This study is conducted with quantitative approach through regression analysis, specifically instrumental variable regression analysis The research method in this study is arranged in two stages. First, the analysis begins with the selection of variables and the formation of a general model of the influence of maternal social capital on the health status of children. Second, analysis the role of maternal social capital on the physical health of children with the instrumental variable (IV) regression analysis employs 2 Stage Least Square (2SLS). This analysis is used to overcome the issue of endogeneity in maternal social capital or our variable of interest.

This study uses IFLS data (Indonesia Family Life Survey) in 1997 and 2014. The use of IFLS wave-2 data in 1997 was chosen because data on maternal characteristics, especially maternal social capital data were available starting in 1997 which provides child health data when the child is an adult and IFLS wave-5 is the last IFLS data.

The general model that will be used in this study refers to the research of (Sujarwoto & Tampubolon, 2013), which is as follows:

(1)

By using IV, equation (1) can be modified to be our standard model:

(2)

Where is children’s physical health as dependent variables. represents maternal social capital define by days of participating in community activities, represents mother’s characteristics consist of mother’s age, mother’s education level. represents household characteristics such as per capita expenditure in household level where mother and children live, number of household member, total asset in the household. defines vector of community variables.

This study uses data on maternal social capital which is thought to contain endogeneity. This is due to the simultaneity problem between maternal social capital and other mother characteristic variables, such as mother’s education, per capita expenditure, and number of household members. This is supported the theory of health production functions by (Grossman, 1972), where human capital (education) and social capital are determinants of health.

Mother’s participation in activities in her community is influenced by the level of per capita expenditure in her household. (Nobles & Frankenberg, 2007) explain that the effect of mother’s participation on child health will be greater when the mother has a low education and per capita expenditure. Mothers who have low socio-economic conditions will take advantage of the value they get from the community compared to mothers who have better socio-economic conditions.

Simultaneous problems can occur in maternal social capital. Maternal social capital can be influenced by other factors not considered in the model. Other factors such as the mother’s obligation to participate in the village cooperatives in her village or the mother’s employment status can affect the maternal social capital (Nobles & Frankenberg, 2007).

The correlation between mother’s social capital and child’s health is simultaneous. Increasing maternal social capital can improve child health. On the other hand, health issue in children will limit mothers from participating in various community activities. Moreover, the healthy condition of the child can support the mother’s participation in various activities.

The descriptive statistics of this study are described in Table 2 observational data used to explain that the average child as an adult has good health. The physical health index has an average value of 5.16 percent, which means that the average child had a healthy physique as an adult in 2014. This index explains children’s physical disorder context; the smaller the index, the better children’s physical health.

Observational data used in this study showed that the average age of the mothers included in the sample was 35 years. The social capital variable (frequency of mothers participating in activities) shows that, on average, mothers participate in activities in their community 12 hours a year. Data on mothers’ education shows that two percent of mothers are primary school graduates. 1.5 percent are high school graduates, and 0.18 percent of mothers are college graduates, while 96.5 percent of mothers are not in school or have not graduated from elementary school in the year 1997.

Because other factors also influence maternal social capital, the social capital variable is a non-independent variable, either due to the issue of the committed variable or the problem of simultaneity. The mother’s education will influence maternal social capital. Mothers who have an education will be more aware of interacting with the surrounding community (Nobles & Frankenberg, 2007). On the other hand, according to (Sujarwoto & Tampubolon, 2013), the maternal social capital will also be influenced by the mother’s place of residence.

Maternal social capital can also experience simultaneity problems. Maternal social capital can be influenced by other variables that are not included in the model, such as the mother’s obligation to participate in one of the activities and the type of mother’s work (Nobles & Frankenberg, 2007). The problem of simultaneity and committed variables is supported by the results of the Wu-Hausman endogeneity test with a value of 3.979 (p-value = 0.046) and a Durbin chi2 value of 4.002 (p-value = 0.045). With both significant values, it proves that the maternal social capital variable is an endogenous variable.

To overcome the problem of endogeneity on maternal social capital used instrument variables. One variable that was supposed to be used as an instrument variable was selected in this study, namely ethnic similarity. Validity testing was carried out to determine the instrument variable to be used.

In Indonesia, the same culture and language have an essential role in shaping community or social group activities. Individuals with the same ethnicity will interact more because of the same culture and language. Intensive interaction will make gatherings or activities easier. Furthermore, the interaction is motivated by a sense of community-owned by the community and the same goals among the people (Bowen, 1986) (Geertz, 1962) (Koentjaraningrat, 1961). The percentage level of ethnic/ethnic similarity in a community will not affect children’s health in the future. So that in this study, ethnic similarity instrument variables will be used in physical health models. Table 3. show that ethnic/ ethnic similarity is correlated with mother’s social capital. This uncorrelation explains that mothers participate in activities partially not based on the same ethnicity.

Maternal social capital uses a proxy for the frequency or intensity of mothers participating in activities in their community in one year. Maternal social capital has a marginal negative effect on children’s physical health disorders as adults. The mother’s presence in her association will reduce the chance of physical disorders by 0.10 percent of physical disorders in children as adults. Maternal social capital has been shown to affect children’s physical disorders as adults, seen from a significant value. These results are consistent with previous studies ((Silva & Harpham, 2007); (Nobles & Frankenberg, 2007);(Sujarwoto & Tampubolon, 2013)) which explained that maternal social capital will improve the physical health status of children.

