently been strongly correlated with inheritance. The patriarchal and inheritance norms favoring only the eldest son even among sons generates a hierarchical gap among the children, and it may lead to a decline in the social hierarchy for younger daughters because of their disadvantages in inheritance (H. Kim, 2008; Lee, 2015). Although it has been studied in India and China (Bhalotra et al., 2018; Deininger, Goyal, & Nagarajan, 2013; Jayachandran, 2015; Jiang et al., 2015), the issues of gender inequalities and gaps in inheritance have hardly been discussed in South Korea, a society built on a similar socio-cultural background.

Strategic Inheritance Motive Theory: Inheritance as Compensation for Caregiving Contributions

Many previous studies have explained the inheritance process from parents to children based on strategic inheritance motive theory (Chang & Luo, 2015; H. Kim, 2008). This theory is also called the exchange inheritance motive theory, and it explains that inheritance allocation is determined by the degree of support provided by children (Caputo, 2005; H. Kim, 2008). Using the Young Women’s Cohort of the National Longitudinal Surveys, Caputo (2005) analyzed 399 young females living in the United States who provided caregiving to their parents in the later part of their lives in terms of whether they received an inheritance. In other words, children try to secure their right to an inheritance by providing care at the end of their parents’ lives. It seems logical and reasonable to suggest that the exchange inheritance motive theory can be applied to South Korea. Since the mid-Joseon Dynasty (the 1600s), the eldest son has received preference has had the inheritance priority (Hwang & Kim, 2014; Kim & Hwang, 2013; Nam et al., 2015). However, the Family Law was revised in 1991 according to the principle of inheritance by equal distribution, and the patriarchal family system was abolished. Moreover, current regulations stipulate that children who have cared for their parents are supposed to inherit but it is unclear whether these children actually received inheritance proportional to the amount of caregiving to the parents. C. Park (2014) analyzed parent –children pairs (596 parents, 1,504 children) using the Korean Labor & Income Panel Study and reported that it was difficult to find evidence that children cared for their parents in order to receive an inheritance from them in the future. Therefore, this study aims to compare the intersectionality between birth order and gender based on the traditional socio-cultural norms of Korea and the contribution as a caregiver before a parent’s death based on the strategic inheritance motive theory, and see which theory works better in reality.

Socio-Demographic Characteristics and Inheritance of Parents and Children

Most studies conducted in South Korea analyzed judicial decisions and requested legal amendments (Lee, 2015; Nam et al., 2015). Not many studies have conducted quantitative statistical analysis by surveying families who have gone through the inheritance process (Hwang & Kim, 2014; Lee, 2015; C. Park, 2014). After conducting in-depth interviews with siblings who experienced parental deaths and inheritance and analyzing their relationships qualitatively based on grounded theory methods, Lee (2015) suggested that quantitative studies should be conducted in the future to identify variables affecting the inheritance process. The quantitative studies carried out in South Korea only measured inheritance intention, inheritance plans, and inheritance perceptions as dependent variables rather than the inheritance behavior itself (H. Kim, 2008; Y. Kim, 2013; C. Park, 2014; Park et al., 2014) but, as far as we know, no study has analyzed inheritance behaviors in Korea.

Prior to discussing the original insights we intend to develop in this study, we wish to review the results, contributions, and suggestions of previous studies for future research, which identified that there was no difference in inheritance intentions, inheritance plans, and inheritance perceptions arising from the characteristics of parents and children. Kim, Hwang, and Kim (2012) phoned 1,000 South Koreans (≥ 50 years old) to understand the public’s perception of the most desirable inheritance procedure. The majority (66%) responded that an inheritance should be evenly distributed to all children, with 15% recommending a relatively even distribution but giving a little more to the eldest son. Other less common answers were returning assets to society (7%), inheritance only to the eldest son (5%), inheritance only to children who looked after parents (5%), and even distribution between sons but omitting daughters (2%). The 2011 National Survey of Living Conditions and Welfare Needs of Older Koreans representing the elderly (≥ 65 years old) in South Korea also showed similar results (Kang, 2012). The majority (67%) responded that it was most desirable to bequeath assets evenly between children, with 21% saying more should be bequeathed to the eldest son, 10% only to the eldest son, and 2% only to sons. From their results, Hwang and Kim (2014) and Kang (2012) concluded that the traditional inheritance method had been dismantled and inheritance had become equal and diverse in South Korea. However, we believe that the results of these studies only reflect opinions about desirable inheritance procedures, not actual inheritance behavior, and that those results may therefore not be reflected in actual inheritance behaviors.

