300 words essay outline

ContentServer1.pdf

Home >History homework help >Ancient history homework help >300 words essay outline

The Long-term Impacts of the Cultural Revolution: A Micro-Analysis

Dong Zhou

Abstract. This paper provides a micro-analysis of the long-term impacts of a particular historical event: the Cultural Revolution in urban labor market of China. Using the life-cycle models and synthetic cohort approach, I demonstrate that the Cultural Revolution produced a lasting negative effect on permanent income for the subjected birth cohorts and this effect was amplified from the middle 1990s to the early 2000s as the Chinese market economy increasingly evolved. Channels in the mechanism of the impact includes productivity determinants (educational attainment, work experience, and health), marriage, and personality. The conclusions are robust to a variety of controls for family background as well as contamination factors, examinations with various control groups, contemporaneous comparisons, and placebo tests.

1. Introduction

In the past decade, a growing branch of literature has emerged that examines the impacts of historic legacies, such as the origins of the colonizers in the Americas, the slave trade, and colonial institutions. These studies provide empirical evidence demonstrating the persistent effects of historical events on current economic developments (Acemoglu et al., 2001; Banerjee and Iyer, 2005; Engerman and Sokoloff, 1997; Glaeser and Shleifer, 2002; Ichino and Winter-Ebmer, 2004; Iyer, 2007; Nunn, 2007, 2008, 2009). A related article that investigated micro perspectives came from Dell (2010), who measured the impact of the institution of forced labor (Mining Mita) on current household consumption in Peru. My paper contributes to the literature by examining how a specific historic event affected individual economic development by quantifying the effects of the Cultural Revolution in urban China. As noted by Nunn (2008), the existing research still has shortcomings in its

The author would like to thank Jorge M. Ag€uero and Mindy S. Marks for their insightful com- ments and suggestions which greatly improved the paper. Moreover, I thank Richard B. Freeman for his insights and useful comments on this paper. I also thank Victor Lippit, David Farris, Aman Ullah, Anil B. Deolalikar, Richard Arnott, and all participants in the Applied Economics Collo- quium for useful comments. Furthermore, I also thank the anonymous referees for useful comments which benefited the paper substantially. At last, I am incredibly grateful to the National Bureau of Statistics of China, the Inter-University Consortium for Political and Social Research of University of Michigan, Chinese Family Panel Studies of Peking University, the World Health Organization and China Health and Retirement Longitudinal Study for sharing their data with me and to the National Natural Science Foundation of China (71503147) for financial support. All mistakes are on my own. Dong Zhou, Assistant Professor, 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: [email protected]

inability to distinguish the channels through which historical events matter today. This study adds dimension to the current literature by exploring the transmission mechanism from microeconomic perspectives, that is, the potential channels through which the Cul- tural Revolution persistently affected the exposed cohorts. As the largest developing country, China has experienced substantial growth since its

reform and opening to international trade. The Cultural Revolution (1966–76) was dee- ply implicated in China’s institutional transition away from a closed, planned economy to an open, market-oriented economy. It lasted for more than a decade, and during this period, dramatic and intricate policies were implemented nationally (Treiman and Deng, 1997). For example, all levels of schools were shut down for certain years. No formal higher education was provided for a decade. Seventeen million urban youth were ‘rusti- cated’ by being relocated to rural areas for years in what came to be known as the Send-Down Movement. Considering the uniqueness and large scale of these programs, it is intriguing to explore their current impacts on the urban labor market and to under- stand the economic gains and losses to the population that was primarily subjected to them. Research on the Cultural Revolution has gradually increased since the 1990s. Most

studies are descriptive, with few empirical papers in publication. One branch is to focus on the educational interruption. Meng and Gregory (2003a) quantified the loss in edu- cation attainment caused by the school closure. Subsequent studies gauged whether the educational interruption has induced heterogeneous returns to schooling between exposed cohorts and the others and they found insignificant effects (Meng and Gregory, 2003b). Han et al. (2011) explored their decisions of human capital reinvest- ment and found that the subjected generations are more likely to reinvest into their adult education. The other branch of study analyzes how the Send-Down Movement affected individual life courses by comparing the sent-down youth with those who were not sent down. Zhou and Hou (1999) showed that the Send-Down Movement has fun- damentally changed the life course of a generation of urban youth. They also observed that the sent-down youth earned more and had more schooling than their counterparts in 1993 which they attributed to economic reforms and changes in personality. Xie et al. (2008) showed that there is no benefits of the Send-Down experience based on sibling comparisons. Li et al. (2013) utilized a twin survey in five cities and found that staying longer in the rural areas has a positive effect on monthly earnings among sent-down twins. Did the forced rustication truly benefit the young people? As will be discussed later, selection problems and simultaneous exposure to different interruptions continue to cast doubts on the validity of the allegedly positive effects of the Send-Down Movement at individual level given the impossibility of observing the counterfactual outcomes.1

The urban population was exposed not only to education interruptions but also to heterogeneous effects of the rustication because of uneven implementation. Other conflicts also simultaneously affected them during the same period (e.g. public shaming). Estima- tions that simply focus on one aspect or measure at individual level might lead to bias and misunderstanding the subsequent influences of the Cultural Revolution since its complexity and the interrelationship of different conflicts. In contrast to the existing literature, this paper identifies the variety of exposures with the synthetic cohort approach from different perspectives. The cohort approach to some extent loses variations of individual idiosyn- crasies and has difficulties in disentangling age, period, and cohort effects (Barr and Lin, 2015; Glenn, 1989; Hanoch and Honig, 1985; Ryder, 1965; Sasaki and Suzuki, 1989;

286 Dong Zhou

Winship and Harding, 2008). Despite the limitations, it is plausible and informative to exploit the cohort approach for capturing how the concurrent policies simultaneously impacted the subjected cohorts’ permanent incomes based on deviations from the life-cycle model at different points of time, given such large-scale interruptions of Cultural Revolu- tion and its uniqueness. In this study, the main urban micro-data utilized is from Chinese Household Income

Project Survey (CHIPS). Estimates from different measures and samples consistently show that the Cultural Revolution significantly lowered the subjected individuals’ annual earnings, and this effect was amplified from the mid-1990s to the early 2000s as the market-oriented economy increasingly evolved. Comparing the school closures with the Send-Down Movement, empirical evidence shows that the latter policy played a signifi- cantly stronger adverse role. After establishing the existence of the impacts, I further look into the channels through which this adverse role matters, including productivity determi- nants (e.g. schooling years, working experience, and health status), marital status and atti- tudes. Educational attainment is the major channel, accounting for approximately 50 per cent of the overall effect. Interestingly, changes in the estimates imply that the Send-Down Movement shaped the sent-down youth’s personalities in a positive way by making them more perseverant. Their attitudes toward the different determinants of personal success contributed to their earnings. These findings are shown to be robust to evaluations with different control groups and to a variety of controls for family background and confound- ing factors (e.g. the Great Famine: 1959–61). Placebo tests for the permanent rural resi- dents and the rural–urban migrants are conducted separately, and the results strongly support a causal effect of the impacts. The remaining part of this paper is organized as follows. Section 2 documents the histor-

ical background and explains the identification methodology. Section 3 describes the theo- retical framework and empirical models. Data and statistics summary are presented in Section 4. Section 5 presents the primary results and examines the potential channels through which the Cultural Revolution matters. Robustness checks are also provided. Sec- tion 6 concludes the paper.

2. Historical background and identification methodology

The Cultural Revolution lasted for more than a decade and involved abundant conflicts (Li et al., 2013; Treiman and Deng, 1997; Zhou and Hou, 1999). It influenced the evolu- tion of social values, political institutions, and individual developments in China. Consider- ing the prime stage of human capital development, I focus on the urban cohorts of school age during this period who were exposed to multiple interruptions to their development (birth cohorts 1946–61). The first large-scale interruption was, as is well known, the educa- tion interruption. Different levels of schools were closed for certain periods, which dis- rupted students’ advancement through the formal education system. Separate from the school closures, the length and substance of education also changed. Second, at the same time, those cohorts were also forced to leave cities and live in rural areas (the Send-Down Movement). Third, there were other conflicts, for example, the Red Guard Campaign and public shaming. Overall, the first two were large-scale and more likely to have had lasting impacts. Details on these two policies and the identification methodology are documented in the following sections.

Historic Legacy Evaluation: the Cultural Revolution 287

2.1 The closure of schools

Before the Cultural Revolution, the formal education system had a six-three-three-four structure in China: 6 years of primary school, 3 years of junior secondary school, 3 years of senior secondary school, and 4 years of university education. Children began formal pri- mary schooling at the age of seven or eight (Li et al., 2013; Zhou and Hou, 1999). Primary and secondary schooling were the most important components and involved large portions of the population. Newly enrolled students in the secondary schools accounted for 17 per cent of all enrollments, and the proportion of primary students constituted 83 per cent in 1965. Meanwhile, the ratio of students enrolled in colleges to all newly enrolled students was far below 1 per cent. In all, the number of schools and students below college level accounted for the major component in the formal education system before the Cultural Revolution (Data source: China Education Statistics Year Books: 1972–2008). Correspond- ingly, the closure of secondary and primary schools would have affected the majority of the population. After the May 16th Notification in 1966, the Cultural Revolution was initiated, and

nearly all urban schools were closed or they stopped offering lessons after June 1966. Pri- mary schools were temporarily affected, and many junior secondary schools did not resume until the fall of 1968 (Source: People’s Daily, which is a central newspaper, October of 1967). Senior secondary schools resumed in September 1971, and they grew in number from 500 to 4000 schools by 1972 (Data source: China Education Statistics Year Books: 1972–2008). The numbers of new enrolled students and full-time teachers in 1972 more than doubled compared with the situations in 1971 for urban areas. A small number of universities, all of which had been closed since 1966, gradually began reenrolling students after 1972. The majority of colleges did not formally reopen until October 1977, the year in which the National College Entrance Examination was reinstituted (Singer, 1971; Treiman and Deng, 1997). Hence, secondary schools and above were shut down for relatively longer periods. Considering both the exposed population and the length of the school closures, the clo-

sure of the secondary schools should have had a greater impact. As is consistently illus- trated in Figure 1, the completion of senior secondary schooling was significantly affected. The increasing trends in senior and junior high attainment were obviously disrupted for cohorts of school age during this period (especially for birth cohorts 1947–59). Based on the regular age for attending secondary school and the timing of the secondary school recovery in urban areas, an index ‘Closure’ for measuring the intensity of being treated by the school closure over birth cohort is constructed as follows:2

Closureij ¼ 6 if individual i was born in the birth cohort j : j = 1952–54 4 if individual i was born in the birth cohort j : j ¼ 1950�51; 1955�56 1 if individual i was born in the birth cohort j : j ¼ 1948�49; 1957�58 0:5 if individual i was born in the birth cohort j : j ¼ 1947; 1959 0 otherwise