Maternal social capital is an investment for everyone (Lin, 2001). In this context, social capital is an investment for mothers in developing themselves to increase their ability to care for and raise children. Mothers have a role in providing nutrition interventions for children, especially the education of the mother and the economic condition of the family (Abuya et al., 2012). According to (Pangaribowo et al., 2019) women who participate in their communities will have a better bargaining position so that in this context, the position of mothers in the family is increasingly important, and the family increasingly needs their role. (Firmansyah & Sihaloho, 2021) find that large investments in women’s education and health are instead intended to lower the birth rate and improve the well-being of the children.

The results show that mothers with elementary education had a significant influence on physical health. This result means that mothers who have elementary education will reduce the chance of physical health problems by 1.51 percent compared to mothers who do not have elementary education. The per capita expenditure variable is divided into four categories. The 25 percent of households with the lowest per capita expenditure were used as a reference in the study. This study found that the 25 percent of households with the highest per capita expenditures will reduce the chance of children’s physical disorders as adults by 84 percent compared to the 25 percent of households with the lowest per capita expenditures.

Another characteristic that can be seen is the mother’s age, which significantly affects the probability of a child’s physical disorder as an adult—one year younger, the greater the 43.3 percent chance of having children’s physical disorders as adults. Household size has a positive effect on increasing children’s physical health disorders as an adult. It implies that the more household members in a house, the more children have physical health disorders.

This study also regards the community- level characteristics. The presence of Puskesmas and Posyandu or health facilities near children’s houses will reduce children’s physical health disorders in 2014. These health facilities will help the mother easily access health services for her children’s and household members’ needs.

In addition to looking at maternal social capital as a whole, this study will also partially examine maternal social capital by dividing the physical health function model into six parts based on the type of association where mothers participate. It is helpful to see the partial impact of the goals of each association that the mother participates in on the physical health of children as adults. The results show that Integrated Health Service Post and Rotating saving and credit association have significant effects on decreasing children’s physical health disorders in the future.

Table 5 describes the impact of the number of mothers’ days participating in activities in their community on their children’s physical health disorder. The more active mothers participating in the Integrated Health Service Post and Rotating saving and credit association will reduce the probability of children’s physical health disorders as adults. Integrated Health Service Post can provide for the needs of mothers and children, especially in health information and nutritional program. Rotating saving and credit association is the association that has higher participation of mothers. This association could provide community information and provide saving management among participants. Rotating saving and credit association is a social association that has the function of a rotational savings scheme, but the specific purpose of the Rotating saving and credit association is to increase social and business networks among members and increase information networks among its members (McCulloch & Grover, 2010).

This study aims to examine the effect of maternal social capital on children’s physical health in the future. It fills the gap that physical health status is not only affected by short-term factors but also affected by long-term factors. This study also improves the previous findings of a mother’s social capital effect on children’s health status.

The main result shows that maternal social capital reduces children’s physical health disorders. This finding is similar to the study result of (Sujarwoto & Tampubolon, 2013) that in the short-term, mother’s social capital has association with children health outcomes. Maternal social capital is one of the factors in determining healthy behavior and preventing health risks that can involve children’s health status and their quality of life as adults. Through social capital, mothers will have more information, since they get knowledge transfer, social support, social leverage, and informal social controls, that will play important role in providing the best nutrition for their children.

This finding relates to prior research which mentioned that social capital is vital for women, especially mothers with children (Caughy et al., 2003). The spillover of maternal social capital to children’s health is using information diffusion that will affect the healthy behavior of individuals. Mother will implement her information from community activities to household member health behavior, for instance, sanitation behavior, keeping the house clean, providing healthy foods, and sending their children to health facilities ((Silva & Harpham, 2007); (Mengesha et al., 2021)) Social capital has been shown to be positively associated with a range of health outcomes, yet no studies have explored the association between maternal social capital and child nutritional status. Using data from the Young Lives study comprising 7242 1-year-old children from Peru, Ethiopia, Vietnam and the state of Andhra Pradesh in India, we find significant differences in the levels of, in particular, structural social capital (group membership and citizenship. In the Indonesian context, information diffusion and maternal social capital signal are vital because access to information in 1997 is very limited due to geographic remoteness.

The finding also shows that Integrated Health Service Post and rotating saving and credit association (locally known as arisan) have significant effects of decreasing the probability of children’s physical health disorders. Integrated Health Service Post can provide for the needs of mothers and children, especially in health information and nutritional program. At the same time, rotating saving and credit association can provide social and business networks among members, even though it indirectly increases household’s food and non-food expenditures (Astuti, R. D., & Hardi, A. S. (2022).

The implication of this study is the importance of women’s participation, especially mothers, in activities and programs held in the village because the more active the mother’s participation, the more children will have a good health status. The government needs to consider increasing the number of programs and activities in the village as a channel to increase information and access to maternal health knowledge, especially in remote areas.

However, this study also has some limitations, such as this study cannot estimate the amount of information provided by the local community to mothers’ participation. Moreover, it does not consider the biological factors, including the initial health of children and other external factors that could influence children’s health.