Gender Inequalities in Labor Market Outcomes of Informal Caregivers near Retirement Age in Urban China

ABSTRACT

This study examines the impacts of unpaid family care on labor supply and earnings of women and men near retirement age in urban China. Using the 2011 China Health and Retirement Longitudinal Study (CHARLS) and ordinary least squares (OLS) and instrumental variable approaches, it finds that grandchild care is negatively associated with both women's and men's labor force participation, while there are no effects for eldercare. For women caregivers, caring for grandchildren substantially lowers paid labor hours compared to noncaregivers. No significant relationships are found between eldercare and paid labor hours of women workers. For men workers, neither grandchild care nor eldercare is significantly associated with labor hours. The study also finds no statistically significant relationships between grandchild care and labor earnings for either women or men. Eldercare, however, is positively associated with the earnings of men workers.

KEYWORDS:

• Informal care
• childcare
• eldercare
• labor supply
• earnings
• China

JEL Codes:

• I11
• J1
• J22

ACKNOWLEDGMENTS

The authors are grateful for valuable suggestions and comments by Xiao-yuan Dong, Yaohui Zhao, Edgard Rodriguez, Margaret Maurer-Fazio, three anonymous referees, and participants at different seminars and conferences. Financial support from the International Development Research Centre of Canada (Project no. 107579) is gratefully acknowledged. Yafeng Wang is grateful for financial support from the National Natural Science Foundation of China (71603013). Chuanchuan Zhang is grateful for financial support from the National Natural Science Foundation of China (71503282) and the Young Elite Teacher Project of Central University of Finance and Economics (QYP1609). All remaining errors are the authors’ own.

Notes

1 We provide a simple theoretical framework in the Appendix to illustrate the effects of care burden on labor supply.

2 See Heckman (1974); Blau and Robins (1988); Ribar (1991, 1992, 1995); Berger and Black (1992); Connelly (1992); Leibowitz, Klerman, and Waite (1992); Michalopoulos, Robins, and Garfinkel (1992); and Kimmel (1998), who use survey data from the US.

3 We use the linear probability model (LPM), rather than nonlinear models, such as probit or logit, because the results of LPM are easier to interpret. Also, we use the instrumental approach in resolving potential endogenous problems, and the IV estimation in the linear model is more robust than that in nonlinear models (Angrist and Pischke 2009). Nevertheless, we also try the two-stage residual inclusion (2SRI) approach, which is an IV-based approach used in nonlinear models, and obtain similar results.

4 The respondent is defined as active if he or she is self-employed, employed, or unemployed and currently looking for a job.

5 In one recent study, Lu Chen et al. (2017) emphasized the intensity of caregiving. To check whether our conclusions are robust to the definition of caregiving, we also use a more restrictive definition of caregiving: that is, if the respondent spends at least 20 hours per week on caring. The results are qualitatively the same as those obtained when we measure the care burden in terms of whether any care services are provided. The results are available upon request.

6 Hukou is the household registration status in China. The type includes agricultural and nonagricultural.

7 Approximately 30 percent of respondents in our sample have a retirement scheme, which may lead to different patterns of retirement decisions. Thus, we explicitly control for whether the respondents have a retirement scheme.

8 Ideally, we would use the existence of younger grandchildren, who are more likely to need intensive care. However, the CHARLS data only record the main respondents’ numbers of grandsons and granddaughters who are younger than age 16.

9 Elderly parents include both parents and parents-in-law.

10 We have also controlled for the number of sons and daughters, which is the main determinant of the existence of young grandchildren.

11 CHARLS is a biennial survey conducted by the National School of Development at Peking University, aiming to be representative of Chinese residents ages 45 and older and their spouses. CHARLS is part of a set of longitudinal aging surveys that include the HRS in the US, the English Longitudinal Study of Ageing (ELSA), the Survey of Health, Ageing, and Retirement in Europe (SHARE), the Korean Longitudinal Study of Aging (KLoSA), the Japanese Study of Aging and Retirement (JSTAR), and the Longitudinal Aging Study in India (LASI). More details can be found at CHARLS (n.d.); please also see Yaohui Zhao et al. (2013) for a complete discussion.

12 China has a mandatory retirement policy, requiring men and women workers in formal sectors to retire at ages 60 and 50, respectively. The labor force participation rate of workers older than age 65 in urban areas is less than 7 percent in the CHARLS sample.

13 Urban and rural in China can be defined either by hukou status or by type of residence. While the hukou status often refers to urban or rural, it is officially recorded as agricultural or nonagricultural in the government system of household registration. The hukou status is typically classified by whether the household owns agricultural land or not. The type of residence, which refers to urban areas or rural areas, is classified by local economic development level. Since the labor market is more closely related to the local development level, we restrict our sample to those living in urban areas.

14 The gap between active and employed includes those self-employed and unemployed.

15 To explicitly test the gender differences in the effects of caregiving on labor supply, we conducted a joint model analysis and present the results in Appendix Table A2. As shown in columns 2 and 4, the effects of grandchild care on paid work hours are statistically different by gender. However, we find no statistically significant gender-specific differences in the effects of grandchild care on labor force participation and labor earnings. There are also no statistically significant differences by gender in the effects of eldercare.

16 See Appendix Table A1 for the first stage estimation results.

17 We also explicitly test the gender differences in the effects of caregiving on labor earnings with a joint model analysis. The results are presented in Table A3. Although the results generally suggest a smaller effect of caregiving on labor earnings for men, the gender-specific differences are not statistically significant due to large standard errors.

Additional information

Notes on contributors

Yafeng Wang

Yafeng Wang is a Research Fellow in the Institute of Social Science Survey at Peking University. His research interests including labor economics, demography, and applied econometrics. He is currently Director of the Data Department for the China Health and Retirement Longitudinal Study (CHARLS), which is based on the Health and Retirement Study (HRS) and related aging surveys, such as the English Longitudinal Study of Aging (ELSA) and the Survey of Health, Aging, and Retirement in Europe (SHARE).

Chuanchuan Zhang

Chuanchuan Zhang is an Associate Professor in the School of Economics at the Central University of Finance and Economics. His research interests include labor economics, health economics, and public economics. He has published in Research in Labor Economics, Health Economics, Journal of Housing Economics, China Economic Review, and other journals.