Unless otherwise specified, the subscript j always represents the birth year and i repre- sents individual i in this paper. The value of the index can be approximately interpreted as number of years of formal secondary education that were denied for cohort j. The birth cohorts of 1952–54 are identified as the potentially most affected group because their junior and senior secondary schooling were both disturbed. For example, if one was born

288 Dong Zhou

in 1953, he was interrupted by the close of junior secondary schools when he was at age of 13 and was about to enter junior secondary schools in 1966. Assumed that he could con- tinue his education after 1967, when he finished 3 years of junior secondary schooling around 1970, his advancement to senior secondary education was interrupted again. There- fore, individuals born in this birth cohort are classified in to the most treated group and their 6 years of secondary schooling potentially got denied due to the closure. The second group (birth cohorts 1950–51 and 1955–56) partially experienced education

interruptions at the secondary school level. For example, kids born in 1950, they were at age of 16 in 1966 and were attending junior secondary school. Their senior secondary edu- cation had been interrupted and marginally affected by the closure of junior secondary schools. If were born in 1955, they were attending primary schools and were about to attend junior secondary school in 1967 so that their primary schooling was marginally affected and junior high schooling were probably denied by the policy. The third group comprises cohorts born between 1957 and 1958 and those born between 1948 and 1949. These cohorts were marginally affected in either primary schooling or their senior sec- ondary, and both were delayed in potential college entrance. Grade repetition and early or delayed school entry could lead to bias in estimations if these cohorts are mistakenly con- sidered to be treated or non-treated. Therefore, the birth cohorts of 1947 and 1959 are also taken into consideration. The mean of this measure is 1.5; that is, approximately on aver- age, the treated cohorts were denied 1.5 years of secondary schooling (see column 2 of Table 1). In addition to the school closures, the length of schooling and the substance of education were also affected. For some regions, the length of each schooling level was shortened by half a year or one year. Half work and half study were required. Admission

Figure 1. Education attainment in 1990, by birth cohort and school level

Sources: Urban residents from China Census 1990. Consistent patterns are observed in our empirical sample of CHIPS 2002.

Historic Legacy Evaluation: the Cultural Revolution 289

policies also deviated from being academic merit based. For example, colleges selected stu- dents from specific social classes (workers, peasants, and soldiers) with no academic merit criterion. These changes caused a downgrade in academic standards and an emphasis on political qualifications (Treiman and Deng, 1997; Zhou and Hou, 1999; Meng and Gre- gory, 2003a). Additionally, because of the large geographic scope of China, the intensity of the

education interruptions varied across rural and urban areas. With the goal of reducing gaps between rural and urban areas, the experience in rural areas was quite different. First, there was no large-scale closure of schools in rural areas such as occurred in urban China. Meng and Gregory (2003a) provided empirical evidence that the exposed rural cohorts were not as significantly affected as their urban counterparts by the school closures. Second, Han et al. (2011), Thogersen et al. (2002) and Zhang (2012) even mentioned that there was an expansion of middle schools in some rural counties in the period of 1972–76. Therefore, my evaluation particularly focuses on the perma- nent urban residents, with the rural residents used for placebo tests.

2.2 The Send-Down movement

The major period of the Send-Down Movement was from 1968 to 1978, and it was implemented nationwide. At the early stage of the Cultural Revolution, the sent-down youth included a small group of adult social elites with higher education as well as young

Table 1. Methodology summary

Interested birth cohort

(1) Age in 1966

(2) (3)

(4) (5) (6) (7)Years of

secondary schooling denied

Probability of being sent down

Length of Stay in

Countryside (2002)

Send- Down (1995)

Send- Down (2002)

Exposure to Famine at Age 0 to 2(1995) (2002)

1946 20 0 0.16 0.14 6.32 0.98 0.96 0 1947 19 0.5 0.18 0.21 6.22 1.05 1.29 0 1948 18 1 0.24 0.24 5.04 1.16 1.26 0 1949 17 1 0.40 0.35 4.90 1.80 1.82 0 1950 16 4 0.45 0.46 4.83 2.17 2.32 0 1951 15 4 0.51 0.41 4.44 2.39 1.92 0 1952 14 6 0.44 0.34 4.60 1.87 1.60 0 1953 13 6 0.41 0.31 4.77 2.04 1.50 0 1954 12 6 0.33 0.36 3.96 1.31 1.48 0 1955 11 4 0.42 0.42 3.67 1.50 1.52 0 1956 10 4 0.41 0.49 3.49 1.42 1.70 0 1957 9 1 0.49 0.44 3.25 1.55 1.47 1 1958 8 1 0.41 0.42 2.75 1.09 1.20 2 1959 7 0.5 0.34 0.33 3.24 1.07 1.07 3 1960 6 0 0.27 0.17 2.28 0.60 0.40 2 1961 5 0 0.10 0.07 2.63 0.27 0.19 1 Mean – 1.5 0.214 0.195 4.09 0.85 0.79 –

Sources: 1995 and 2002 waves of CHIPS (empirical sample) and values of measure ‘Send-Down’ in 1995 are com- puted by the product of probability of being sent down in 1995 wave and length of stay in the rural areas merged from 2002 wave because of data limitation.

290 Dong Zhou

adults from families with particular social status (‘chengfen’). During the major period, the predominant population that was rusticated was current students or graduates from secondary schools. Specifically, the highly exposed cohorts were those who were born between 1946 and 1961 regardless of province (see Figure 2, Table 1 and Appendix A). In Appendix A, I examine the determinants of being sent down or not. Estimation results confirm that the probability of being sent was highly correlated with family background (for example, parents’ education attainment and social status). Moreover, details of the sending policy varied across regions and changed continuously over time. For some regions, the local governments required at least one child from each family to be sent to a rural location. Meanwhile, some cities established the criterion that only one child must be sent. An extreme case was Wu Han City in Hubei Province, where the government sent all age-eligible youth to the countryside in 1974 (Bernstein, 1997). The return policy also var- ied across regions and over time. Returns began sparsely in 1973 but were limited before 1978 (Treiman and Deng, 1997; Li et al., 2013; Zhou and Hou, 1999). After 1977, the returns of the sent-down youth became common, and most of them returned to their origi- nal locations.3

The duration of stay was not exogenously determined (Xie et al., 2008). Different poli- cies, families’ bargaining power and individual performances led to varying durations of stays in the rural areas among the sent-down youth. On average, the older birth cohorts were likely to be sent down earlier and required to stay longer in the rural areas than were the younger birth cohorts (see Appendix A). During their stays, some of the sent-down youth were sent to live with farmers, perform manual labor, and spend years in rural areas (Bernstein, 1997; Zhou and Hou, 1999). There were various exposures during their stays. First, they might have performed different economic activities because of their own capaci- ties or family backgrounds. For example, the sent-down intellectuals could also work as part-time teachers in rural schools in addition to performing manual work while they

0 .2

.4 .6

.8 P

ro b.

o f b

ei ng

s en

t- do

w n

0 .5

1 1.

5 2

2. 5

3 S

en d-

D ow

1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 Year of birth

Measure of send-down Prob. of being sent-down

Figure 2. The intensity of being treated by the Send-Down movement

Source: Urban residents of CHIPS 2002 and 1995.

Historic Legacy Evaluation: the Cultural Revolution 291

stayed in the countryside. Some might have been assigned to harsh manual work. Second, any number of geographic and demographic factors in rural areas also endowed abundant heterogeneities across the sent-down population. Different living environments might have had fewer resources or fewer social conflicts than urban areas. A small group of sent-down youth could even continue schooling in some rural areas (Xie et al., 2008). Additionally, because of different adaptive capacities among individuals, the cohorts might also have developed differently during their stays. Overall, this policy altered their life courses and affected their human capital accumulation, not only with regard to their receipt of formal schooling but also with regard to their accumulation of working experience, and it further shaped their future economic opportunities. On the whole, there were heterogeneities of treatment effects, selection problems, and

various exposures during their stays for the sent-down population. Meanwhile, since the non-sent-down youth were simultaneously exposed to other interruptions in urban areas, it is difficult to find the counterfactual outcomes for those who were sent down. Therefore, the simple comparisons between sent-down urban youth and those who were not sent down at individual level in the existing literature might have captured biased estimations of the Send-Down Movement. Even for the twin comparison at individual level, it has its lim- itations, which is not only related with insufficient observations lack of representativeness as well as unobserved heterogeneities even for twins but also associated with the selection problems within a family and across families as well as the heterogeneous lives in the rural areas for the sent-down youth. In addition, the positive marginal effect of staying one extra year among twins might in itself capture the subsequent evolvement of adaptability or per- formances of the more able twin rather than a true effect of the Send-Down experience on the whole exposed population. Correspondingly, I separately calculate the probability of being sent down and the average length of stay at the cohort level and construct their pro- duct, ‘Send-Down’. This measure gauges the expected length of staying in the countryside, reflecting the intensity of being treated by the Send-Down Movement over birth years. It alleviates selection bias and potential measurement errors at the individual level. Consider- ing the regional differences, I also consider regional variation: the product of the probabil- ity of being sent down and the average length of stay in the countryside for one specific birth cohort born within a specific province. Consistent conclusions were found.

Send � Downj ¼ PjLj

Pj: the probability of being sent down for birth cohort j. Lj: average length of staying in the rural area for birth cohort j. Figure 2 visually describes the measure based on the individual dataset from CHIPS

2002. Compared with other cohorts, the birth cohorts of 1946–61 were more likely to be rusticated and to have a larger Send-Down value, that is, a longer expected length of stay in the rural areas.

2.3 The measure of the cultural revolution

As shown in the above paragraphs, the primary birth cohorts who were exposed to the school closures were also those who were exposed to the Send-Down Movement. These are individuals who were of school age during the Cultural Revolution and were at the prime stage of human capital accumulation, and they were relatively vulnerable to their

292 Dong Zhou

surrounding environments. I further construct a single binary variable (CR) to capture the overall impact: CR = 1 for agents born between 1946 and 1961.4 The control group is the younger and older birth cohorts who potentially experienced no interruptions to their human capital accumulation during the Cultural Revolution. Given the identified affected birth cohorts (1946–61), the Great Famine (1959–61) might

be one confounding factor.5 Although it impacted the rural areas directly, the Great Fam- ine also marginally affected the urban areas. Without controlling for it, the resulting esti- mates could have been biased. Therefore, I constructed a discrete variable, Famine, to control for its effects based on the number of years the birth cohort was exposed to the Great Famine during infanthood (age 0 to 2; see column 7 of Table 1).

