Are there heterogeneous impacts of social support on subjective well-being?

1.   Introduction

Subjective well-being is a global health issue exacerbated by the COVID-19 pandemic. Social support has a positive effect on subjective well-being. Studies concerning the correlation between social support and subjective well-being have been conducted from different perspectives. Firstly, social support and active social relations have positive influence on human mental and physical health. Literature suggests that individuals could achieve optimal psychological well-being when they have strong and forceful social support (Mellor et al., 2018; Moeini et al., 2018). Social support can reduce stress or negative influence resulted from negative events or difficulties by acting as a hindrance factors (Cassel, 1976; Cobb, 1976; Liu et al., 2014). With more types of social ties, people might live longer, have less cognitive decline with age, more resistant to infectious diseases, and a better prognosis in life-threatening chronic diseases (Cohen and Janicki-Deverts, 2009; Itzick et al., 2018). Secondly, different types and sources of social support also have positive impacts on happiness for different people. The supports from colleagues and supervisors are significantly positively correlated with the well-being of employees (Merida-Lopez et al., 2019). Support from closest family members, mentors and best friends can improve the subjective well-being of vulnerable groups (Mendonca and Simoes, 2019). Emotional support from children was more important than social support from outside for the older immigrants (Peng et al., 2015). Thirdly, researchers identified other personal and social attributes that also affect the degree of social support to subjective well-being, such as gender (Hori and Kamo, 2018), age (Ju et al., 2018), loneliness (Tian, 2016), emotional intelligence (Gallagher and Vella-Brodrick, 2008), social status (Yoon et al., 2012), social trust (Helliwell and Huang, 2011). For example, Kong and You (2013) examined that overall self-esteem had a positive regulating effect on the relationship between social support and life satisfaction, but had no effect on the negative effect between them by the investigation of Chinese college students.

However, subjective well-being can vary even receiving the same social support. For example, in highly developed countries, people may have strong subjective well-being when they get social support because rich materials and complete social security can satisfy their spiritual demands. Conversely, in less-developed countries where there is a lack of survival necessities, poor medical conditions, weak social security system, people could not achieve subjective well-being because social support cannot meet the requirement of their material need (Ko and Kuo, 2009). It can be further argued that differences in thinking modes, language, religious belief and personal values, could lead to different subjective well-being. For example, a culture of collectivism may associate with a higher degree of subjective well-being; whereas in countries with individualism culture, people's subjective well-being may not reach their expectations with different types of social support. Generally, the economy offers the most important material basis, while culture makes values more diversified, which have significant impact on subjective well-being. Based on the above, it is possible that different levels of economic development and pluralistic culture can lead to different degrees of subjective well-being even with the same social support.

Economy is the material basis of subjective well-being. Income as the most important key factor, determines the level of people's happiness to a certain extent. According to Maslow's Hierarchy of Needs, human happiness comes from satisfaction of needs from lower levels to higher levels (Benson and Dundis, 2003). During this process, income level plays a profound role among other factors. In highly developed countries, high income provides a solid material basis for people's life (Huang et al., 2020). Social support along with high income provides individuals energy and power for spiritual pursuit, ambitions realization, and ultimately enables improvement of subjective well-being. Thus, people's subjective well-being may be stronger when social support reaches or exceeds their own expectation. In low income countries, people are likely to gain lower subjective well-being because of limited material support, such as insufficient survival necessities, poor medical care, scarce educational resources and weak social security system. Under this circumstance, their expectations of personal self-esteem, status and self-realization of spiritual needs are likely to be low. Research evidence supports that working long hours often leads to fatigue and stress (Valente and Berry, 2016). Based on that, people have relatively low subjective well-being when there is a large gap between the level of social support and people's expectations.

Pluralistic cultural background is also considered a factor influencing subjective well-being and social support. Cultural dimensions, such as: power distance, uncertainty avoidance, masculinity and femininity, individualism and collectivism, influence human behaviors of different countries (Hofstede, 1980). People with individualism culture tend to emphasize goals and interests' pursuit that requires less social interaction or cooperative support, thus, their subjective well-being with emotional social support may be low. Anderson et al. (2012) assert that people with high social status gain more respect and appreciation from friends, which suggests people with high social status have more subjective or emotional support by means of self-satisfaction of being respected. With collectivism culture, people pay more attention to maintaining harmonious social relationships, emphasizing cooperation over competition, and developing unity of groups' interests. Thus, their subjective well-being is largely determined by social contacts or relations with others. For example, marital status is one of these relations. It reported that in Japan, full-time work is negatively correlated with women's happiness (Hori and Kamo, 2018). Wang et al. (2020) also assessed that the dual pressures of work and family make female doctors less happy and more likely to experience burnout.

