Comparison of second-child fertility intentions between local and migrant women in urban China: a Blinder–Oaxaca decomposition

ABSTRACT

With China's termination of the longstanding one-child policy and its implementation of a universal two-child policy since 2016, it remains an open empirical question whether the Chinese, including more than 200 million rural-to-urban migrants, are willing to have a second child. Using the 2017 China Migrants Dynamic Survey data, this study compares the intention of having a second child between urban local women and rural-to-urban migrant women in Chinese cities. We find significantly lower second-child fertility intentions among migrant women, despite their younger average age than local women. Employing the Blinder–Oaxaca decomposition technique borrowed from labour economics, we reveal that education and son preference both play particularly prominent roles in explaining the lower second-child fertility intentions among rural migrants. First, more education is found to promote second-child fertility intentions in urban China. Rural migrants’ fertility intentions are depressed by their lower educational levels. Second, in urban China, when the first child is a boy, a couple tends to have a lower intention to have a second child. This fertility-depressing effect of already having a son is particularly pronounced among rural migrants, and moreover, compared with urban locals, a higher percentage of rural migrants’ first child is a son.

KEYWORDS:

• Rural-to-urban migration
• fertility intentions
• second child
• China
• housing
• Blinder–Oaxaca decomposition

Acknowledgements

Dr. Wei Guo is particularly grateful to his mother, Mrs. Bifang Wang, for firing his impulse to co-write this research paper on fertility policy in China with his best friend. The brave mother ignored the one-child rule and bore the burden of heavy financial penalties to have continued childbearing in pursuit of her original aspiration of producing one more offspring to carry on the family line and to provide for old-age security in western rural China in the late 1980s.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Based on latest research on patterns of rural-to-urban migrants in China, the age distribution of migrants is mainly concentrated between 15 and 49 (accounting for 89.5% of the total migrant population) (Xue, Wang, and Chang 2019). Rural-to-urban migrants display a very typical ‘young and middle-aged’ characteristic, and the average sex ratio is between 113.1 and 141.3 across different cities (Xue, Wang, and Chang 2019). The composition of the migrant population has shifted significantly over time and recent migrants are more likely to be female and to be married (Colas and Ge 2019). Compared with rural locals, rural-to-urban migrants marry significantly later (Mu and Yeung 2019).

2 While the CMDS employs various techniques to minimize non-responses (such as training interviewers, offering respondents incentives, follow-up visits, personalizing the contact, etc.), it does not provide a definite response rate, however. The support and supervision from the National Health Commission, a cabinet-level executive department of the State Council, also helped implement the survey and lower the non-response rate. There is no obvious concern for missing data as very few missing data are present in our data analyses. Only 79 out of 6942 observations (only 1.14%) miss data on some variables. We also tried multiple imputation to handle these few missing data and all the results were substantively the same.

3 In demography or population studies, women of reproductive age refer to all women aged 15–49 years. However, the age requirement for marriage is 20 years of age for women and fertility mainly occurs within marriage in China, so we limit our sample to those respondents who are married and aged from 20 to 49 years.

4 We exclude urban-to-urban migrants (urban hukou holders residing in a city different from the city where their hukou is registered) due to the focus of this study and the small sample of urban-to-urban migrants in the CMDS data. Nevertheless, urban-to-urban migrants are a unique group distinct from both rural migrants and urban locals, whose fertility intentions are worth studying in future research with new suitable data.

5 The Blinder-Oaxaca decomposition works the best for linear regression models, especially when we need to identify the contribution of each explanatory variable to the overall disparity in the dependent variable (O’Donnell et al. 2008). While scholars (Bauer and Sinning 2008; Sinning, Hahn, and Bauer 2008) have developed methods that can perform the Blinder-Oaxaca decomposition for nonlinear models, these methods can ‘not separate the contributions of single variables’ (Sinning, Hahn, and Bauer 2008, 481). We thus use the OLS regression when conducting the Blinder-Oaxaca decomposition. We find no substantive differences in the results from the ordinal logistic regression and from the OLS regression.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China under grant number 71921003, the Insight Grant from the Social Sciences and Humanities Research Council (SSHRC) of Canada, and Zijin Media Think Tank.