Understanding the Intergenerational Transmission of Human Capital: Evidence from a Quasi‐natural Experiment in China

• Author: Aparajita Dasgupta,Dong Zhou
• Date: 01 May 2017
• Published date: 01 May 2017
• DOI: http://doi.org/10.1111/rode.12266

Understanding the Intergenerational Transmission

of Human Capital: Evidence from a Quasi-natural

Experiment in China

Dong Zhou and Aparajita Dasgupta*

Abstract

This paper exploits the closure of senior secondary schools in urban China from 1966 to 1971 in order to

identify the causal intergenerational transmission effects of education. The paper uses the instrumental

variable approach to examine the intergenerational causality of educational transmission at the senior

secondary schooling level in urban China. The exogenous variation in parental senior secondary

educational attainment both over time and across regions allows us account for selection bias and thus

identify the causal intergenerational transmission effect in education. We further show that our

conclusion is robust to alternative identification strategies and data sets.

1. Introduction

It is widely documented that there exists a positive correlation between parental

education and children’s schooling outcomes. However there is limited evidence on

what discerns the causality and selection effects in this positive correlation,

1

which

are crucial to identify from a policy perspective. First, if children’s education

attainment can be improved by their parents, because better schooling of parents

has shaped their behaviors and personal traits in a certain way, the social benefits

of supply of public education might be underestimated. This would be important in

the sense that it can provide insurance for initial conditions to curb lifetime

inequality reaching further. Second, understanding the important factors of nurture

is also broadly beneficial from a policy standpoint. It can direct the policies that pay

attention to the working life time of parents (e.g. unemployment insurance or

subsidy with necessary requirements of educational expenditure imposed). Better

designed policies can help alleviate parents’ current budget constraints, avoid

potential ignorance of children’s schooling deficits during downturns and improve

the outcomes of their children for the sake of long-run development. In this paper,

we study empirically the intergenerational effects of transmission of education

taking the case of urban China and examine its causal impact using instrumental

variables (IV). We are also able to detect the patterns of gender heterogeneity in

this transmission that are consistent with the existing literature.

*Zhou: Department of Cultural Industry and Management, School of Media and Design, Shanghai Jiao

Tong University, No. 800 Dongchuan RD., Minhang District, Shanghai, 200240, China. E-mail: dong.

zhou@email.ucr.edu. Dasgupta: Department of Economics, Ashoka University, New Delhi, India. The

authors are grateful to Mindy S. Marks, Jorge M. Ag€

uero, Ullah Aman and Victor Lippit for their

insightful comments and suggestions which greatly improve the paper. We would like to express our

gratitude to the anonymous referees for their insightful comments. Also, they are incredibly grateful to

the National Bureau of Statistics of China, the Inter-University Consortium for Political and Social

Research of University of Michigan and the Chinese Family Panel Studies of Peking University for

sharing their data. All errors are the responsibility of the authors.

Review of Development Economics, 21(2), 321–352, 2017

DOI:10.1111/rode.12266

©2016 John Wiley & Sons Ltd

The formal education system was dramatically impacted from 1966 to 1976 by the

Cultural Revolution in China. The majority of senior secondary schools were shut

down between 1966 and 1971, especially in urban areas. Despite the school closures

being nationwide, intensities of implementation and exposure to the educational

interruptions varied across residential regions. Given the regional and time series

variations, we instrument this shock, which is arguably exogenous to parental ability

or genetic factors. Specifically, the interactions between parental ages and

residential area are used as instruments for parental senior secondary completion in

the empirical model. Both the IV estimate and difference-in-differences estimate

indicate the existence of a causal link between senior secondary attainment of

parents’ and children’s senior secondary attendance. Further, estimates consistently

indicate that the transmission from parents to daughters is slightly stronger as

compared with sons. The results are robust to alternative model specifications,

subsamples and data sources.

The primary contribution of this paper is that it is one of the first papers to

identify the causal relationship between parental senior secondary schooling

attainment and childrens’ senior secondary attendance, taking evidence from the

very disaggregated household data in the largest developing country, China. Most

of the existing literature is limited by its focus on reforms at lower levels of

education, for example, reforms at primary level schooling (Haveman and Wolfe,

1995; Holmlund, 2007; Oreopoulos et al., 2006). The focus of this paper is on a

relatively higher education level, the senior high attainment, which provides people

with important skills in the labor market of developing countries. Second, the

policy is involved with the large-scale closure of senior secondary schools and a

large affected population, and our empirical estimations utilize the 1990 wave of

China Population Census, a large, detailed and national representative data set. We

argue that implications of this paper would be very informative, especially in the

context of developing countries. In addition, the paper also adds to the literature on

identifying the long-term impacts of the Cultural Revolution on long-term

consequences of human capital formation, taking here the case of the

intergenerational transmission mechanism in higher education.