Although the above literatures suggest an impact of relationship between social support and subjective well-being, there is apparently a gap on the potential heterogeneous effect due to economic and cultural differences. This paper aims to make a tentative study with the following specific contributions:

Firstly, based on the Gallup survey data, this paper constructs a panel fixed effect model to further verify the differences of the influence of social support on subjective well-being under different conditions. On the one hand, under the background of economic differences, high-income countries have positive impact of social support on subjective well-being; whereas the lower ones have no significant impact. On the other hand, from the perspective of cultural differences, social support in countries with high and low generosity of cultural characteristics has a significant positive impact on subjective well-being; however, the higher ones have greater impact than the lower ones.

Secondly, the influencing mechanism of social support on subjective well-being has heterogeneity underpinned by different economic and cultural levels. The research shows that the difference of economic and cultural condition had a consistent effect on the influencing mechanism by the regulating variables of democracy and freedom. However, these regulating variables have different effects when their economic and cultural backgrounds are different. Specifically, under economic differences, democracy and freedom both have a significant positive regulating effect in low-income countries; However, under different culture background, democracy and freedom both have a significant positive adjustment effect in both high and low generosity countries, but the moderating effect of the higher ones is greater than the lower ones.

The remainder of this paper is arranged as follows: Section 2 introduces model, variables and data, which including variables description, data description and grouping, as well as stationary test. Section 3 reports test results, which including heterogeneous impacts of social support on subjective well-being by economic and cultural differences, and their influencing mechanisms as well. This is followed by discussion in Section 4. Lastly, section 5 concludes our paper with a note on limitation and implications.

2.   Materials and methods

2.1.   Model setting

This paper aims to explore the differences of the influence of social support on subjective well-being under different conditions. Social support and subjective well- being are the key variables in this research. The following benchmark panel regression model is constructed based on the panel data of this paper.

where SWB_it stands for subjective well-being for the t year of individual i; Control_it denotes for a series of control variables for the t year of individual i, which includes health life expectancy, generosity, corruption perception, freedom, democratic quality, positive affect and negative affect. represents random error terms; represents the constant term, and are the regression coefficients, are individual effects, are time effects.

2.2.   Variable and data

2.2.1.   Variables description

In this model, subjective well-being is the explained variable and social support is the explanatory variable. Some control variables and moderating variables are: health life expectancy, generosity, corruption perception, freedom, democratic quality, positive affect and negative affect. Data for all the variables are obtained from World Happiness Report 2019 (Helliwell et al., 2021). The data sources and description of the variables are shown in Table 1 as below.

The sample data is drawn from 53 selected countries that cover the period from 2007 to 2018. The criteria for the sample selection is considering the continuity of indicators in the Global Happiness Index series reports, and data integrity, regional representation and parameter estimation of sample size. In terms of time span, the data is available from 2006 to 2019 (https://worldhappiness.report/ed/2019/) with useful continuity indicators available from 2007. From the sample countries, the survey scope has expanded from 27 in 2005 to 102 in 2007, then to 146 in 2011, and finally stabilized at around 140 at present. Due to missing values for data continuity in some countries, the sample frame of 53 countries covering a period from 2007 to 2018 has been finally confirmed.

2.2.2.   Data description and grouping

Firstly, the descriptive statistics of all variables data of 53 countries from 2007 to 2018 involved in this paper under complete sample are shown in Table 2. The mean, standard deviation, minimum and maximum of each variable are given as below:

Secondly, diversified backgrounds could make this impact of social support on subjective well-being heterogeneous. As aforementioned, economic and cultural differences are the key factors likely to cause heterogeneity, thus, the samples are grouped according to economic and cultural variables to explore heterogeneity. Generally, national income is more directly related to subjective well-being among the indicators in economic differences grouping. According to the World Bank, the distribution of grouping samples presents a serious asymmetry. By analyzing the data distribution characteristics, the samples are divided into high-income and low-income countries based on national income grouping. The high-income countries include upper or middle-income ones and low-income countries include low-income and lower-middle income ones. There are totally 30 high-income countries and 23 low-income countries after grouping. These 38 high-income countries are Argentina, Azerbaijan, Belarus, Brazil, Canada, Chile, Colombia, Denmark, Dominican Republic, Ecuador, Georgia, Germany, Guatemala, Indonesia, Italy, Japan, Kyrgyzstan, Lithuania, Mexico, Panama, Peru, South Africa, South Korea, Spain, Sweden, Thailand, Turkey, United Kingdom, United States and Uruguay. The rest 23 low-income countries are Bangladesh, Bolivia, Cambodia, Cameroon, Chad, El Salvador, Ghana, Honduras, India, Kenya, Kazakhstan, Mauritania, Moldova, Nepal, Nicaragua, Niger, Philippines, Senegal, Tajikistan, Tanzania, Uganda, Ukraine and Zimbabwe.

Cultural differences are with many indicators. According to the hypothesis of rational man, generosity degree is closely related to human subjective well-being, which reflects the differences on cultural values. Under this premise, the indicator of "Generosity" from World Happiness Report is selected as the proxy variable of cultural difference, and then group samples with data sources. Specifically, all sample data of generosity value is ranked in the order from small to large, and divide them into two groups by 50% quantile: one is the upper quantile which represents the group with higher degree of generosity, and the other is lower quantile which means the lower degree generosity group.

The following descriptive statistical indicators of key variables after grouping are shown in Table 3. It shows that the average level of subjective well-being and social support are different from different groups of income and generosity.

2.2.3.   Stationary test of data

This study is going to take a further step to test the different impact by the panel model. Before testing, in order to avoid the occurrence of false regression, two methods of Levin-Lin-Chu unit-root test and Fisher ADF unit-root test are selected respectively to conduct stationarity tests on all index data. The test results are shown in Table 4.

Table 4 presents the results of unit root test for subjective well-being, social support, and the control variables. It is worth to report in particular, two test results for the null hypothesis that each sequence contains unit roots. The first one is the Levin-Lin-Chu unit root test, and the second is the Fisher augmented Dickey Fuller t test. The criterion is: if the unit root test confirms that it is 0 at 5% significance level, the variable is stationary. In both cases, the null hypothesis is rejected at 1% significance level. Therefore, all variables are stationary for panel model.

3.   Results

3.1.   Heterogeneity by economic differences

Based on the above, this part is to take high and low level income as indicators to verify the heterogeneity of social support on subjective well-being by economic differences. Based on the samples from high and low income countries, the baseline panel model (1) with the time and individual fixed effects is estimated by selecting the hausman test. At the same time, the GMM model is used to test its robustness. Table 5 shows the estimated results of the benchmark regression model with different income levels.

It is confirmed that income differences cause heterogeneous influence of social support on subjective well-being. Based on the estimation of baseline regression model with different income samples, Table 5 shows that the regression estimation coefficients of social support in the panel model for high-income is 1.491, which is significant under the significance of 1%. However, the impact for low income is not significant. The empirical results indicate that high-income countries have positive impact of social support on subjective well-being; whereas the lower ones have no significant impact. In addition, panel GMM regression model estimation results in column (3) and (4) also show the same result. Hence, the estimation results of base panel regression model are stable.

3.2.   Heterogeneity by culture differences

Then, cultural characteristics of generosity levels are taken as an indicator to verify the heterogeneity of social support on subjective well-being by cultural differences. Based on sample data from high-generosity and low-generosity countries, we estimate the baseline panel model (1) with time and individual fixed effects, which is selected by hausman test. At the same time, the panel GMM model is also used to test its robustness. The estimated results of the benchmark regression model with different generosity levels are shown in Table 6.

The results reveal that generosity levels cause heterogeneous influence of social support on subjective well-being. In Table 6, columns (1) and (2) show regression coefficients under high and low generosity sample are 1.053 and 0.885 respectively, both are significant respectively under 10%. The empirical results show that social support in countries with high and low generosity both has a significant positive impact on subjective well-being. In addition, the countries with high generosity have greater impact than in countries with low levels. Besides, the estimation results of column (3) and (4) in panel GMM regression model also show the same results. It indicates that the panel fixed effect model estimation results are stable.