3. Theory and empirical model

It is common wisdom that the individual life-cycle earning profile commonly follows an inverse U-shape pattern, increasing during the working lifetime and declining later in the range of retirement age (Deaton and Paxson, 2000; Modigliani, 1986). Similar hump shapes of age-earning profiles are widely documented in empirical studies that examine cross-sectional data. The cross-sectional profile also works as a surrogate for an individual life-cycle profile in non-stationary economies with increasing productivity growth over birth cohorts (Irvine, 1981). Given a growing economy such as that of China, the Cultural Revo- lution severely interrupted the human capital accumulation for a range of birth cohorts. The permanent loss in human capital led to deviations from the bump-shape profile because of the interruption to the increasing dynamics of cohort productivity growth over birth year. Following Shorrocks (1975), it is assumed that individuals complete formal human capi-

tal formation at age of 25 and begin to work for pay. The earnings of the representative individual born in cohort j in year t in a cross-sectional profile are denoted as follows:

Iðt; jÞ ¼ ¼ 0 if 0 � t � j � 25¼ hf ðt � j � 25ÞekðjÞ if t � j � 25 �

where h represents the resources that are common to all generations; f(�) is a concave func- tion of age that can also be represented by a polynomial function of age; and k(j) repre- sents the cohort productivity growth for cohort j. The cohort effect in the regressions for the log income increases almost linearly with the year of birth over the generation (Jappelli, 1999). One prime determinant is the increasing trend of human capital produc- tion over time (e.g. educational attainment) identified in the rich empirical studies (Jappelli, 1999; Mincer, 1997). Given the generation-specific productivity growth and no distur- bances, the cross-section age-earning profile can be reduced to a pure age effect when exhorting all of the generation-specific resources (King and Dicks-Mireaux, 1982). From this aspect, cohort productivity growth reveals the pattern of the generation-specific resources that contribute to human capital accumulation. In addition to productivity growth, the cohort effect can also reflect heterogeneities of preferences and mortality rates across birth cohorts (Jappelli, 1999; Masson, 1996; Shorrocks, 1975). Based on the above theoretical framework, with no interruptions or uncertainties, the

cohort productivity growth should increase over the year of birth, and the life-cycle earning

Historic Legacy Evaluation: the Cultural Revolution 293

profile should exhibit the widely observed inverted U-shape. If particular generations expe- rienced adverse interruptions during human capital production, the positive correlation of cohort effects with year of birth would be interrupted, giving rise to deviations from the life-cycle model. In other words, we can introduce cohort-level productivity disturbances into the model when large-scale interruptions destroy generation-specific resources and per- manently cause the loss of human capital within a particular range of cohorts. Therefore, I can identify the impact of this plausible natural experiment, the Cultural Revolution, by testing the predictions of the life-cycle model using cross-sectional micro-data. Considering the far-reaching influences and complexity of the Cultural Revolution, my

baseline estimating model to evaluate the existence of the impact is

Yij ¼ a þ bPolicyij þ d1agei þ d2age2i þ d3Genderi þ d4Provincei þ d5Famineij þ eij ð1Þ

where the dependent variable Y is the natural log of annual income6 for individual i of cohort j. Policy represents the Cultural Revolution, measured as three alternatives (CR, Closure and Send-Down). These capture different intensities of exposure to policies across cohorts (see Section 2). Other independent variables include age, age squared, gender, and provincial indicators.7 In addition, the Great Famine is also taken into consideration for an accurate estimation concerning mortality influence and potential contamination. The coefficients of the three measures are of interest, and consistent estimates will indicate the qualitative and quantitative effects of the Cultural Revolution. Moreover, with repeated cross-sectional datasets (CHIPS 1995 and 2002), I can track

the same cohorts over time and examine the differences between the affected and the control cohorts. In the theory of human capital production (e.g. schooling and working skills), various exposure to adverse childhood environments could have also resulted in differences in the slopes of the earnings life-paths among different population. Hence, the second proposition is to examine how the impact changed over time as the Chinese market economy increasingly evolved. In a more market-driven economy, workers should earn closer to their marginal product to labor so that the negative impacts should increase. Third, I will further study the possible channels and identify how the Cultural Revolu-

tion created differences in generations using extended models of equation [1] using sequen- tial addition.8 Following the basic technique to uncover the mechanism (Nunn, 2008), I control for individuals’ characteristics in the regressions one by one and analyze how the coefficients of interest change. Since the complexity of the Cultural Revolution, it corre- lated with most aspects of the life-course for the exposed population. The pattern of changes interprets the mechanism and shows how the Cultural Revolution affected the population through particular micro-perspectives in long term. In other words, the baseline regressions are extended by controlling for different channel variables: educational attain- ment, current marital status, work experience, schooling years, health status, attitudes, etc., one by one. After controlling for one specific channel, the interested coefficients should increase

with smaller absolute values if the Cultural Revolution negatively impacted through that channel or vice versa. For example, cohorts affected by the Cultural Revolution had lower education attainment and were thereby likely to earn less. After controlling for this channel, the absolute value of the interested estimates will shrink. Based on pat- terns of changes, dominant channels can be identified. The interested estimates would

294 Dong Zhou

become insignificant if the complete mechanism could be captured. Furthermore, I will examine subsamples of different control groups, control for family background and uti- lize datasets of rural residents as well as rural–urban migrants to construct placebo tests for robustness checks.

4. Data and statistics descriptions

4.1 Data

This paper draws mainly on data from the CHIPS 2002.9 It comprises multiple question- naires and provides rich information on permanent rural residents, permanent urban resi- dents, and rural–urban immigrants at the household as well as individual levels. The urban total from the 2002 wave randomly selected 20,632 permanent urban individuals and 6,835 urban households from 12 provinces. Most importantly, it includes Send-Down experience at the individual level, such as whether one was rusticated and how long one had stayed in rural areas, which makes it possible to measure this treatment. In addition to its credibility and capacity, another advantage of this dataset is that the identified cohorts were of work- ing age, whereas the most recent datasets might have fallen short of the requirements for the interested cohorts who were out of the labor force. The urban data sheet of CHIPS 2002 also contains a wide range of demographic and

economic variables. First, the data provide detailed information about annual personal income10 in 2002 and retrospectively for 2000 and 2001. I take the average of the three years’ incomes as the dependent variable. To check robustness, regular wage, and current annual income are also taken as dependent variables and consistent findings are found. Information about current employment status, current marital status, occupation and years of schooling is provided, and I utilize it to explore channels. Guided by this theoretical framework, I narrow the empirical sample to permanent urban residents age 25–60. Indi- viduals who had not completed their human capital accumulation (full-time students) or who were full-time homemakers in 2002 are excluded. To answer the question of how the impact changed over time, I also check CHIPS 1995. The same strategy for the sample is applied to permanent urban residents taking out full-time student, full-time homemakers, and the disable group. In addition, urban residents of Chinese Family Panel Studies (CFPS) 2008 and CHIPS 2007 are utilized to explore the channels in the transmission mechanism of the Cultural Revolution. For CHIPS 2007, I compute the average height at birth cohort level by gender to merge into CHIPS 2002 for the channel of health status. Similarly, I compute the ratio of values toward hardwork to luck using information of per- sonal attitudes toward different determinants in personal success in CFPS 2008, and take cohort averages to merge into the main data for studying the channel of attitude.

4.2 Statistics description

As presented in Table 2, for the 2002 wave, the full empirical sample is 12,304 observa- tions. Approximately 50 per cent of the samples are female, and 19.5 per cent report being sent down. For the sent-down youth, the average length of their stay in rural areas is approximately 4.1 years. The natural log of average annual income from 2000 to 2002 is 8.966 on average, and in 2002, the average years of schooling are approximately 11. From

Historic Legacy Evaluation: the Cultural Revolution 295

the middle 1990s to the early 2000s, the natural log of average annual income of urban res- idents increased dramatically by 7.4 per cent for cohorts of 1942–70 as the market-oriented Chinese economy rapidly grew. Based on mean values of income, the older cohorts earned

Table 2. Summary statistics

Urban residents CHIPS 2002

All (obs. 12,304)

Birth cohorts 1942–45

Birth cohorts 1946–61

Birth cohorts 1962–77

Log (income) 8.966 (0.95) 9.063 (0.59) 9.006 (0.74) 8.900 (1.06) Age 43.12 (8.72) 58.44 (1.41) 48.099 (4.23) 34.235 (4.433) Female 0.495 (0.50) 0.429 (0.50) 0.486 (0.50) 0.517 (0.50) Schooling years 10.944 (3.1) 10.300 (3.52) 10.318 (2.95) 11.885 (2.99) Sent-down youth 0.195 (0.40) 0.055 (0.23) 0.347 (0.48) 0.009 (0.097) Current marital status (composition) Other (per cent) 0.04 0.00 0.01 0.08 Single (per cent) 4.71 0.43 0.39 11.14 Married with spouse (per cent) 93.18 96.14 97.27 87.24 Separationa (per cent) 2.07 3.43 2.32 1.54 Marital instability of sent-down youth

(per cent) 5.13 2.89 2.13

Current employment status (composition) Other (per cent) 0.24 0.57 0.28 0.12 Employed (per cent) 78.02 30.14 72.99 91.61 Unemployed (per cent) 2.80 0.71 2.50 3.48 Retired (per cent) 15.25 66.00 20.43 1.05 Laid-off (per cent) 3.7 2.57 3.79 3.73

Work experienceb (2002) 21.025 (8.95) 33.898 (9.28) 26.686 (6.58) 14.301 (5.83) Working experience of sent-down youth 35.9 (4.12) 28.00 (4.97) 21.381 (4.91) CFPS 2008 attitudec (obs. 3855) 1.377 (0.979) 1.341 (0.94) 1.350 (0.93) 1.414 (1.04) Diligence 5.968 (1.21) 6.09 (1.12) 5.96 (1.22) 5.96 (1.21) Lucky 5.135 (1.59) 5.300 (1.53) 5.170 (1.58) 5.070 (1.61) CHIPS 2007 height (obs. 8865) 165.215 (7.677) 164.214 (7.48) 164.708 (7.893) 165.800 (7.457)

CHIPS 1995 (urban areas) All (obs. 1581)

Birth cohorts 1935–45

Birth cohorts 1946–61

Birth cohorts 1962–70

Work Experience 21.839 (9.096) 32.123 (7.076) 22.239 (5.568) 10.805 (4.125) Proportion of smokers (per cent) 30.6 30.9 30.9 29.8 Smoking proportion of sent-down youth (per cent) 43.3 32 33.3

CHIPS 2002 and 1995 All mean

Older group (1942–45)

Birth cohorts 1946–61

Younger group (1962–70)

Log (income) 1993–95 8.400 (0.867) 8.518 (0.650) 8.480 (0.716) 8.160 (1.175) Log (income) 2000–02 9.023 (0.741) 9.063 (0.765) 9.006 (0.739) 8.995 (0.770) Growth rate (per cent) 7.42 6.4 6.2 10.2

Notes: aThe group of separation includes observations who is currently divorced or separated from her or his spouse for other reasons; bstatistics summary for working experience only works for employed population of the wave of 2002. cVariable of Attitude represents attitude towards determinants of success (the impor- tance of diligence to luck in determining personal success). The importance of different factors is measured from 1 to 7. The higher the value is, the more important the individual values the determinant. (1) Summary statistics is computed based on CHIPS 2002 and 1995. Means of variables are shown with the standard deviation within the parentheses. (2) Income growth rate is computed after the 2000–02 average annual income is deflated by national con- sumer price index.