Many previous studies have consistently reported that inheritance decisions are closely related to the socio-demographic characteristics and death-related characteristics of parents (Bae, 2008; Kang, 2012; H. Kim, 2008; Y. Kim, 2013). Age, gender, education, marriage status, and assets are among parents’ socio-demographic characteristics that affect preferences for how bequests are made (Kim & Hwang, 2013; Park et al., 2014). They showed a stronger preference for a son-centered traditional inheritance approach when they were older, had a lower level of education, did not have a spouse, and had fewer bequests (Kim & Hwang, 2013; Y. Kim, 2013; Park et al., 2014). Moreover, the inheritance allocation plan was influenced by a parent’s gender and the number of children (Hwang & Kim, 2014; Kim & Hwang, 2013; Kang, 2012). Some parents were willing to bequeath more assets to a child who had cared for them. Lee (2015) carried out in-depth interviews with children who had inherited after their parents’ deaths and reported that death-related characteristics such as sudden death and the amount of an inheritance also affected decisions regarding the disposition of bequests. Therefore, this study also included these characteristics as control variables in this analysis model.

Previous studies have also revealed that the socio-demographic characteristics of children, as well as those of parents and the death characteristics of parents, were closely related to inheritance, and the effects of the characteristics of parents and children on inheritance varied even within families (Kang, 2012; Park et al., 2014). Lee (2015) reported, using qualitative research methods, that parents made decisions regarding inheritance allocation taking into consideration how well they had educated their young children, whether a child was employed and able to make a living by himself or herself, and whether they themselves had children, in addition to factors such as birth order, gender, and the child’s contribution to parents’ care at the end of their lives. Nam et al. (2015) analyzed the proportion of bequest allocations by birth order, gender, age, sibling composition, and caregiving contribution by analyzing 17 juridical decisions on the legacy of deceased parents. However, it is difficult for studies based on a qualitative analysis of in-depth interview data or an analysis of juridical decisions to guarantee representativeness. It is also hard to find quantitative studies analyzing inheritance behaviors that take the individual characteristics of children into consideration. The study by Erixson and Ohlsson (2018) conducted in Sweden is the only available study whose model analyzed actual inheritance behaviors by including all the characteristics of all the children in a family. This study statistically analyzed the representative administrative data for Sweden, and reported that an equal allocation of bequests occurred more commonly when children were well educated, when someone was the youngest child, and when someone had a grandchild (Erixson & Ohlsson, 2018). However, daughters received more or larger bequests than sons, with a youngest daughter with a child particularly inheriting more assets than other siblings (Erixson & Ohlsson, 2018). By contrast, sons received more and larger bequests than daughters in India and China (Bhalotra et al., 2018; Deininger, Goyal, & Nagarajan, 2013; Jiang et al., 2015). A quantitative study in China also showed that education, age, and marital status were significantly related to the inheritance possibility projection (Jiang et al., 2015). In summary, the results of previous studies conducted in many countries varied according to the socio-cultural context, and no quantitative study on inheritance behavior has been carried out in South Korea. Therefore, we believe that it is necessary to statistically analyze how inheritance behaviors are affected by the characteristics of children in the South Korean socio-cultural context.