3.3.   Heterogeneity of moderating effects mechanism

3.3.1.   Model

The study continues to investigate the moderating effect of regulatory variables. In this part, only the democracy and freedom variables are selected. Based on the benchmark regression model, we construct the following regulatory effect model:

where SWB means subjective well-being and the interactive item Social support*Mediate means the regulating effect of regulating variable. Mediate is a regulating variable, such as democracy and freedom. The meanings of other variables are consistent with the baseline regression model. If β₃ significant, it indicates that democracy or freedom has a significant regulating effect. And if β₃ > 0, it shows that democracy or freedom has a positive regulating effect. In other words, the higher the level of democracy or freedom, the greater the positive influence of social support on subjective well-being.

3.3.2.   Heterogeneity by economic differences

Taking a similar approach as before, regulatory effect model further test on the moderating effect of democracy and freedom is conducted to examine the heterogeneity of regulatory mechanisms by economic differences. The panel fixed effect model is adopted using samples of different income levels. The results are shown in Table 7.

The moderating effects of democracy and freedom on the relationship between social support and subjective well-being in countries with different income levels are heterogeneous. According to column (1) and (2) in Table 7, the estimated regression coefficients of Social support*Democracy in high and low income sample are 0.623 (not significant under the level of 10%) and 3.333 (significant under the level of 1%) respectively. It shows that democracy has no significant regulating effect in high-income countries, but has a significant positive regulating effect in low-income countries. At the same time, according to columns (3) and (4), the estimated regression coefficients of Social support*freedom in high and low income sample are 0.179 (not significant under the level of 10%) and 6.104 (significant under the level of 5%) respectively. The results show that freedom has no significant regulatory effect in high-income countries, but has a significant positive moderating effect in low-income countries.

Using the same approach, the heterogeneity of moderating mechanisms by cultural differences is further examined. The results are shown in Table 8:

The moderating effects of democracy and freedom on the relationship in countries with different generosity degrees of cultural characteristics are heterogeneous. From columns (1) and (2) in Table 8, democracy has a significant positive adjustment effects both in the samples of high and low generosity countries, which are 3.776 and 1.387, and the moderating effect of the higher ones are greater than the lowers. At the same time, based on column (3) and (4), it told us that freedom also has a significant positive moderating effect both in the samples of high and low generosity countries, which are 6.374 and 4.774, but the moderating effect of the higher ones are greater than the lowers.

4.   Discussion

4.1.   Heterogeneity analysis

The results in section 3 reveal that the differences of economic levels cause heterogeneous impact of social support on subjective well-being. Social support in high income countries has a significant positive impact on subjective well-being; however, it has no significant impact in low-income countries. This may be explained that people in high-income countries generally have a good material base and living conditions, the spiritual level of self satisfaction will be greatly improved with the increase of the channel and strength of diversified social support (tool and information support). This is similar to existing studies. For example, people with higher social capital and social relationships may have more chances to maintain better overall health or better be able to overcome stress and then improve their subjective well-being (Liang et al., 2011; Kumar et al., 2012). Conversely, in low-income countries, people may tend to obtain a large amount of social capital or material resources as their first task because of insufficient survival necessities, poor medical conditions and weak social security system and so on. Some kinds of emotional support, social relations as well as instrumental support cannot meet the requirement of their material needs. For instance, through a five-year longitudinal study, Moskowitz et al. (2013) confirm that social support does not uniformly mitigate the effects of stressors on health for individuals living in urban poverty. Because social networks of the poor tend to include others who are resource deprived, there may be limits to the extent that they buffer individuals from the harmful effects of stress on health. From the above, this add to the understanding that social support, such as emotional support, tangible or instrumental support, interaction or exchange support, and community support may influence subjective well-being outcomes to some extent no matter in high or low income countries. However, it is pretty certain that more diversified especially hard and accurate social support may definitely improve the possibility of subjective well-being in high income countries.

Besides, social support in countries with high and low generosity with cultural characteristics also has a significant positive impact on subjective well-being; however, countries of high-level generosity have higher positive effects than that in low-level ones. This may be based on the reason that people in high generosity countries are more willing to help others. Generosity itself as a voluntary behavior can bring people high subjective well-being. The more social support people receive; the more generous act they give to society, which can improve more subjective well-being to themselves finally. This is similar to the existing literature. For example, Kyung-Hee (2012) denoted that generosity (prosocial behavior) means try one's best to help or benefit others with happy, and their voluntary behavior of helping others do not expecting a return. However, in low generosity countries, people are less likely to give, unwilling to dedication but more inclined to ask for help or support for free. As a result, their subjective well-being stays at the material level, and they will not get real spiritual pleasure and realize their personal values in this society. This finding tells us social support also can influence subjective well-being outcomes to some extent no matter in high or low generosity levels countries with different cultural characteristics. People would respond positively to the needs of others, which in turn has the potential to generate more prosocial behaviors (Twenge et al., 2007). In view of this, it is a good way to advocate prosocial behavior because generosity itself can improve individual's subjective well-being.