296 Dong Zhou

more than the younger cohorts but the growth rate of treated cohorts was relatively lower than the younger and older control groups. Examining the educational attainment across groups, the treated cohorts have obtained 1.5 fewer years of schooling than the younger control group which is similar to the mean of the measure Closure, while they only have 0.018 years more than the older cohorts. The difference in the differences implies a sig- nificant loss in education attainment of the exposed cohorts. Generally, patterns of these variables follow the life-cycle profile: the older the cohorts, the more working experience they possess, relatively less schooling years they have obtained, and the more of them are retired. Deviations from the trend to some extent reflect specialties of the exposed cohorts. This survey has skipped the questions of working experience and occupation for the

unemployed population. Hence I examine current employment status for the full sample as one channel and further study the working experience channel for the subsample of the employed population. There are multiple classes of employment status in the 2002 wave of CHIPS: currently employed, unemployed, retired, special status, and other.11 As the data show, there are more laid-off observations of the treated group than the control groups. The treated cohorts work for more years than the average of the control groups no matter in 1995 or 2002. In particular, the sent-down youth have more working experience no mat- ter within which group they are in (e.g. on average 28 years of working experience which is 2 years more than the non-sent youth within the treated cohorts 1946–61). Contemporane- ous comparisons between the sent-down and non-sent-down youth show that marriage status of the sent-down youth are more instable, and the treated group as well as the sent- down population are likely to smoke. Average height of treated cohorts are slightly shorter than average of the two control groups. Additionally, the variable representing attitudes toward diligence against personal luck in determining personal success is computed and the relatively higher value of the treated cohorts implies they value diligence relatively more than luck. In the next section, estimation results will be presented. To visually illustrate the Cultural

Revolution’s effects on shaping life-cycle profile, I first graph the coefficients of age indica- tors from two basic regressions using CHIPS 2002: regressions of log income and school- ing years on gender, age indicators, and region indicators (see Figure 3). In Figure 3, the two vertical lines highlight the interested cohorts. The right and left axes represent the coefficients from the two separate regressions. As it shows, there was an overall increase in educational attainment trend overtime, but there is a dip in the range of the affected cohorts. The pattern of cohort effects in the regression of log income exhibits a quasi- bumped pattern over the age profile but with a similar dip for the interested group. These violations against the positive relationship between cohort productivity growth and year of birth as well as the inverted U-shape of age-earning profile visually imply the existence of impacts of the Cultural Revolution.

5 Estimation results

5.1 The impact on income

The results of the empirical model established in equation [1] are presented to show the existence of the long-term impact in Table 3. Different estimations of interest are sepa- rately obtained utilizing different measures for the Cultural Revolution. Note that the interested estimates, regardless of which Cultural Revolution measures or which

Historic Legacy Evaluation: the Cultural Revolution 297

cross-sectional life-cycle profiles were adopted, are significantly negative. Our main focus is the impacts in the early 2000s. The estimate of CR indicates that the affected cohorts experienced lower average annual incomes by roughly 11 per cent [see regression (1) in Table 3]. The estimates for Closures and Send-Down were also negatively significant [re- gressions (2) and (3) in Table 3]. Empirically, one year of secondary schooling denied decreased annual income among the affected by 1.8 per cent in the early 2000s. As for the role of the school closures, frequent interruptions in formal education on average lowered the earnings of the affected cohorts by 2.7 per cent among the empirical sample (�1.8 per cent 9 1.5; 1.5 is the mean of Closure). The magnitudes of the interested estimates imply a stronger and statistically negative influence of the Send-Down Movement. On average, the Send-Down Movement lowered the incomes of the affected population by 4.8 per cent in the early 2000s (�6 per cent 9 0.195 9 4.09; �6 per cent is the coefficient of the estimate; 0.195 is the proportion of the full sample that was sent down; 4.09 is the average length of stay) among the full empirical sample. Compared with the school closures, the effect of the forced immigration was nearly double. However, one may argue that the economic environment might have varied and the

interested estimates may mainly capture the effects of changes in macroeconomic environ- ments rather than the role of the Cultural Revolution in shaping the generation specifics. CHIPS 199512 is also explored to investigate this concern as well as to examine how the costs of the Cultural Revolution evolved in response to macroeconomic environment changes. As shown in Table 3, I examine the age-earning profiles at two points in time. The Cultural Revolution significantly lowered the subjected cohorts’ average annual incomes by 9.1 per cent in the mid-1990s. Comparing the estimates of CR, the negative effects are amplified by 20 per cent from the mid-1990s (�9.1 per cent) to the early 2000s

Figure 3. Deviations from the life-cycle model

Notes: The scatters are the coefficients of the cohort indicators in the regression of natural log of income as well as schooling years on cohort indicators, gender and region indicators based on the empirical sample. And the coefficients are smoothened by taking three continuous cohorts’ moving average for graphing. Two vertical lines highlight the interested birth cohorts.

298 Dong Zhou

(�10.9 per cent). I also track the same birth cohorts 1942–70 and increased impacts are consistently found over time (see Appendix C: Table 1C). This amplification could be asso- ciated with the fact that as the Chinese market economy increasingly evolved, the factor returns were more determined by their marginal productivity in the market. Historically, China began to transition to a market economy immediately after the Cultural Revolution (1979), and this market-oriented economic reform gradually deepened and induced rapid economic growth in China after the 1980s. The affected cohorts were at a disadvantage as the result of various interruptions in human capital accumulation during the Cultural Revolution. The loss in human capital was further amplified as the market became more and more competitive. Furthermore, to study the simultaneous policies in terms of multiple controls, two or

three measures are simultaneously controlled for. Estimate results shows that at least one measure remained significantly negative, implying that the negative effect strongly persisted as the result not only of the policies measured in this paper (school the closures and the Send-Down Movement) but also of other characteristics of the Cultural Revolution. The

Table 3. The existence of long-term impacts

Sample: 25–60 age cohorts

CHIPS 2002 CHIPS 1995

(1) (2) (3) (1) (2) (3)

CR �0.109*** �0.091*** (0.020) (0.027)

Closure �0.018*** �0.006 (0.003) (0.004)

[�0.027] Send-Down �0.060*** �0.068***

(0.014) (0.02) [�0.048]

Female �0.351*** �0.351*** �0.351*** �0.272*** �0.272*** �0.272*** (0.015) (0.015) (0.015) (0.024) (0.024) (0.024)

Famine 0.001 �0.038** �0.021* 0.018 �0.0002 �0.0004 (0.001) (0.015) (0.013) (0.01) (0.0002) (0.008)

Age 0.0711*** 0.070*** 0.0654*** 0.143*** 0.119*** 0.152***

(0.008) (0.008) (0.007) (0.025) (0.020) (0.023) Age square �0.001*** �0.001*** �0.001*** �0.0016*** �0.0013*** �0.0017***

(0.00008) (0.00009) (0.00008) (0.0003) (0.0002) (0.0003) Constant 8.077 8.126*** 8.195*** 5.833 6.294 5.664

(0.162) (0.171) (0.157) (0.510) (0.423) (0.482) R2 0.101 0.101 0.101 0.12 0.12 0.12 Obs. 12,304 12,304 12,304 12,581 12,581 12,581

Notes: The bold values indicate the average treatment effects computed based on the empirical sample. (1) The three measures of the Cultural Revolution (CR, Closure and Send-Down) are examined separately. The alternative Send-Down which considers regional variation is also estimated with data of 2002 wave and results are consistent. For the 1995 wave, the average duration of stay in the rural areas for the sent-down youth is merged from the 2002 wave to construct the measure of Send-Down because of data limitation. It is same to all the coming tables. (2) Other control variables which have not been shown in the tables are the group of province indicators. Dependent variable are natural log of individual’s average total annual income between 2000 and 2002 or from 1993 to 1995. (3) ***, **, and * represent significance at 1, 5, and 10 per cent, respectively. The robustness standard error is reported in the parentheses and adjusted for clusters in age.

Historic Legacy Evaluation: the Cultural Revolution 299

above conclusions are consistent when the probability of being sent down in 2002 is instru- mented with that of the 1995 wave, when other constructions of measures are applied and when different age-spans are utilized (e.g. observations at ages 16–60). Additionally, when regular wage and current annual income are taken as dependent variables, consistent con- clusions are found. In all, the negative effects are consistent with the fact that the Cultural Revolution inter-

rupted human capital production and thereby led to lower earnings in the current urban China. All of the evidence demonstrates the existence of a long-lasting negative effect of the Cultural Revolution on the permanent incomes of the exposed population, and this negative impact was amplified as the Chinese market economy increasingly evolved. In addition, esti- mates show a stronger role of the Send-Down Movement in lowering the affected cohorts’ incomes than the role played by the school closures. This stronger role is also demonstrated by different specifications with multiple controls and principal factor analysis.

5.2 The transmission mechanism

To explore the channels that drove this negative impact, I proceed to estimate extended empirical models of equation [1] in the spirit of Nunn (2008).13 Specifically, I introduce productivity determinants (e.g. education attainment, work experience, and health), marital status and revealed cohort traits one by one to investigate how the treated population developed differently. The main estimation results for the full sample are presented in panel A of Table 4 from regressions (1) to (6). Three measures (CR, Closure and Send-Down) are controlled for respectively. The primary factor of human capital, schooling years, is controlled for in regression

(2), and regression (3) contains current employment status. Further, current marital sta- tus is included in regression (4). In regression (5), a proxy for health, the average height at cohort level, is controlled for and provides a quantified estimate of health condition. Cohorts’ attitudes toward the role of luck and hardwork in determining indi- vidual success are controlled for in regression (6) to explore the channel of cohort-spe- cific attitudes. Testing is conducted to control for the channels by different orders into the model, and consistent results are found. Panel B presents the estimation results using the subsample of the employed population. Rather than current employment sta- tus, working experience and occupation are available and are examined for this subsam- ple. In the following paragraphs, I will discuss specific channels through which the Cultural Revolution matters in detail.

