Social Norms and Gender Differences in Labor Force Participation in China

ABSTRACT

Since China’s transition to a socialist market system, women’s labor force participation has declined sharply. Using data from the Chinese General Social Survey (CGSS) 2010, the authors re-examine China’s gender gap in labor force participation with a focus on social norms. Probit model estimates of the gender gap in labor force participation probability confirm the contribution of conventional factors such as health and other human capital as important explanations for the resurgent gender gap in China. However, even after extensive controls for human-capital differences, the gap remains mostly unexplained in the data. The Blinder–Oaxaca decomposition analysis confirms that the gender gap in labor force participation is predominantly explained by behavioral differences between women and men. Gender-related community social norms account for 41.4 percent of the unexplained gap. The study results are robust to alternative measures of social norms and additional controls for community characteristics.

HIGHLIGHTS

• Recent studies have examined why women’s labor force participation has declined sharply in post–economic reform China, but research on the effects of social norms has been limited.

• The gender gap in participation remains mostly unexplained by differences in human capital between women and men.

• Social norms account for almost half of the unexplained portion of the gender gap in labor force participation.

• Along with expanding women’s access to education and improving their health, policies should promote favorable attitudes toward women’s employment, particularly during times of economic crisis.

KEYWORDS:

• Gender inequality
• social norms
• health capital
• China

JEL Codes:

• I14
• J16
• J22

Acknowledgments

This study is the outcome of “The China Model: Implications of the Contemporary Rise of China (Ministry of Higher Education, Government of Malaysia, High-Impact Research Grant)” project UMC/625/1/HIR/MOHE/ASH/03. Data analyzed in this paper come from the research project “Chinese General Social Survey (CGSS)” of the National Survey Research Center (NSRC), Renmin University of China. We would like to thank participants at the 2017 International Conference on Applied Economics and Policy (ICAEP) in Kuala Lumpur, Malaysia, and seminar at Universiti Brunei Darussalam (UBD) for useful comments. The views expressed herein are the authors’ own.

SUPPLEMENTAL DATA

Supplemental data for this article can be accessed at https://doi.org/10.1080/13545701.2020.1758337

Notes

1 The rate of women’s participation based on population census data for urban China was even higher – 89.4 percent in 1990 (Wu and Zhou 2015).

2 At the world level, between 1995 and 2015, the female labor force participation rate fell from 52.4 to 49.6 percent. The male–female gap in labor force participation and employment rates has also widened in other East Asian and South Asian countries (ILO 2016).

3 Based on the World Bank World Development Indicators (WDI) data, the Gini coefficient jumped from 0.291 in 1980 to 0.421 in 2010.

4 For the causes and consequences of changes in female labor force participation in developing countries see Rachel Heath and Seema Jayachandran (2016).

5 In addition, the reform on collective firms led to the loss of approximately 13 million Chinese jobs between 1996 and 2000 (Fewsmith 2001).

6 This study tested the hypothesis that women in China are more attached to family and have less incentive to work when experiencing a wealth gain. The authors find that a 100-thousand-yuan increase in housing value leads to a significant decrease in female homeowners’ labor force participation probability and an increase in the probability of becoming housewives.

7 Wei Chi and Bo Li (2014) also provided potential causes for this widening gender gap in labor force participation and employment, which include high childcare costs, diminished availability of publicly funded childcare centers, the re-emergence of the traditional gender division of labor, and income effects. In addition to discriminatory measures against women in the workforce, such as denied employment or promotion, or being laid off more quickly, all contribute to their lower employment rate and the observed gender gap.

8 Existing research found social norms to be correlated with gender differences in female labor force and paid work participation decisions in developing countries (Contreras and Plaza 2010; Dildar 2015; Asadullah and Wahhaj 2019; Codazzi, Pero, and Sant'Anna 2018; Bursztyn, González, and Yanagizawa-Drott 2018) and developed countries (Bertrand et al. 2018).

9 For a review of the social science literature on competing definitions of social norms, see Rebecca Pearse and Raewyn Connell (2016).

10 For more theory of labor supply please see Mark R. Killingworth and James Heckman (1986), Richard Blundell and Thomas MaCurdy (1999), and Janna Besamusca et al. (2015).

11 Apart from social norms, no former studies have simultaneously considered health human capital and cognitive skill as factors that could affect female labor force participation (for example, see Mammen and Paxson [2000]; Klasen and Pieters [2012]; and Cunningham [2001]) For a review of the available research on gender inequality by scholars in China, see Qi Wang, Dongchao Min, and Bo Ærenlund Sørensen (2016).

12 There is only one exception for the case of urban China; Xi Chen and Suqin Ge (2018) attribute the low labor force participation among married women to intergenerationally transmitted social norms in terms of men’s gender role preferences.

13 An action plan “Beijing Declaration and Platform for Action” which aims to ensure women’s equal rights in economic, social, cultural, and political fields and remove their obstacles in participation in both public and private fields, was adopted in that conference.

14 During the 1980s and early 1990s, the proportion of the labor force with a tertiary education increased significantly due to the reopening of universities to the public after the Cultural Revolution. This proportion experienced only a slight increase during the 1990s. Therefore, in 1999, the Chinese government decided to increase the enrollments in higher education by approximately 40 percent.

15 Although the majority of the former studies restricted the sample to individuals ages 16–55 for women and 16–60 for men (16 years old is the lowest legal working age in China; 55 and 60 are the official retirement ages for women and men, respectively), in this study we follow T. Paul Schultz’s (2002) standard: women ages 25–55 and men ages 25–60.

