Motherhood delay, polarized return of education and fertility supporting policies

LIU Feng1 HU Chunlong2

(1.School of Economics, Dongbei University of Finance and Economics, Dalian 116025, China 116025)
(2.School of Economics, Shanghai University of Finance and Economics, Shanghai 200433, China 200433)

【Abstract】The implementation of the universal two-child policy and the complete release of dividend policy reform need to actively improve all kinds of fertility supporting policies on the basis of the top-level design. From the perspective of motherhood delay, the paper provides empirical evidence on how to improve fertility supporting policies. Based on the revised Mincer wage function and using the CHIP2008 and CHIP2013 datasets, the impact and the mechanism of motherhood delay on the return of education are studied. The paper proposes a research hypothesis that motherhood delay will polarize the return of education. Theoretical analyses show that there are two reasons—active and passive delay—for women delaying childbearing age, which corresponds to the two results—increase and decrease of income. The difference comes from different jobs chosen due to education. The empirical results are consistent with theoretical analyses by econometric empirical studies about effects of motherhood delay on women’s income. Motherhood delay has polarized effects on the return of different education, which increases the return of high education and reduces the return of low education. In order to guarantee robustness of empirical results above, the paper performs deep analyses from four perspectives: extending the length of data, dealing with the ability variable omitted, changing the measure index of motherhood delay, and considering nonrandom behaviors of fertility. Specifically, CHIP2008 is added on the basis of CHIP2013, the problem of omitted abilities is solved using the instrumental variable method, motherhood delay is measured using continuous variables, and self-selection is solved using the framework of counterfactual analyses and the method of propensity score matching. The robust test shows that the empirical results still hold. Further studies show that polarized effects come from the difference in job stability and difficulty of professional title or leading roles. Specifically, women with high education mainly engage in jobs with more stable and easily obtaining professional titles or leading roles, while women with low education usually engage in jobs with the opposite. When choosing motherhood delay, the income of the former increases, while the latter decreases. Hence, the paper argues that different fertility supporting policies should be made among different education groups to improve the women childbearing willingness. Specifically, income compensation policies should be made for women with low education to overcome the “willing but unable to give birth” problem, and professional development compensation policies should be made for women with high education to settle the “able but unwilling to give birth” phenomenon.

【Keywords】 motherhood delay; polarized return of education; fertility supporting policies;

【DOI】

【Funds】 General Project of National Natural Science Foundation of China (71773012) Youth Project of National Social Science Fund of China (13CJY010)

Download this article

    Footnote

    [1]. ① Source: National Bureau of Statistics: Sixth Census, http://www.stats.gov.cn/tjsj/pcsj/rkpc/6rp/indexch.htm [^Back]

    [2]. ① Scholars in China and other countries use various measurement methods to correct the above biases to ensure the accuracy of the estimation results of educational yield. In recent years, Sun (2014) used the intra-group difference method to estimate the rate of return on education based on twin data, and found that omitted variables such as ability will lead to OLS overestimation. Jian and Ning (2013) used the propensity score matching method to find that the OLS estimation results of the 1997 data were upward biased, while the OLS estimates of the 2006 data were downward biases. Deng (2013) conducted a systematic discussion on related adjustment methods and problems. Based on existing research experience, it is known that omitted variables will generally cause overestimated OLS, while measurement errors will cause underestimated OLS. [^Back]

    [3]. ① The choice of optimal childbearing age depends mainly on the comparison between the marginal income and the marginal cost of the childbearing age. The benefits of delayed childbearing age are mainly due to the premium of wages, the increase of income, the success of the career (the rise of women’s social status) and the promotion effect of children’s cognitive ability (Miller 2011); while the cost of delayed childbearing age comes from physiological effects (difficulties in conception after the missing of women’s vigorous ovulation period), increased cost of matching spouses, and increased health risks for mothers and their children (Bratti and Tatsiramos, 2012). In addition to the determination of income and cost, the choice of childbearing age is also influenced by social culture and policy systems. For example, the lack of family-friendly businesses and employers’ discriminate against child-rearing employees can lead to delays in women’s childbearing age, advocation of later marriage and later childbearing and late-childbearing holiday system will weaken women’s preference of early childbearing (Budig and England, 2001). [^Back]

    [4]. ② By dividing the whole nation into the eastern, central and western regions, the sample data of CHIP2013 is gained from the National Bureau of Statistics’ 2013 large-scale sample library of urban and rural integration regular household surveys (this sample library covers 160,000 households of all 31 provinces, municipalities and autonomous regions) according to the systematic sampling method. The CHIP2013 sample contains 18,948 household samples and 64,777 individual samples. [^Back]