Current Study

We try various innovations to take a leap forward from these previous studies. First, most researchers interested in inheritance discussed, argued about, or insisted on the need for an improvement in the law or system (Nam et al., 2015). Moreover, even when they performed a survey, they collected data only on living people (Kim & Hwang, 2013; Park et al., 2014), and little is known about actual inheritance decisions and behaviors after a parent’s death. In order to examine inheritance allocation behavior, we utilized repeated tracking measurements of a representative selection of deceased South Koreans who died in middle or old age, including information regarding them from when they were alive, and information obtained after their death from those close to them , as well as inheritance allocation measurements (Erixson & Ohlsson, 2018; Lee, 2015).

Second, although previous studies conducting surveys on either parents or children or analyzing parent –child pair structure data quantitatively mainly used the absolute inheritance amount as a dependent variable (Erixson & Ohlsson, 2018; Kim et al., 2012; C. Park, 2014), it is not an appropriate variable for identifying relative allocations; inheritance amount and inheritance allocation proportion are indices with different qualitative characteristics. This is because the inheritance amount is proportional to the parents’ absolute asset level, but this asset level is irrelevant when assessing the inheritance allocation proportion. In other words, inheritance allocation proportion is far more appropriate as an indicator for analyzing resource allocation in a family than the inheritance amount. Therefore, this study controlled the amount of inheritance and used inheritance allocation proportion measurements as a dependent variable.

Third, since the proportion of one child’s inheritance affects the values of the other siblings, it is necessary to consider this interconnectivity while collecting information on all the siblings and analyzing it (Erixson & Ohlsson, 2018; Girardin et al., 2018; Jiang et al., 2015; Lee, 2015). The most reasonable statistical method to overcome this problem is a multi-level analysis that consists of parents–all children as a multi-level structure (Chang & Luo, 2015; Erixson & Ohlsson, 2018; Jiang et al., 2015; Lee, 2015). In order to take a multi-level structure that interconnects all siblings into account, a multi-level analysis using the % inheritance for each child and taking the relationship between the children in a family into consideration was conducted.

Lastly, although the inheritance allocation proportion is a value that can only be measured after death, the study should use independent variables and control variables relating to values before death and close to the time of death, because the contribution to caregiving for parents before their death, the individual characteristics of parents and children, and the characteristics associated with a death can affect inheritance allocation (Caputo, 2005; C. Park, 2014). In this respect, inheritance allocation proportions and the characteristics related to parents’ deaths were collected from a survey that interviewed individuals close to the deceased following their death and information regarding other explanatory variables was obtained from the last survey participated in by a parent before death.

Subject and Data

The subjects of this study were 1,587 middle or elderly Koreans (level 2) who had died while at least one of their children was alive, and all their children (6,054 people). This study used KLoSA, which contains inheritance allocation information from after death of South Korean middle or elderly parents and their children. KLoSA provides longitudinal data from the first wave in 2006 to the sixth wave in 2016. The first wave targeted middle-aged and elderly people (≥ 45 years old) representative of the South Korean population and living in local communities as of 2006. The KLoSA data were collected by professionally trained investigators who visited each subject at home. They gathered each subject’s personal information as well as information on the gender, age, education, marriage, and employment status of all their children. If a subject had passed away since the previous survey, they also investigated the end of the subject’s life, the process leading up to their death, and post-death processes (e.g., funeral and inheritance allocation) by questioning those close to them, such as a spouse, child, relative, or neighbor. As a consequence, the KLoSA data are representative and highly reliable. It is also the only available data in South Korea that contain the necessary details (the death, inheritance, and socio-demographic characteristics of deceased parents and their children) to answer our research question.

One thousand, six hundred fifty-eight subjects who participated in the first wave of KLoSA died within 10 years of it, meaning that someone close to them answered in one of the subsequent surveys on behalf of the deceased person: second wave=187 subjects, third wave=309 subjects, fourth wave=327 subjects, fifth wave=438 subjects, and sixth wave=397 subjects. To maximize the sample size, this study combined all the available data and created a single dataset. This study excluded 60 subjects who did not have a living child at the time of their death because the objective of this study was to identify the inheritance allocation proportions of the children. When the deceased had a child, the parents became level 2 and the children became level 1, which is a multi-level structure. Inheritance allocation proportion, a dependent variable, is closely connected to the characteristics of level 1, so that it was transformed into a multi-level-structured dataset (1,598 parents–6,082 children), which treats each child as an observation case. However, 11 parents–28 children had missing values. Since this represented less than 1% (parents=0.7% and children=0.5%) and the effects on the statistical results were negligible, they were excluded from the analysis (Listwise deletion). The final target subjects of this study were 1,587 deceased parents (level 2) and their 6,054 children (level 1).