16 We follow the division of Shuaizhang Feng, Yingyao Hu, and Robert Moffitt (2015) on employment and labor force participation.

17 The original third statement in the CGSS 2010 questionnaire is actually, “Housework should be shared by couples equally.” Here, in order to present results consistently with the first and second statements, we paraphrased the third statement as, “Housework should not be shared by couples equally” and the answers correspondingly.

18 We follow the recoding method of Chen and Ge (2018) on recoding responses to these social-norms related statements. If we recode “neutral” as “0,” the community norms show larger impacts on women’s waged work participation decisions compared with current results as shown in Table 3.

19 Here, community (she qu) refers to a residential area under a district according to the geographical/administrative division. Our data set covers 478 communities in mainland China, and around twenty-five households are sampled in one community. For studies that use this same concept of “community,” please see Qingwen Xu and Julian C. Chow (2006).

20 Another study on the effects of gender inequality in China on education, healthcare, and pension benefits is Ke Shen, Feng Wang, and Yong Cai (2016).

21 For an empirical examination of gender differences in cognitive abilities in China see Xiaoyan Lei et al. (2014).

22 Some existing studies found that young children are more likely to affect the decision of labor force participation instead of older children. For example, Dante Contreras and Gonzalo Plaza (2010) found that “children under the age of 4: number of preschool children living in the home” has a significantly negative impact on women’s participation in the labor market in Chile. As this variable is unavailable in the CGSS dataset, we are only able to investigate the effect of the number of children regardless of their age.

23 Although cognitive skills have multiple dimensions, here our use of English language proficiency is only a proxy measure of cognitive skills. Existing papers in the literature not only consider the important roles of formal schooling and cognitive skills on labor market participation and wages, they also examine the impact of noncognitive skills in China (see Glewwe, Huang, and Park [2017]).

24 There are significant spatial variations in social norms relating to gender relations in China. For instance, Yang Hu and Jacqueline Scott (2016) note that individuals from the industrialized regions in China exhibit significantly less traditional attitudes toward patrilineality (with the exception of eastern towns/villages) compared to those from the western rural towns/villages. In general, major cities and central towns are least likely to show support for traditional gender beliefs.

25 Other recent studies on China that implemented the Oaxaca method to study gender differences include Rui Zhao and Yaohui Zhao (2018), Liangshu Qi and Xiao-yuan Dong (2016), and Maria Karamessini and Elias Ioakimoglou (2007) on gender pay gaps in other countries.

26 Now, in equation (3)(3) ${\overline{P}}_M\text{--}{\overline{P}}_F=\left({\overline{X}}_M\text{--}{\overline{X}}_F\right){\widehat{\text{β}}}_M+\overline{X}\left({\widehat{\text{β}}}_M\text{--}{\widehat{\text{β}}}_F\right)$(3) , X vector includes the English language skill variable; three measures of health capital: height, self-reported health status and body-mass index (BMI); and three community social norms covariates as well.

27 We acknowledged the problem that the estimates for the community social norms might suffer endogeneity. However, in this dataset, we are not able to find any appropriate IV to exploit exogenous shock to attitudes, for example. Therefore, instead of using the average answer for social norms in the community, we also test the effect of social norms on labor force participation decisions based on different reference groups (against women only and against men only in the community), and the results do not change. So, our results are robust to different choice of reference group.

28 Using nationally representative survey data from twenty-nine countries Kathleen L. McGinn, Mayra Ruiz Castro, and Elizabeth Long Lingo (2019) find that adult daughters raised by employed mothers are more likely to be employed than daughters whose mothers are full-time housewives.

29 Although Chen and Ge (2018) indicate that the paid work behavior of a woman’s own mother is not significant in explaining the woman’s participation decision in urban China, the labor force participation rate of married women with nonworking mothers-in-law is significantly lower than those with working mothers-in-law.

30 This unexplained percentage is quite high; other papers based on data from other developing Asian countries found that the gap is also dominantly explained by the behavioral differences between women and men, where the unexplained part accounted for 90.88 percent (Abdulloev, Gang, and Yun 2014; Asadullah and Wahhaj 2019).

31 Detailed results for these subsamples are also available upon request.

Additional information

Funding

This work was supported by MOHE High-Impact Research Grant, Government of Malaysia: [Grant Number UMC/625/1/HIR/MOHE/ASH/03].

Notes on contributors

Saizi Xiao

Saizi Xiao is a final-year PhD student at the University of Malaya, Malaysia and an affiliate of the Global Labor Organization (GLO). Her field of research is gender and development economics with a focus on the Chinese labor market.

M. Niaz Asadullah

M. Niaz Asadullah received his doctorate in economics from Oxford University. He is Professor of Development Economics at the University of Malaya, Malaysia. He has held research/teaching/visiting positions at Oxford, Harvard, Reading, Manchester, Kent, Dhaka, Mindanao State, and BRAC Universities. His research relates to poverty, well-being, labor market, skills formation, education, and gender issues in South and East Asia. Professor Asadullah is an Associate Editor of the Journal of Human Development and Capabilities, Asia Pacific Education Review, and BMC Public Health. Other affiliations: Fellow, the IZA Institute of Labor Research; Southeast Asia Lead, the Global Labor Organization; and Co-founder, Initiatives on Education, Gender and Growth in Asia (IntEGGrA).