This paper is organized as follows. Section 2 reviews the current literature and

discusses the historical background. Section 3 describes the data and the empirical

strategy. Section 4 presents the empirical results along with robustness checks.

Section 5 offers concluding remarks.

2. Research Background

Literature

The literature on identifying the intergenerational “causality” from “association”

can be classified into several types sorted by identification strategies: identical twins,

adoptees and instrumental variables (including regression discontinuity designs).

Most of the twin-difference estimation shows a significant and positive effect of

fathers’ education on children’s schooling achievement (Ashenfelter and Krueger,

1994; Ashenfelter and Rouse, 1998; Antonovics and Goldberger, 2005; Behrman

and Rosenzweig, 2002; Pronzato, 2009). However, the causal effect of maternal

education on children’s school outcomes is ambiguous. The adoption strategy also

aims at taking out the genetic transmission between parents and biological children

and identifying the causal impacts. The reason for this is that the absence of genetic

322 Dong Zhou and Aparajita Dasgupta

©2016 John Wiley & Sons Ltd

transmission in the case of adopted children would enable the difference estimators

to capture the causal intergenerational effect in educational outcomes. Bjorklund

et al. (2006) and Sacerdote (2000, 2007) find supporting evidence for the causal

relationship between paternal education achievement and children’s educational

performance. By contrast, Behrman and Rosenzweig (2002) with twins’ data and

Plug (2004) argue that the positive influence of a mother’s schooling on that of her

child disappears when heritable abilities and assortative mating are taken into

account. Haegeland et al. (2010) using the twin identification strategy does not find

any impacts of intergenerational transmission on examination marks and grade

repetition. Apart from contesting conclusions, there are still two other concerns in

these papers: sample limitation (e.g. small sample size) and individual specifics (e.g.

unobserved differences within twins, unobserved parental characteristics and

selective placements of adopted children).

Third, studies utilize educational reforms as instruments for parental education

and examine the existence of causal effects.

2

Most of the exploited reforms are at

the lower level of schooling, for example, primary schooling reforms (Chevalier,

2004; Black et al., 2005; Holmlund et al., 2010; Oreopoulos et al., 2006). Using

higher levels of schooling as outcome variables, Maurin and McNally (2007)

explore the reform of expanding a more academic track before entering into

college in Northern Ireland. Additionally, Carneiro et al. (2013) resort to regional

variation in schooling costs as instruments. However, the results from these papers

vary across regions as well as institutions. For example, Black et al. (2005) find an

insignificant intergenerational causal effect of the Norwegian Primary School

Reform. In contrast, Holmlund et al. (2010) suggest that intergenerational school ing

associations are largely driven by selection. Ag€

uero et al. (2010) evaluate changes

in Zimbabwe and find the presence of causal intergenerational transmission using

the RDD methodology. The above discrepancy might be caused by the limitation of

the instruments used and differences in economic environments that vary by setting

in the developed and developing countries. This study will follow the IV approach

to examine the intergenerational causality of educational transmission at a

relatively higher educational level, looking at senior secondary schooling, which has

important labor market consequences.

Historical Background: Educational Interruptions

The formal education system in China in general includes 6 years of primary

education, 3 years of junior secondary education, 3 years of senior secondary

education and 4 years of college. After completing junior secondary school,

individuals can choose between technical secondary school and regular senior

secondary school. Additionally, after graduation from senior secondary school,

individuals can enter junior college (2–3 years) instead of regular university study.

However, this trajectory was interrupted nationally by school closure during the

Cultural Revolution (Treiman and Deng, 1997; Li et al., 2013; Meng and Gregory,

2002; Zhang et al., 2007; Zhou and Hou, 1999).

The May 16 Notification in 1966 formally initiated the Cultural Revolution in

China. At the very beginning, all levels of school stopped offering lectures in urban

areas and most of the students at secondary schools as well as at universities, in

particular, the urban youth in large cities, such as Beijing, Wuhan, Guangzhou and

Shanghai (Chan et al., 1980; Macfarquhar and Schoenhals, 2006) arbitrarily joined

the Red Guards to fight against the “Four Olds,” which included old customs, old

HUMAN CAPITAL: INTERGENERATIONAL TRANSMISSION 323

©2016 John Wiley & Sons Ltd