5.2.1 Educational attainment14. The overall effect of the Cultural Revolution, regardless of the measures adopted, reduced by more than 50 per cent after controlling this channel (see panel A; from �10.9 to �5.4 per cent for CR; from �1.8 to �1 per cent for Closure; from �6 to �2.3 per cent for Send-Down). Rather than schooling years, I also use education attainment level to capture this channel, and consistent results are found. Note that the Closure estimates become insignificant or trivial as long as educational attainment is controlled for. In the labor market, education attainment is the primary signal for ability, skill, and

knowledge and plays a critical role in determining earning. In the scenario of the Cultural Revolution, what had happened are all levels of schools were closed, there were campaigns against intellectuals as well as social elites, and the rustication movement disrupted regular attendance in formal schools. Thereby, the subjected individuals acquired fewer years of schooling and were placed at a disadvantage in the labor market. This loss was established

300 Dong Zhou

T ab le

4. E xa m in e th e m ec h an

is m

o f th e cu lt u ra l re vo

lu ti o n (C

H IP S 20 02 )

P an

el A : re gr es si on

F ul l sa m pl e an

al ys is

(1 )

(2 )

(3 )

(4 )

(5 )

(6 )

C R

�0 .1 09

* * *

�0 .0 54

* *

�0 .0 44

7* *

�0 .0 43

5* *

�0 .0 43

5* *

�0 .0 44

4* *

(0 .0 2)

A lt er n at iv e m ea su re s

C lo su re

�0 .0 18

* * *

�0 .0 07

* �0

.0 03

�0 .0 03

�0 .0 02

(0 .0 03

) (0 .0 03

) (0 .0 04

) S en d -D

o w n

�0 .0 6*

* *

�0 .0 23

* *

�0 .0 2*

* �0

.0 19

* �0

.0 19

* �0

.0 23

* *

(0 .0 1)

S ch o o li n g ye ar s

Y es

E m p lo ym

en t- st at u s

Y es

M ar it al

st at u s

Y es

H ea lt h

Y es

A tt it u d e

Y es

R 2

0. 1

0. 17

0. 35

O b s.

12 ,3 04

P an

el B : re gr es si on

S ub

sa m pl e of

em pl oy ed

po pu

la ti on

an al ys is

(1 )

(2 )

(3 )

(4 )

(5 )

(6 )

(7 )

C R

�0 .1 22

* * *

�0 .0 72

* * *

�0 .0 68

* * *

�0 .0 51

* * *

�0 .0 38

* *

�0 .0 36

* �0

.0 35

* *

(0 .0 25

) (0 .0 26

) (0 .0 23

) (0 .0 19

) (0 .0 16

) (0 .0 14

) A lt er n at iv e m ea su re s

C lo su re

�0 .0 17

* * *

�0 .0 09

�0 .0 08

�0 .0 06

�0 .0 05

�0 .0 04

�0 .0 00

1 (0 .0 05

) (0 .0 05

) (0 .0 04

) (0 .0 03

) S en d -D

o w n

�0 .0 71

* * *

�0 .0 44

* * *

�0 .0 42

* * *

�0 .0 34

* * *

�0 .0 26

* *

�0 .0 23

* �0

.0 29

* *

(0 .0 14

) (0 .0 14

) (0 .0 13

) (0 .0 12

) (0 .0 1)

(0 .0 09

) S ch o o li n g ye ar s

Y es

W o rk in g ex p er ie n ce

Y es

O cc u p at io n

Y es

M ar it al

st at u s

Y es

H ea lt h

Y es

Historic Legacy Evaluation: the Cultural Revolution 301

T ab le

4. C o n ti n u ed

P an

el B : re gr es si on

S ub

sa m pl e of

em pl oy ed

po pu

la ti on

an al ys is

(1 )

(2 )

(3 )

(4 )

(5 )

(6 )

(7 )

A tt it u d e

Y es

R 2

0. 13

0. 18

0. 25

0. 29

O b s.

9, 63

7 9, 63

N ot es : (1 ) M ea su re s o f C R , S en d -D

o w n , an

d C lo su re

ar e co n tr o ll ed , re sp ec ti ve ly , in to

em p ir ic al

m o d el s;

M ar it al

S ta tu s is

a gr o u p in g va ri ab le

an d re p re se n ts

d if fe r-

en t m ar it al

st at u s in

20 02 . H ea lt h p ro xy

is av er ag e h ei gh

t at

co h o rt

le ve l m er gi n g fr o m

C H IP S 20

07 . A tt it u d e re ve al s h o w

re la ti ve ly

im p o rt an

t in d iv id u al s

va lu e d il ig en ce

ag ai n st

lu ck

in d et er m in in g p er so n al

su cc es s at

co h o rt

le ve l w h ic h is m er ge d fr o m

C F P S 20

08 .

(2 ) W o rk in g- E xp

er ie n ce

is h o w

m an

y ye ar

in d iv id u al s h av e b ee n w o rk in g an

d m is si n g in d ic at o r is

co n tr o ll ed

fo r 67

o b se rv at io n s w h o h av e n o t re p o rt ed

th e

in fo rm

at io n . E m p lo ym

en t- st at u s is a st at u s va ri ab le

gr o u p in g b ei n g u n em

p lo ye d , em

p lo ye d , la id -o ff , re ti re d , an

d o th er s in

20 02

. (3 ) O th er

in d ep en d en t va ri ab le s: ge n d er , ag e, ag e sq u ar ed , fa m in e, an

d p ro vi n ce -f ix ed

ef fe ct s.

(4 ) * * * , * * , an

d * re p re se n t si gn

if ic an

ce at

1, 5,

an d 10

p er

ce n t, re sp ec ti ve ly . T h e ro b u st

st an

d ar d er ro r is

re p o rt ed

in th e p ar en th es is

an d ad

ju st ed

fo r 36

cl u st er s in

ag e.

302 Dong Zhou

as an important and permanent channel through which the Cultural Revolution affected the subjected cohorts’ current average incomes.

5.2.2 Working experience. Limited by the survey, current employment status is examined for the full sample, and years of employment and occupation are tested for the employed subsample. The estimate of CR changes from �5.4 to �4.5 per cent when I control for current employment status, indicating that the Cultural Revolution affected the interested cohorts negatively through the current employment status channel. Individuals among the treated group were more likely to be retired, laid-off, or unemployed. Estimations support the negative impact on employment status. For the subsample of the employed population, the coefficients of CR and Send-Down also change to be negatively smaller when I control for years of employment in regression (3) of panel B. From regression (3) to regression (4), the estimates continue to shrink when controlling for occupation type. The changes imply that the Cultural Revolution played a negative role for the affected population through the working experience channel. Interestingly, statistics summary and estimation results show the treated cohorts and

sent-down population possess more years of working. The relatively more working years are probably associated with the school closure, school calendar change, and the forced immigration to rural areas leading to early entry into the labor market. But early entry contributed to more working experience but might have led to a lower paid path caused by insufficient education attainment. Moreover, the sent-down youth spent years in rural areas, and their accumulated working skills might have been unfit for the demands of the later urban market. Additionally, living in rural environments and doing manual work might have affected their working skills through increasing health hazards. If the data allowed for quality adjustment for the working experience variable, the sent-down youth’s real work experience would be lower, and we could obtain a more accurate understanding of the working skills channel.

5.2.3 Marriage. Current marital status is controlled for to quantify the effect of this channel. However, it is commonly argued that a better-matched marriage or a stable relationship can enhance efficiency, marital instability (for example, separation and divorce) will contribute to more uncertainties and depressed economic performance. Although current marital status is different from marriage history, it can still give us an idea of this channel. According to our empirical results, individuals’ current marital stability was negatively influenced by the interruptions and it accounted for a part of the overall negative effect of the Cultural Revolution on current earnings. There are a number of mediators for the marital instability caused by the Cultural Revolution. First, disturbed environmental factors psychologically influenced the interested cohorts’ commitments to relationships. Second, interruptions and uncertainties increased their searching costs in the urban marriage market, especially for those sent-down youth who returned to cities after years. One direct consequence is that overall, the interested cohorts’ first marriages were on average 2 years later than the before and after cohorts. Last but not least, separations caused by the Send-Down Movement and policies that prohibited the sent-down youth who married local residents from returning led directly to divorces.

5.2.4 Health. The average height at cohort level by gender is controlled for as a health proxy to provide a quantified estimate for the health condition channel [see regression (5) in panel A and regression (6) in panel B of Table 4]. I also compute the mean body mass

Historic Legacy Evaluation: the Cultural Revolution 303

index at cohort level by gender to control for as the health status channel. The estimates, regardless of which Cultural Revolution measures are used, do not change significantly (see panel A). However, for the employed subsample, the estimates of CR and Send-Down imply a significantly negative impact on the affected cohorts’ health conditions (CR: from �3.8 to �3.6 per cent; Send-Down: from �2.6 to �2.3 per cent). Considering the unclear picture of the synthetic cohort analysis, I turn to other pieces of evidence to illustrate the potential effects of the Cultural Revolution through this channel: violence experienced during adolescence could have affected the subjected cohorts’ physical development, and bad habits could have been cultivated by depression during this period. For example, the highly exposed cohorts had a higher proportion of smokers than other cohorts (Table 2). Third, when the rusticated youth were sent to rural areas, they were forced to live in rural environments and perform manual labor (Li, 2013 and Zhou and Hou, 1999). Inferior environments might have hindered their human capital accumulation, for example, lack of nutrition and harsh manual labor.

5.2.5 Revealed attitudes. The last regression controls for the cohorts’ attitudes toward diligence versus personal luck in determining success. For the full empirical sample, the estimate of CR does not change significantly. Interestingly, the coefficient of Send-Down changes from �1.9 to �2.3 per cent and becomes more significant. For the employed population, the estimate of Send-Down also becomes more significantly negative after controlling for this channel (changes from �2.3 to �2.9 per cent). The patterns imply that the Send-Down Movement played a positive role in the exposed observations’ current performance through this channel which is confirmed by results in Table 1B in Appendix B. Life in rural areas made the sent-down youth more perseverant, and they utilized their diligence to conquer personal misfortune. In summary, there are multiple conclusions based on the above results. First, the esti-

mates of interest consistently trend toward zero from the first to the last regression as additional potential channels are added, regardless of the measures (CR, Closure and Send-Down) or samples. Numerically, the impacts are larger within the employed popula- tion on average (CR, 10.9 per cent for the full sample and 12.2 per cent for the employed subsample). Second, the changes in the estimates’ magnitudes indicate that educational attainment is the most important channel, accounting for roughly 50 per cent of the over- all effect. Third, channels are unexhausted because the interested estimates (CR and Send- Down) were still significantly negative in the last regression for both panels.