    [5]. ① The first category of industries includes 09 (information transmission, software and information services), 10 (financial industry), 13 (scientific research and technical services), 14 (water, environment and public facilities management) and 16 (education); the second category includes 02 (mining), 04 (electricity, gas and water production and supply), 05 (construction), 11 (real estate) and 19 (public administration, social security and social organization); the third category includes 03 (manufacturing), 07 (transportation, warehousing and postal services), 12 (lease and business services), 17 (health and social work) and 18 (culture, sports and entertainment); and the fourth category includes 01 (agriculture, forestry, animal husbandry and fishery), 06 (wholesale and retail), 08 (accommodation and catering), 15 (residential services, repairs and other services) and 20 (international organizations). [^Back]

    [6]. ① According to the practices of Liu (2008) and Huang and Yao (2009), we set dummy variables: ethnic minorities, members of the Communist Party of China, healthy, poor health, first category of industries, second category of industries, third category of industries, first category of employer, second category of employer, unit leader, technician and service personnel. [^Back]

    [7]. ① In theory, there are many ways to achieve matching, and they are all progressively equivalent. However, in the actual application process, Caliendo and Kopeinig (2008) pointed out that the trade-offs between deviation and efficiency are different for each method. Therefore, the actual matching results of various methods may have some differences. In view of this, in order to ensure the robustness of the matching results, this paper considers a variety of matching methods, including the K-nearest-neighbor (1-k matching), kernel matching method, caliper matching method and spline matching method. [^Back]

    [8]. ② The common support domain is mainly to reduce the samples loss of intervention group in the matching process as much as possible. The standardized deviation test method of Rosenbaum and Rubin (1985) suggests that if the standardized deviation of the explanatory variable between the intervention group and the control group is greater than 20% after matching, it means the failure of the matching. [^Back]

    References

    [1] Deng, F. Education & Economy (教育与经济), (5): 42−48 (2013).

    [2] Huang, Z. & Yao, X. The Journal of World Economics (世界经济), (7): 74–83 (2009).

    [3] Jia, N., Gan, L. & Zhang, J. Economic Research Journal (经济研究), (5): 61–72 (2013).

    [4] Jian, B. & Ning, G. Economic Research Journal (经济研究), (2): 83–95 (2013).

    [5] Li, S. & Ding, S. Social Sciences in China (中国社会科学), (6): 58–72 (2003).

    [6] Liu, Z. Economic Science (经济科学), (2): 119–128 (2008).

    [7] Sun, Z. China Economic Quarterly (经济学(季刊)), (3): 1001–1020 (2014).

    [8] Ashenfelter O, Rouse C. Income, schooling, and ability: Evidence from a new sample of identical twins. The Quarterly Journal of Economics, 1998, 113(1): 253−284.

    [9] Blackburn M L, Bloom D E, Neumark D. Fertility timing, wages, and human capital. Journal of Population Economics, 1993, 6(1): 1−30.

    [10] Brandt L, Holz C A. Spatial price differences in China: Estimates and implications. Economic Development and Cultural Change, 2006, 55(1): 43−86.

    [11] Bratti M, Tatsiramos K. The effect of delaying motherhood on the second childbirth in Europe. Journal of Population Economics, 2012, 25(1): 291−321.

    [12] Buckles K. Understanding the returns to delayed childbearing for working women. American Economic Review, 2008, 98(2): 403−407.

    [13] Caliendo M, Kopeinig S. Some practical guidance for the implementation of propensity score matching. Journal of Economic Surveys, 2008, 22(1): 31−72.

    [14] De Brauw A, Rozelle S. Reconciling the returns to education in off-farm wage employment in rural China. Review of Development Economics, 2008, 12: 57−71.

    [15] Erosa A, Fuster L, Restuccia D. Fertility decisions and gender differences in labor turnover, employment, and wages. Review of Economic Dynamics, 2002, 5(4): 856−891.

    [16] Kalist D E. Does motherhood affect productivity, relative performance, and earnings? Journal of Labor Research, 2008, 29(3): 219−235.

    [17] Li H Z, Luo Y. Reporting errors, ability heterogeneity, and returns to schooling in China. Pacific Economic Review, 2004, 9(3): 191−207.

    [18] Miller A R. The effects of motherhood timing on career path. Journal of Population Economics, 2011, 24(3): 1071−1100.

This Article

ISSN:1001-9952

CN: 31-1012/F

Vol 44, No. 08, Pages 31-45

August 2018

Downloads:0

Share
Article Outline

Abstract

  • 1 Introduction
  • 2 Research hypothesis and theoretical analysis
  • 3 Empirical research design and data description
  • 4 Empirical analysis
  • 5 Mechanism verification of polarizing rate of return on education
  • 6 Conclusions and policy recommendations
  • Footnote

    References