Measurement

Independent Variables: Birth Order, Gender, and Caregiver before Death

This study combined birth order and gender and used it as an independent variable in order to examine the effects of the traditional patriarchal system and gender norms on inheritance allocation proportion. The birth order and gender of a child is a categorical variable that has four classes (eldest daughter, eldest son, other daughters, and other sons). It was converted as a dummy variable, and the eldest daughter was used as a reference group. In order to estimate the effect of pre-death caregiving contribution on inheritance allocation, this study divided caregivers into primary caregivers and secondary caregivers and used them as independent variables. The primary caregivers were defined as children who were primary caregivers at one month before the parent’s death and they were coded as 1 (=yes), while non-primary caregivers were coded as 0 (=no). Similarly, secondary caregivers were coded as 1 or 0 depending on whether they were secondary caregivers at one month before the parent’s death.

Characteristics of Study Subjects

Table 1 shows the descriptive statistics of the parent–child characteristics. The characteristics of the parent (G1) are the same as the number of death cases since they are level two in the multi-level model. Moreover, all deceased parents are included in the descriptive statistics of Table 1 regardless of the presence of inheritance. Deceased parents were almost equally divided by gender (male 50.5%, female 49.5%). Their most common education level was elementary school graduation (73.9%), followed by middle school (11.1%), high school graduation (10.4%), and junior college and over (5.2%). The mean age of deceased parents was approximately 80 and 61% were married at the time of death. The number of G2 children averaged 4.6. A somewhat expected death was 40.1%, followed by expected (34.8%) and sudden death (25.1%). The mean inheritance value converted to a natural logarithm was 0.6 (SD = 2.3). The mean amount bequeathed by all deceased parents was 12.7 million KRW, and the median amount for those who bequeathed more than 10,000 KRW was over 140 million KRW.

Table 1

Descriptive Statistics of the Deceased Parents and Their Children’s Characteristics

Unlike parental characteristics, the children’s characteristics are based on all instances since they are nested at level 1 in the multi-level model. The results showed that the mean inheritance allocation was 3% including all children (N = 6,054) whose inheritance allocation was 0%, but the average was 36.7% for children (n = 488) whose inheritance allocation was 1% or more. About 14.7% of children were the eldest son, 11.5% the eldest daughter, 37.5% the other sons, and 36.3% the other daughters. The primary caregiver children were 4.6% of all children, while the secondary caregiver children were 3.0% of all children. The proportion of primary and secondary caregivers was low because nurses and professional caregivers were primary caregivers when parents were hospitalized and spouses were commonly caregivers when they lived at home. The mean age of children was 50.8 years old (range: 6–85). In terms of children’s education, 42.2% of them were high school graduates or lower, 26.7% were middle school graduates or lower, 24.8% were 4-year college graduates, 4.6% were junior college graduates, and 1.7% graduate school or higher. There were more employed (62.7%) than unemployed (36.3%) children and most were married (82.5%), followed by single (11.7%), and only 5.8% divorced, bereaved, or separated. Only 1.8% of children did not have a sibling. The mean number for G3 was 1.8 people. For reference, the descriptive statistics of those with siblings are presented on the right side of Table 1. Since only 1.8% of subjects did not have a sibling, the results of descriptive statistics were almost identical whether or not subjects without siblings were included.

Inheritance Allocation Proportion According to Birth Order, Gender, and Caregiving before the Death

Table 2 shows the results of the multi-level tobit model analysis. This study divided the model into Model 1 and Model 2 and analyzed them separately in order to examine whether the relationship between independent variables and inheritance allocation proportion varied by the presence of inheritance. Model 1 includes all deceased parent–children regardless of the presence of inheritance. On the other hand, Model 2 excludes all cases where there was no inheritance at all, showing cases where parents did make bequests but a child did not receive anything.