5.3 Robustness checks

The estimation results are consistent with our hypothesis: past adverse interruptions dur- ing the human capital accumulation process will cause violations of life-cycle model predic- tions. They interrupted the positive increase in cohort-specific productivity growth and negatively affected the subjected groups’ current economic performance. In this section, the robustness of the above results is checked. I gauge the effects among different subsamples with different control groups to check

the existence of the negative impact in panel A of Table 5. First, I study a shorter age span with control groups of birth cohorts four years earlier and five years later than the treated cohorts (total empirical sample: birth cohorts 1942–67) for both waves, which can be con- sidered contemporaneous comparisons. We consistently find that the subjected group earned significantly less on average and the Send-Down Movement played a stronger

304 Dong Zhou

negative role than the school closures. Estimates of both measures negatively increased from the 1990s to the 2000s. Second, subsamples only including the control group born after 1961 are considered. Estimates for these samples (birth cohorts of 1946–70) are also significantly negative. For the 2002 wave, the overall effect is �11 per cent, whereas for the 1995 wave, it is �6.6 per cent. In all, persistent negative impacts of the Cultural Revolution are found regardless of the control group chosen and regardless of the time point being examined. Moreover, in comparing the impacts in 1995 with those in 2002, the effects are amplified over time. I further examine the subsample with family background considering potential selection

bias, and the results are reported in panel B of Table 5. The survey provides family back- ground information for household heads and their spouse, for example, their parents’ social status (‘chengfen’), education attainment, and occupations. All of these predetermined fam- ily background variables are controlled for in regressions simultaneously and respectively. Consistently, the significantly negative estimates of interest support the existence of long- term impacts (�12.5 per cent for CR; �1.3 per cent for Closure; and �7.8 per cent for

Table 5. Robustness of the existence of the impact on incomes

Panel A

Different control groups

2002 CHIPS 1995 CHIPS

Birth cohorts 1942–67 1946–70 1942–70 1942–67 1946–70 1942–70

CR �0.070*** �0.111** �0.103*** �0.05** �0.066** �0.061** (0.024) (0.04) (0.03) (0.02) (0.03) (0.023)

Closure �0.01** �0.01 �0.01 �0.003 �0.008** �0.007 (0.004) (0.005) (0.005) (0.003) (0.004) (0.004)

Send-Down �0.04** �0.02 �0.04** �0.025* �0.03** �0.033*** (0.01) (0.02) (0.02) (0.01) (0.01) (0.01)

R2 0.13 0.15 0.15 0.2 0.24 0.24 Obs. 9,878 10,198 10,896 10,081 9,748 10,801

Panel B

Controlling family background (CHIPS 2002)

(1) No (2) Yes (3) No (4) Yes (5) No (6) Yes

CR �0.141** �0.125** (0.04) (0.04)

Closure �0.017* �0.013* (0.007) (0.007)

Send-Down �0.83*** �0.78*** (0.025) (0.025)

R2 0.15 0.19 0.15 0.19 0.15 0.26 Obs. 3,788 3,788 3,788 3,788 3,788 3,788

Notes: (1) Panel A tests subsamples of different age-spans from both waves of CHIPS. Panel B contains urban residents at age of 25–60 with respective parents’ information in 2002. This subsample with family back- ground is constructed by combining the household information with the individuals’ data. Family back- ground includes parents’ occupation, parents’ social status (‘chengfen’), and parents’ educational attainments. (2) Other independent variables are gender, age, age squared, effects of famine, and provincial-fixed effect. (3) ***, **, and * represent significance at 1, 5, and 10 per cent, respectively. The robust standard errors are reported in the parenthesis and adjusted for clusters in age.

Historic Legacy Evaluation: the Cultural Revolution 305

T ab le

6. P la ce b o te st s w it h in

C h in a

D at a so ur ce

C H IP

S 20

02 C H IP

S 19

95 C H IP

S 20

M ea su re s

R ur al

re si de nt s 20

02 R ur al

re si de nt s 19

95 R ur al – ur ba

n m ig ra nt s 20

C R

�0 .0 15

(0 .0 50

) �0

.0 34

(0 .0 75

) 0. 20

1 (0 .1 63

) C lo su re

�0 .0 08

(0 .0 07

) 0. 00

4 (0 .0 13

) 0. 00

6 (0 .0 35

) S en d -D

o w n

�0 .0 25

(0 .0 30

) 0. 04

3 (0 .0 51

) 0. 04

9 (0 .1 08

) R 2

0. 27

0. 36

0. 16

O b s.

7, 36

9 7, 36

9 3, 81

9 3, 40

7 3, 40

N ot es : (1 ) E m p ir ic al

sa m p le

o f p er m an

en t ru ra l re si d en ts

in cl u d es

o b se rv at io n s th at

cu rr en tl y w o rk

an d at

ag e o f 25 – 60

. T h e su m

o f th e an

n u al

w ag e in co m es

an d

n o n -w

ag e in d iv id u al

in co m es

(n o n -a gr ic u lt u re

ea rn in gs ) ar e co m p u te d as

th e d ep en d en t va ri ab le .

(2 ) E m p ir ic al

sa m p le

o f ru ra l– u rb an

m ig ra n ts

in cl u d es

m ig ra n ts

th at

ar e at

ag e o f 25 – 60

an d m ig ra te d to

u rb an

af te r 19

90 s. N at u ra l lo g o f th e to ta l an

nu al

in co m es

ar e ta ke n as

th e d ep en d en t va ri ab le .

(3 ) O th er

in d ep en d en t va ri ab

le s ar e ge n d er , ag e,

ag e sq u ar ed , ef fe ct s o f fa m in e,

an d re gi o n al -f ix ed

ef fe ct . F o r ru ra l ar ea s,

th e co u n ty -f ix ed

ef fe ct s ar e co n -

tr o ll ed

fo r w h il e fo r ru ra l– u rb an

m ig ra n ts , b ir th

p la ce -f ix ed

ef fe ct s ar e co n tr o ll ed

fo r.

(4 ) V ar io u s sa m p le s ar e ex am

in ed

fo r ro b u st n es s ch ec k s,

fo r ex am

p le , al l p o p u la ti o n at

ag e o f 16 – 60

. A n d al so

in d iv id u al ’s

w ag e in co m e is

ta ke n as

th e

d ep en d en t va ri ab le

to ev al u at e th e im

p ac ts . C o n si st en t fi n d in gs

ar e fo u n d .

306 Dong Zhou

Send-Down). Differences in the interested estimates between before and after controlling for family background show that household background played a role to some extent. As discussed above, the rural residents had very different experiences. Compared with

permanent urban citizens, they were not exposed to such intensive human capital disrup- tions (large-scale school closures and the Send-Down Movement). Meanwhile, from the 1990s to 2002, annual income also largely increased for them because of the rapid growth in China. Therefore, it is expected that the Cultural Revolution did not have significantly persistent effects on the cohorts in the rural areas as the urban counterparts. In Table 6, I conduct multiple placebo tests with rural–urban migrants and permanent rural residents from CHIPS 1995 and 2002. First, the results for rural residents show that the CR, Clo- sure and Send-Down estimates are all insignificant and close to 0. Note that rural house- hold income from agriculture production as well as other family production is recorded collectively at the household level and is indivisible. Therefore, observations with non-farm individual revenue are examined. Second, as shown in Table 6, no negative impacts are found, and the coefficients of interest are insignificantly positive for rural–urban migrants. The sample of rural–urban migrants might have worked better as a counterfactual compar- ison because they migrated from rural areas after the 1990s, currently work in urban areas and share similar macroeconomic environments with the affected cohorts in the urban market. The insignificant estimates of interest that were obtained in all of the placebo-con- trolled studies provide convincing evidence supporting the causal effects of the Cultural Revolution on urban residents.

6. Conclusion

Following the existing literature, this paper documents and examines the impact of his- torical events on current economic development by focusing on one specific historic event in one specific country (the Cultural Revolution in urban China). Specifically, I gauge the impacts of the Cultural Revolution by testing its micro-effects on individuals’ average annual earnings and explore the channels through which the effects persist. Moreover, I examine how the impacts changed from the 1990s to the 2000s as the Chinese economy became more market-oriented. To accurately capture the impact, I designed three measures for the Cultural Revolution

(CR, Closure and Send-Down) based on synthetic cohort approach and utilized multiple datasets. Send-Down and Closure revealed different intensities at the cohort level from being subjected to two specific policy shocks: the closure of schools and the Send-Down Movement. In this way, selection bias and measurement error at the individual level were alleviated. In addition, the Great Famine was controlled for to avoid contamination. Consistent evidence supports that a Cultural Revolution effect lowered urban residents’

average annual income in the middle 1990s and early 2000s, and the effect was amplified as the market-oriented economy increasingly evolved. Based on the results from CHIPS 2002, the magnitude of the negative impact on average annual earning was approximately �11 per cent. The coefficient of school closure implies that one year of formal secondary schooling denied reduced the subjected individuals’ annual incomes by approximately 1.8 per cent. On average, the school closures lowered the exposed cohorts’ annual income by 2.7 per cent. Furthermore, the forced immigration (the Send-Down Movement) played a stronger role in lowering current earnings, by �4.8 per cent overall within the empirical sample. The existence of long-term impacts is robust to contemporaneous population

Historic Legacy Evaluation: the Cultural Revolution 307

comparisons and to a variety of controls for family background and different control groups. Placebo tests of rural residents as well as rural–urban migrants also support the existence of a negative causal effect of the Cultural Revolution on the subjected popula- tion’s permanent incomes. I also study the mechanism through which the impacts of the Cultural Revolution persist

based on sequential covariate addition. Evidence from the extended models reveals that the Cultural Revolution shaped many aspects of individuals’ life courses: education attain- ment, work experience, marriage history, health conditions, and attitudes toward the deter- minants of personal success. Among all of these explored micro-channels, education achievement is the major channel, accounting for approximately 50 per cent. Additionally, it is intriguing to find that the Send-Down Movement affected the exposed observations’ current incomes negatively through the health status channel and positively through the revealed attitudes channel. Their experiences made the sent-down youth more perseverant. Considering the complexity of the Cultural Revolution (which changed virtually every aspect of the exposed cohorts’ life courses), the results also cast doubt on the existing empirical papers that use the school closures as an IV to measure the real return of school- ing or that conclude a positive effect of the Send-Down Movement through simply com- paring the sent-down youth with the non-sent-down youth. There are still parts of the black box that remain unopened. Additional questions requir-

ing more study include the following. How did the Cultural Revolution impact the exposed populations’ health conditions, preferences, and personalities? How did the Cultural Revo- lution continue to have effects through institutional persistence? Separate from the impacts on labor market performance, we can also explore the impacts of the Cultural Revolution on other economic outcomes, for example, precautionary savings, consumption patterns, and intergeneration transmission of human capital.