4.2.   Moderating effects analysis

The moderating effect on the relationship between social support and subjective well-being at different income and generosity levels are also heterogeneous. To begin with, democracy and freedom both has a significant positive regulating effect in low-income countries. This may be explained that promoting a high-quality democratic government may be the best way for people in low income countries with conditions of insufficient living necessities, poor medical conditions, and weak social security system in order to get a positive impact of social support on subjective well-being. This is similar to the previous research, such as: democracy is positively related to national well-being and has the meaning of good governance or the quality of government (Ott, 2010; Ott, 2015). Besides, freedom also has a significant positive regulatory effect in low-income countries. This may be based on the fact that economic scarcity is one of the most important factors limiting people's choices. In low income countries, freedom can increase the growing demand for resources greatly. The findings of Inglehart also show the similarities. The extent to free choice can be improved by economic development, democratization and increasing social security, which in turn has led to higher levels of happiness around the world (Inglehart et al., 2008).

Moreover, democracy and freedom both have more significant effect in high generosity countries with cultural characteristics than the lower ones. This may be based on the reasons that high generosity countries generally have higher influence on speech rights and public decision-making, as well as more moral advantages on political and moral aspects, which promotes more unconditional gratitude and support from their citizens when compared with low generosity countries. This is the same to the existing literature that generosity is not only promoting harmony between the government and people, but also increasing more support for their dominant democracy in turn when it acted as a pro-social behavior (Aristotle, 2011).

5.   Conclusions

The key findings of this study are summarized as below: the differences of economic levels and culture are the two influential factors on the heterogeneous impacts of social support on subjective well-being. However, the degree of impact varies when it is associated with levels of income and generosity. Democracy and freedom have a constant significant positive adjustment effect on the relationship in both high and low generosity culture-characterized countries; such moderating effect is also associated with the level of income.

This has significant implications. It not only confirms the notion that social support significantly impacts on subjective well-being, but also contributes to new insight that the impact is heterogeneous according to economic and culture differences that have different moderating effect. The significance of this finding shapes the conception of economic dominated social support for well-being. It can be suggested that future social and public health policy shall focus on balancing (or shifting) towards building generosity and democratic societies alongside economic support. Countries should enrich non-economic social support channels in order to increase subjective well-being based on its own economic status and cultural characteristics. For example, increasing more functional social support after the income level exceeds the critical value; combining social entrepreneurship and entrepreneurial behavior to fight poverty and inequality (Kruse, 2020); considering PPS (Private Pension System) to help individuals make savings and make contributions to the nation's economy (Demirtas and Kececi, 2020). In terms of culture, advocating caring for people, and establish good social atmosphere of personality equality and mutual respect. Moreover, for the moderating effects, such as establishing social trust mechanisms, carrying out the public construction of public resources and effective supervision in countries with higher levels of generosity; whereas in lower generosity countries, enhancing the personal ethical values of the wealthy class, so as to promote the ethics of resource sharing and fulfill the obligation of public well-being.

It is worthy to note the limitations of this study—due to the limited data available and the specific model and variables selected. Future research could further analyze the heterogeneous influence of social support on subjective well-being from different perspectives, e.g. consider other indicators such as: energy ecology (Goers and Schneider, 2019), human resources (Sánchez-Ollero et al., 2020; Syed et al., 2020), etc. other economic indicators in the field of economic differences, such as economic credit (Kartal, 2020; Li et al., 2020), financial cycle (Liu et al., 2020), economical risk (Sukharev, 2020), etc. Moreover, study can consider alternative variable measures of subjective well-being in different countries from Gallup, diversified background with specific country/context so that specific policy and intervention mechanisms can be initiated and implemented.

Acknowledgments

We would like to thank the funding support from Education Special Project of National Social Science Fund of China [grant number: BIA180197].

Conflicts of interest

All authors declare no conflicts of interest in this paper.