Table 2

The Results of the Multi-Level Tobit Model on Children’s Inheritance Allocation Proportion According to Birth Order, Gender, and Caregiving Before the Death of Parents

In Model 1, the number of deceased parent cases (level 2) was 1,587, and that of their children (level 1) was 6,054. Among the 6,054 observed cases, left-censored observations (0% inheritance allocation) amounted to 5,566 people, and uncensored observations with 1%–100% inheritance allocation were 488 people. The Wald chi-square value was 363.1 (DF = 25) and significant at a 99.9% level. Moreover, the likelihood-ratio test value was 783.8 (DF = 1) and significant at a 99.9% level, implying that it was a statistically appropriate model. In Model 1, the eldest son had a 26.9 percentage points higher inheritance allocation proportion than the eldest daughter (p < .001), while the other daughters had 18.9 percentage points lower inheritance allocation proportion than the eldest daughter (p < .01). However, the inheritance allocation proportion of the other sons as compared to the eldest daughter was not significantly different. The inheritance allocation proportion of primary caregivers as compared to secondary caregivers was also not significantly different.

In Model 2, the number of cases of deceased parent with inheritances (level 2) was 457 people, and the number of cases of associated children (level 1) was 1,692 people. The Wald chi-square value of Model 2 was 274.4 (DF=25), significant at a 99.9% level. Moreover, the likelihood-ratio test value was 182.5 (DF=1), also significant at a 99.9% level, indicating that it is a statistically appropriate model. Among independent variables, the eldest son had a 31.5 percentage points higher inheritance allocation than the eldest daughter (p < .001), which was approximately 5 percentage points higher than that of Model 1. Second, the other daughters had a 17.4 percentage points lower inheritance allocation than the eldest daughter (p < .01). As was the case with Model 1, in Model 2 the inheritance allocation proportion of the other sons as compared to the eldest daughter, and the inheritance allocation proportion for primary caregivers as compared to secondary caregivers were not significantly different.

Limitations and Future Directions

We would like to suggest directions for future studies based on the limitations of this study so this research can be advanced. First, this study could only use the data from five waves (KLoSA’s death investigation 2nd–5th waves) and could not analyze temporal variation because the number of dead for each wave was small. However, the relationships between inheritance and traditional patriarchy, gender norms, and contributions as primary caregiver at the end of life could vary depending on the period. Therefore, future studies need to evaluate the changes in trends from the past to the present and project into the future by conducting a panel model analysis that can estimate the temporal effects accurately using more waves and data from target subjects.

Second, this study measured the contribution only by evaluating whether the subject was a primary or secondary caregiver one month prior to death and reported that this does not significantly affect inheritance allocation proportion. However, the level of contribution can be measured in various ways and over a longer term. In this context, future studies should subdivide and examine the effects of various factors such as resource exchange between parents and children, emotional exchange, support, and cohabitation. Although the coefficient of the primary caregiver was large (Table 2), it was not significantly different. This could be because of a large standard error, induced by the small proportion of cases (5%) where children acted as the parents’ primary caregiver before their death. The statistical analysis would be much more reliable if the number of KLoSA waves were increased and more cases of children acting as primary caregivers are added. Moreover, it is also possible that temporal variation may exist in this effect and the effect may have increased in recent years. Therefore, the effects of caregivers on the inheritance distribution will need to be analyzed in more detail in future studies after more KLoSA data have been accumulated.

The results of this study show that traditional patriarchal attitudes toward birth order and gender norms still play a major role in actual inheritance allocation behaviors despite the current law stipulating that this allocation should be based on the caregiving contribution at the end of life and regardless of birth order and gender. This study is meaningful in that it has raised the issues of eliminating the gender inequality that descends to the next generations at the end of life and the need to respect the value of caregiving.