Notes

1Selective Sending and Returning: the well-educated elite and the youth with particular family backgrounds were more likely to be sent down. Early returns to urban areas were related to regional policies and household social networks. Heterogeneous Exposures: the sent-down experi- ences varied depending on which regions youth had been sent to. Some places were difficult to live in because of the cumbersome manual labor. However, some rural areas had better environ- ments, for example, there were expansions of construction of senior high schools in some rural areas (Zhang, 2012) and less violence. Meanwhile, the youth who were not sent down during the same period encountered different conflicts in urban areas that were caused by the Cultural Revolution, for example, education interruptions, the Red Guard Campaign, and public shaming (Li et al., 2013; Zhou and Hou, 1999).

2In essence, the principle for categorizing the birth cohorts is from the most to the least affected. Considering different impacts among groups, for example, when one was at the first year of junior high schools (as well as senior high schools) or the last year of junior high schools (as well as senior high schools), I do regressions of the outcomes (schooling years) on group indicators for the measure of Closure as well as indicators for each birth cohorts controlling for time trend, gender, and pro- vince-fixed effects. The estimators of affected group indicators or birth year indicators follow a quasi- symmetric u shape with the lowest points at the birth cohorts of 1952, 1953, and 1954. Limited by information, construction of Closure in this paper is not perfectly index to catch the intensity of exposure to all levels of school closure but still a plausible proxy measuring the persistent effect. Different scales for the measure Closure are also considered for robustness checks. In all, the empirical results are qualitatively consistent.

308 Dong Zhou

3Those who were married to local residents or who worked in local government might have never returned (Li et al., 2013; Xie et al., 2008; Zhou and Hou, 1999).

4This identified group is the same as that in the existing literature (Han et al., 2011; Li et al., 2013; Meng and Gregory, 2003a; Xie et al., 2008; Zhang et al., 2007; Li et al., 2013). Other specifica- tions have also tested, for example, the birth cohorts 1946–1959 as well as 1947–1961.

5According to the literature, being exposed to famine during infanthood is likely to have a long- term effect on individuals’ future performance. Meng and Qian (2009) show that in utero and early childhood exposure to the Great Famine had large negative effects on adult health conditions, educa- tion attainment, and labor supply in rural areas.

6In this paper, I compute the average annual income between 1993 and 1995 as well as between 2000 and 2002 to measure the individual’s permanent income in the middle 1990s and the early 2000s, respectively, in consideration of measurement error.

7Different model specifications with various age polynomials are tested, and the statistics support the age-earning profile. As the parameter tests cannot reject the null hypothesis that the estimated cubic age is equal to 0 at any significance level below 35 per cent, the empirical model (1) is a better specification and fits better. Moreover, the estimation results obtained are consistent even when the cubic age is controlled for.

8To check the empirical robust estimation of sequential covariate addition, different orders of channels are examined and the conditional decomposition by Gelbach (2014) is applied. Consistent findings are found.

9This survey is funded by the Ford Foundation and institutions from multiple countries. Its fund- ing agencies are as follows: the Chinese Academy of Social Science, the Asian Development Bank, City University of New York, Leverhulme Trust (United Kingdom), Columbia University, and the University of California, Riverside.

10The total annual income contains regular wages, bonuses, other revenues from working units and subsidies from other sources. A trivial number of observations have not reported their incomes and these missing observations are eliminated.

11Special status includes people who were ‘Laid-off’ or currently ‘Lixiu’. I group them together into the class of ‘Laid-off’. These special groups are more likely to have worked in the state-owned sector, and the layoffs likely resulted from the state-owned enterprises’ (SOEs) efforts to deepen reform of the market-oriented economy after the late 1990s (Hung and Chiu, 2003). This population was given related subsidies and compensations. I do regression of these compensations on the policy measures and the results indicate that the affected cohorts have more these sources of income. Also regular wage is taken as a dependent variable to evaluate the long-term impacts. Consistent conclu- sions are found. The sample size of employed population is 9637 and evaluations of the impact within this subsample are also done to deeply understand the mechanism.

12Note that no information on length of stay in the rural areas is reported in CHIPS 1995. Hence, Send-Down is computed based on the average duration at cohort level merged from CHIPS 2002 for the 1995 wave. For robustness checks, the probability of being sent down in one wave was instru- mented by the other wave. Consistent conclusions are obtained.

13Evaluation of impacts on the channels as outcomes is provided in Appendix B to qualify and support the discussion. No matter for the empirical sample or employed empirical sample, consistent results are found. In addition, 1990 China census is utilized for demonstrating the credibility of the loss in educational attainment and more evidences supporting the effects on channels of working experience and employment status are also provided in Appendix B.

14To prove the negative impact of the school closure on the affected cohorts’ human capital accu- mulation, placebo tests are conducted across Asian countries (urban China, urban Indonesia, and urban Malaysia). Placebo tests for the impacts on income are also conducted through applying similar estimation strategies to other countries (for example, urban Brazil, urban India, urban Canada, and urban US). No similar impacts are found in other countries without such large-scale interruptions. Results can be provided if required.

Historic Legacy Evaluation: the Cultural Revolution 309

Appendix A The Send-Down movement and determinants of being sent down

A.1. Variations over birth cohorts and regions

Each scatter represents the proportion of the sample being rusticated at specific cohort level in Figure A1. The patterns over birth cohorts are consistent based on datasets of CHIPS 1995 and CHIPS 2002. The bumps indicate that the interested birth cohorts, 1946–61, were more likely to be rusticated during the Cultural

0 .1

.2 .3

.4 .5

P ro

b. o

f b ei

ng s

en t d

ow n

1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 Year of birth

1995 CHIPS 2002 CHIPS

Note: CHIPS, Urban Residents

Figure A1. The probability of being sent down, by birth cohort

Note: CHIPS, Urban residents.

Figure A2. The length of stay in the rural, by cohort (CHIPS 2002)

Note: CHIPS (2002), the sent-down youth.

310 Dong Zhou

Revolution. The older the cohorts, the longer they stayed in the rural areas statisti- cally (see Figure A2). I also checked different provinces and found that the interested birth cohorts, 1946–61,

were more likely to be rusticated during the Cultural Revolution, no matter for which pro- vinces. For the cohorts of 1946–61 as a group from different provinces have different likeli- hood to be sent on average (see Figure A3). Therefore, I also construct a new measure taking into provincial variations into considerations, and consistent conclusions are obtained. But, for keeping the consistency for comparisons among the three indexes, I choose to present the results of the index Send-Down focusing on cohort variations.

Send � Down intensityjo ¼ PjoLjo

Pjo: the probability of being sent down for birth cohort j and in province o. Ljo: average length of staying in the rural area for birth cohort j in province o.

A.2. The selection problem

In Table A1 (see next page), I examine determinants of being sent down for the interested cohorts 1946–61. They were graduates or current students during the Cultural Revolution. The results show that predetermined factors, for example: parents’ social status (‘Cheng fen’) and parents’ education attainments, played significant role in deter- mining the probability of being sent down for those children. Firstly, if parents were with higher education background, their kids were more likely to be sent down. Sec- ondly, if parents belonged to some particular social status (‘chengfen’), for example, rich peasants, office workers, petty proprietor, and revolutionary cadre, their children were more likely to be sent down to the rural areas. Thirdly, some regions have higher probabilities to send the youth to the rural areas. Additionally, I also explore their length of stay as the dependent variable. And their length of stay also to some extent significantly correlated with some characteristics of their family backgrounds. During

Figure A3. The probability of being sent down, by region (CHIPS 2002)

Historic Legacy Evaluation: the Cultural Revolution 311

the period from 1960s to 1980s, China was a planned economy and family income dif- ferences were trivial. Therefore, without controlling family income will not bias our esti- mation. As discussed in the Section 2, selection problem and heterogeneities among sent-down youth both cast doubts on the liability of the positive effects of the Send- Down Movement through simply comparing the sent-down and non-sent-down youth in the existing literature.

Table A1. Determinants for being sent down

Dependent variable

Sent-down

Father’s education Mother’s education

Base group: below primary Primary 0.062*** �0.0081

(0.02) (0.021) Junior high 0.057** 0.064*

(0.025) (0.033) Senior high �0.033 0.133***

(0.037) (0.042) College and above 0.074* 0.167***

(0.044) (0.077)

Father’s ‘Cheng fen’ Base group: poor peasant or landless

Lower middle peasant �0.079* (0.034)

Rich-middle peasant 0.083* (0.048)

Manual worker 0.131***

(0.043) Office worker 0.139***

(0.051) Petty proprietor 0.269***

(0.05) Revolutionary cadre 0.163**

(0.065) Parents’ occupations No/yes Mother’s ‘Cheng Fen’ Yes R2 0.1 Obs. 3,287

Notes: (1) The sample includes the urban residents born between 1946 and 1961. (2) The standard error is reported in the parenthesis. Only significant coefficients are reported in the table. (3) The dependent variable is a dummy represents whether one had been sent to the rural or not. Indepen- dent variables: father’s age, mother’s age, father’s occupation, mother’s occupation, father’s social status, mother’s social status, and province-fixed effect. (4) Considerable models specifications have been applied. The conclusion is consistent. Results of regres- sion are consistent with or without controlling parents’ occupations. (5) ***, **, and * represent significance at 1, 5 and 10 per cent, respectively.

312 Dong Zhou

Appendix B Impacts on the channels in the mechanism

B.1. Evaluations with channels as outcomes

B.2. Channel of education attainment

Table B1. The impacts on channels as outcome variables, CHIPS 2002

Dependent variables

Urban birth cohorts: 1942–77

Schooling yearsa

Working experiencea

Working experienceb

Marital stabilitya Heightc Attituded

CR �0.637*** 0.474 0.186 �0.125*** �0.188 0.011 (0.099) (0.288) (0.168) (0.036) (0.18) (0.04)

Closure �0.122*** 0.085** 0.074** �0.0211*** �0.004 �0.005 (0.019) (0.038). (0.034) (0.006) (0.04) (0.008)

Send-Down �0.419*** 0.262 0.230** �0.078*** 0.04e 0.042** (0.081) (0.16) (0.093) (0.022) (0.122) (0.021)

R2 0.11 0.72 0.75 0.21 0.58 0.03 Obs. 12,304 9,571 12,424 12,304 8,856 3,855

Notes: (1) Other control variables include gender, age, age squared, famine effects, and regional-fixed effects. One cell represents result of one baseline estimation. (2) aCHIPS 2002, bCHIPS 1995, cCHIPS 2007, dCFPS 2008, ethe measure of Send-Down is merged from the 2002 wave of CHIPS. In the wave of 2002 CHIPS, only employed population are asked the question associated with how many years they have been employed or working. (3) ***, **, and * represent significance at 1, 5 and 10 per cent, respectively. The robustness standard error is reported in the parentheses and adjusted for clusters in age.

