In contemporary China, the overly rapid social transformation, the increasing social gap between rich and poor, and the changing family structures and lifestyles have made the mental health problems of elderly groups increasingly serious. Women, in particular, have always been seen as a group with a high prevalence of depression.
Among the many factors influencing the degree of depression in women, retirement has been identified as an important factor. On the one hand, the current retirement system in China deprives them of formal labor opportunities earlier, and on the other hand, longer life expectancy further increases their probability of becoming alone. Compared to the research on retirement behavior on depression in male groups, there is a lack of research on the effect of retirement on depression levels of female workers, based on which this paper examines the effect of retirement on depression levels of women.
The direction and extent of the impact of retirement on depression have been controversial in current studies. Some scholars believe that work can bring not only economic benefits but also many noneconomic benefits, such as structured activities, social platforms and opportunities, a sense of personal accomplishment, a sense of purpose, and a sense of daily meaning. Leaving the workplace may imply both economic and emotional crises, which in turn lead to unpleasant emotional experiences for individuals. However, many studies have come to the opposite conclusion, suggesting that individuals may become more enjoyable after retirement because they are away from the negative effects of work. Some scholars also believe that individuals can make the transition from work to retirement smoothly and thus do not result in significant changes in mental health.
Although the number of studies on the mechanisms of retirement on depression is relatively small, a number of empirical articles analyze the channels of retirement effects on health, and the same applies to depression as a part of mental health. In summary, the literature has explored the mediating mechanisms by which retirement affects health from three main perspectives.
The first perspective presupposes a change in individual time allocation before and after retirement and argues that the sum of nonwork time spent as family businesses, socializing, exercising, and sleeping increases after retirement, which in turn improves the well-being of retirees. For example, Zhao et al. (2017) found that employees' participation in exercise increases after they stop working, and individual health improves as a result. Binh Tran and Zikos (2019) in their analysis of the causal impact of retirement on health, noted that for women, the positive relationship between retirement and health can be attributed to an increase in physical activity. Heo et al. (2014) found that postretirement participation in active leisure and socialization with friends improves subjective well-being in older adults. Hagen et al. (2016) showed that individuals have longer sleep durations during the transition from work to retirement.
The second perspective reveals the path of the impact of retirement on health in terms of health investment motives. Most of these studies regard health as an investment, then people have a strong motivation to invest in health during working years due to high health investment returns, such as maintaining a job and obtaining wage rewards, and therefore have higher levels of health(Galama et al., 2013; Garrouste and Perdrix, 2021). After exiting the labor market, although the return to health investment may decrease, the increase in leisure time leads to lower health investment costs, such as having more time to sleep, exercise, and visit the doctor(Motegi et al., 2016).
The third perspective uses changes in resources, primarily economic and social resources, as an entry point, arguing that retirement affects individual resource ownership and availability before it affects health outcomes. Numerous studies have found that an individual's income and economic status are significantly and negatively impacted by retirement, and this abrupt reduction in economic resources is highly likely to lead to a decline in the quality of life of retirees, which is detrimental to their physical and mental health(Palomäki, 2019). On the other hand, social capital, such as social relationships, organizational activities, and social support, which depend on the work structure, may also be reduced or even lost due to retirement, causing individuals to feel lonely and lack of belonging, which may lead to negative emotions(Börsch-Supan and Schuth, 2014; Kauppi et al., 2021). Of course, if individuals are more capable of regaining these resources, the negative effects of retirement on health will not exist(Kim and Moen, 2002).
It is worth noting that both the direct effects of retirement on health and the indirect effects exerted through mediating variables vary from person to person. As highlighted by Gill et al. (2006), the causal relationship between retirement and depression is heterogeneous. On the one hand, it can be influenced by individual characteristics, e.g., gender, marital status, educational attainment, income, and perceptions of retirement. In a review of 85 empirical studies from all continents except Antarctica, Girgus et al. (2017) found widespread gender differences in depression among individuals aged 60 and older. For example, Brandão and Cardoso (2020) found a protective effect of retirement on depression in women, with retired women having a significantly lower probability of depression than those who were employed. However, among men, employment or retirement was not significantly associated with depression. Szinovacz and Davey (2004) found that marital status played an important role in retirement well-being. For individuals with a spouse, time away from work means increased time with their partner, and their lifestyle, activity participation, and physical and mental health are affected by each other (Lam and Bolano, 2019). Cohen and Janicki-Deverts (2012) found that years of education and income were negatively associated with psychological distress. Older adults with higher education have greater psychological adjustment and are better able to perceive physical, life, and occupational changes and thus have higher well-being (Belo et al., 2020). A study by Doorslaer and Koolman (2004) reported an association between income inequality and health inequality, with higher income groups enjoying more and higher quality than lower income groups. Laberon et al. (2019) found that the more retirees viewed retirement as a negative event, the less able they were to adjust to retirement, the lower their life satisfaction, the higher their depression, and the worse their physical health.
Based on the above summary, we found that the current academic research on retirement and depression still has two limitations. First, the research subjects are not specific enough and lack relevance. In terms of gender, most scholars currently tend to study both retired male employees and retired female employees and compare the cross-sectional differences between the two. This approach can help us visualize the gender differences in retirement utility, but it is not conducive to longitudinal comparisons between individuals of the same gender. The large differences between men and women, both in terms of job nature and personal characteristics, suggest that they should not be studied using the same theoretical framework and empirical model. For example, it is clearly inappropriate to use the dominant life behaviors of men in the analysis of women. In terms of occupation types, most of the existing studies are based on the labor market as a whole, and there are fewer specialized or heterogeneous studies on a particular occupation or industry. Second, the analysis of heterogeneity was relatively fragmented and unsystematic. Although an increasing number of studies have begun to focus on theoretical derivation, there are still many studies in which the more common practice in selecting influential variables is to directly base on the findings of previous studies, which may lead to some important factors being overlooked.
In this regard, this paper makes corresponding adjustments in the research design and theoretical analysis section, which is the contribution of this paper. On the one hand, this paper restricts the study population to the female group, focuses on analyzing the impact of retirement on the mental health of female workers, and conducts a heterogeneity analysis based on different job natures. On the other hand, this paper compares the popular theories in the field of retirement, selects the most suitable role transition theory for this paper, and adjusts it to the actual situation to form a more general retirement model. Specifically, this paper attempts to examine the effect of retirement behavior on women's depression and assess the heterogeneity of the effect across groups from the retirement model using data from the China Family Panel Studies (CFPS) for 2012, 2016 and 2018, combined with a fuzzy regression discontinuity and fixed effects model. The paper is structured as follows: the second part presents the research hypothesis of this paper based on role transition theory; the third part presents the research design and introduces the data sources, variable measures, and measurement methods; the fourth part reports the empirical results; and the fifth part provides further analysis, including the analysis of family, personal, and work-level heterogeneity among women and the analysis of the effect of retirement on depression among men. The sixth part presents conclusions and policy implications.