Table B2. Evaluation of channels within employed urban population, CHIPS 2002

Dependent variables Schooling years Working experiences Marriage status

CR �0.792** 0.355 �0.129** (0.168) (0.28) (0.04)

Closure �0.147** 0.084* �0.021** (0.024) (0.032) (0.006)

(0.005) Send-Down �0.416** 0.2442 �0.071**

(0.11) (0.157) (0.022) Obs. 9,637 9,637 9,637 Other independent variables: gender, age, age square, famine effect, and province-fixed effect

Notes: (1) The sample includes the employed urban residents aged between 25 and 60 in 2002. (2) The standard error is reported in the parenthesis and adjusted for 36 clusters in age. (3) Constructions of CR, Send-Down, Closure and Famine are discussed in Section 2. (4) The marriage status is a dummy variable and represents whether one is married with spouse in 2002. Working experience is how many years they have been working. (5) ***, **, and * represent significance at 1, 5 and 10 per cent, respectively.

Historic Legacy Evaluation: the Cultural Revolution 313

B.3. Channels of working experience and employment status

B.4. More robustness for the mechanism

Table B3. The impact on education attainments (Urban residents, 1990 Census)

Dependent variable: different levels of education attainment (China Census)

Measure Higher Senior Junior Primary

CR �0.01*** �0.06*** �0.02*** �0.002*** (0.002) (0.008) (0.008) (0.0005)

Obs. 831,124 831,124 831,124 831,124 Other independent variables: age, age square, famine effect, birth cohorts, and province-fixed effect

Notes: (1) Sample selected: birth cohorts 1935–75 in permanent urban residents. (2) The standard error is reported in the parenthesis and adjusted for clusters in age. (3) *** represent significance at 1 per cent, respectively.

Table B4. Impacts on employment status in urban labor market

Dependent variable: employment status (CHIPS 2002)

Probability models CR Closure Send-Down

Panel A: male sample Linear model �0.027** �0.007*** �0.022***

(0.01) (0.001) (0.007) F-statistics 7.49 8.26 6.53 Probit model �0.196** �0.053*** �0.16***

(0.083) (0.01) (0.047) Wald v2 136.11 151.79 125.31 Logit model �0.401** �0.107*** �0.325***

(0.177) (0.02) (0.097) Wald v2 129.24 149.7 120.39 Obs. 5,784 5,784 5,784

Panel B: female sample Linear model �0.034* �0.008* �0.013

(0.02) (0.004) (0.011) F-statistics 14.41 15.09 17.61 Probit model �0.169** �0.036* �0.066

(0.085) (0.02) (0.055) Wald v2 174.47 137.02 201.0 Logit model �0.3* �0.068 �0.111

(0.16) (0.042) (0.103) Wald v2 168.99 167.41 191.41 Obs. 4,873 4,873 4,873

Notes: (1) The sample includes permanent urban residents at ages of 25–60 in the labor force in 2002, excludes the full-time students and homemakers and also drops those retired and the disable. (2) The standard error is reported in the parenthesis and adjusted for 36 clusters in age. Other control vari- ables include gender, age, age squared, famine effects, and province indicators. (3) ***, **, and * represent significance at 1, 5, and 10 per cent, respectively.

314 Dong Zhou

Appendix C Changes of the impact through tracking the same cohorts over time

References

Acemoglu D., Johnson S. and Robinson J. A. (2001) ‘The Colonial Origins of Comparative Develop- ment: An Empirical Investigation’, American Economic Review 91: 1369–1401.

Banerjee A. and Iyer L. (2005) ‘History, Institutions, and Economic Performance: The Legacy of Colonial Land Tenure Systems in India’, American Economic Review 95(4): 1190–1213.

Barr T. and Lin C. (2015) ‘A Detailed Decomposition of Synthetic Cohort Analysis’, Economics Let- ter 127: 76–80.

Bernstein P. T. (1997) Up to the Mountains and Down to the Villages: The Transfer of Youth From Urban to Rural China. New Haven, CT: Yale University Press.

Deaton A. and Paxson C. (2000) ‘Saving and Growth Among Individuals and Households’, The Review of Economics and Statistics 82(2): 212–225.

Dell M. (2010) ‘The Persistence Effects of Peru’s Mining Mita’, Econometrica 78(6): 1863–1903. Engerman L. S. and Sokoloff L. K. (1997) Factor Endowments, Institutions, and Differential Paths of

Growth Among New World Economies: A View From Economic Historians of the United States. Palo Alto, CA: Stanford University Press: pp. 260–304.

Gelbach J. B. (2014) ‘When Do Covariates Matter? And Which Ones, and How Much?’, Journal of Labor Economics, Forthcoming.

Glaeser L. E. and Shleifer A. (2002) ‘Legal Origins’, Quarterly Journal of Economic 117: 1193–1230. Glenn D. N. (1989) ‘A Caution About Mechanical Solutions to the Identification Problem in Cohort

Analysis: Comment on Sasaki and Suzuki’, American Journal of Sociology 95: 754–761. Han J., Suen W. and Zhang J. S. (2011) ‘Picking up the Losses: The Impact of the Cultural Revolu-

tion on Human Capital Reinvestment in Urban China’, Working Paper. Hanoch G. and Honig M. (1985) ‘True Age Profiles of Earnings: Adjusting for Censoring and for

Period and Cohort Effects’, Review of Economics and Statistics 67(3): 383–394. Hung P. W. and Chiu W. K. (2003) ‘The Lost Generation: Life Course Dynamics and Xiagang in

China’, Modern China 29(2): 204–236. Ichino A. and Winter-Ebmer R. (2004) ‘The Long-run Educational Cost of World War II’, Journal

of Labor Economics 22(1): 57–86.

Table C1. The changes of the impacts

Samples Birth cohorts: 1942–70

Measures CHIPS 1995 CHIPS 2002

CR �0.061** �0.103*** (0.023) (0.030)

Closure �0.007 �0.01 (0.004) (0.005)

Send-Down �0.033*** �0.04** (0.010) (0.017)

R2 0.24 0.15 Obs. 10,801 10,896

Note: The robustness standard error is reported in the parentheses and adjusted for clusters in age; other control variables include gender, age, age squared, famine effects, and province indicators.

Historic Legacy Evaluation: the Cultural Revolution 315

Irvine J. I. (1981) ‘The Use of Cross-Section Microdata in Life Cycle Models: An Application to Inequality Theory in Nonstationary Economies’, The Quarterly Journal of Economics 96(2): 301–316.

Iyer L. (2007) ‘Direct Versus Indirect Colonial Rule in India: Long-Term Consequences’, Harvard Business School Working Paper, No. 05-041.

Jappelli T. (1999) ‘The Age-Wealth Profile and the Life-Cycle Hypothesis: A Cohort Analysis With a Time Series of Cross-Sections of Italian Households’, Review of Income and Wealth 45(1): 57–75.

King, Mervyn A., and Louis Dicks-Mireaux. Portfolio composition and pension wealth: an econo- metric study. National Bureau of Economic Research, 1982.

Li H. B., Rosenzweig M. and Zhang J. S. (2013) ‘Altruism, Favoritism, and Guilt in the Allocation of Family Resources: Sophie’s Choice in Mao’s Mass: The Send-Down Movement’, Journal of Political Economy 118: 1–38.

Masson A. (1996) ‘A Cohort Analysis of Wealth-age Profiles Generated by a Simulation Model in France (1949-75)’, Economic Journal 96: 173–190.

Meng X. and Gregory R. (2003a) ‘The Impact of Interrupted Education on Subsequent Educational Attainment–A Cost of the Chinese Cultural Revolution’, Economic Development and Cultural Change 50(4): 935–959.

Meng X. and Gregory R. (2003b) ‘Exploring the Impact of Interrupted Education on Earnings: The Educational Cost of the Chinese Cultural Revolution’, IZA Discussion Papers: 2548.

Meng X. and Qian N. (2009) ‘The Long Term Consequences of Famine on Survivors: Evidence From a Unique Natural Experiment Using China’s Great Famine’, NBER Working Paper: 14917.

Mincer J. (1997) ‘The Production of Human Capital and the Life Cycle of Earnings: Variations on a Theme’, Journal of Labor Economics 15(1): S26–S47.

Modigliani F. (1986) ‘Life Cycle, Individual Thrift, and the Wealth of Nations’, American Economic Review 76: 297–313.

Nunn N. (2007) ‘Historical Legacies: A Model Linking Africa’s Past to its Current Underdevelop- ment’, Journal of Development Economics 83: 157–175.

Nunn N. (2008) ‘The Long-Term Effects of Africa’s Slave Trades’, Quarterly Journal of Economics 123(1): 139–176.

Nunn N. (2009) ‘The Importance of History for Economic Development’, NBER Working Paper, 14899.

Ryder N. B. (1965) ‘The Cohort as a Concept in the Study of Social Change’, American Sociological Review 30: 843–861.

Sasaki S. M. and Suzuki T. (1989) ‘A Caution About the Data to be Used for Cohort Analysis: Reply to Glenn’, American Journal of Sociology 95: 761–765.

Shorrocks A. F. (1975) ‘The Age-Wealth Relationship: A Cross-Section and Cohort Analysis’, Review of Economics and Statistics 57: 155–163.

Singer M. (1971) Educated Youth and the Cultural Revolution in China, Vol. 10. Ann Arbor, MI: Center for Chinese Studies, University of Michigan.

Thøgersen S. (2002) A county of culture: twentieth-century China seen from the village schools of Zouping, Shandong: University of Michigan Press.

Treiman J. D. and Deng Z. (1997) ‘The Impact of the Cultural Revolution on Trends in Educational Attainment in the People’s Republic of China’, American Journal of Sociology 103(2): 391–428.

Winship C. and Harding J. D. (2008) ‘A Mechanism-Based Approach to the Identification of Age- Period-Cohort Models’, Sociological Methods and Research 36(3): 362–401.

Xie Y., Jiang Y. and Greenman E. (2008) ‘Did Send-Down Experience Benefit Youth? A Reevalua- tion of the Social Consequences of Forced Urban-Rural Migration During China’s Cultural Revolution’, Social Science Research 37: 686–700.

316 Dong Zhou

Zhang J. S., Liu P. K. and Yung L. (2007) ‘The Cultural Revolution and Returns to Schooling in China: Estimates Based on Twins’, Journal of Development Economics 84(2): 631–639.

Zhang S. (2012). Mothers education and infant health: evidence from closure of high schools in China. Cornell University working paper.

Zhou X. G. and Hou L. R. (1999) ‘Children of the Cultural Revolution: The State and the Life Course in the People’s Republic of China’, American Sociological Review 64(1): 12–36.

Historic Legacy Evaluation: the Cultural Revolution 317

Copyright of LABOUR: Review of Labour Economics & Industrial Relations is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.