Education's role in China's structural transformation

https://doi.org/10.1016/j.jdeveco.2012.10.006Get rights and content

Abstract

We explore education's role in improving the allocation of labor between China's agricultural and nonagricultural sectors and measure the portion of China's recent growth attributable to this channel. Using detailed micro-level data and an empirical model that allows for the endogenous selection of education and sector of employment, we estimate the relationship between an individual's educational attainment, sector, and income. We find that about 11% of aggregate growth in output per worker from 1978 to 2004 is accounted for by increased education, with 9% coming through the labor-reallocation channel and 2% attributable to increased within-sector human capital.

Introduction

A burgeoning macro-development literature has emphasized that the distribution of factors of production significantly affects economy-wide output. In particular, many recent studies have focused on the allocation of labor between the agricultural and nonagricultural sectors and have used the cross-sector difference in labor productivity (i.e., output per worker) as a measure of the (in)efficiency of this allocation. A corollary to the finding that countries in which inputs are allocated to their most-productive uses have higher incomes is that improvements in the input distribution can lead to periods of rapid economic growth (e.g., Restuccia et al. (2008) for cross-country income differences; Caselli and Coleman (2001) for the United States; Hayashi and Prescott (2008) for Japan; Brandt et al. (2008) for China). Economists have used the term “structural transformation” to describe such periods, when inputs are reallocated from the low-productivity agricultural sector to non-agriculture.

China has experienced a particularly remarkable structural transformation over the past three decades, with rapid economic growth accompanied by a decrease in the employment share of the agricultural sector. Young (2003) estimates growth in aggregate output per worker of 5.2% from 1978 to 1998, while Brandt et al. (2008) report growth of 7% from 1978 to 2004. Concurrently, the agricultural employment share dropped from 71% in 1978 to 47% in 2004 (China Statistical Yearbooks). Because labor productivity was much higher in the nonagricultural sector than in agriculture, estimates of the fraction of growth in aggregate output per worker accounted for by labor reallocation are sizable, ranging between 25 and 33% (Brandt et al. (2008) and Dekle and Vandenbroucke (2012), respectively).

Although many forces have undoubtedly contributed to labor reallocation, in this paper we aim to quantify the role of one particular driving force, increased educational attainment. Our interest in education stems from the following observations. First, the educational attainment of Chinese workers has increased smartly in recent years. For example, the share of workers who had completed middle school was 37% in 1982 and 63% in 2005 (Chinese Census, 1% Population Survey). Second, a worker's educational attainment and her likelihood of working in the nonagricultural sector are positively correlated. According to the 2005 1% Population Survey, the fraction of workers employed in nonagricultural jobs was about 40 percentage points higher among workers who had completed middle school than among workers who had not. Of course, correlations do not establish that education drives sector choice. We thus estimate a structural model of education and sector choice using detailed micro-level data to investigate the relationship further.

Our hypothesis is that some factors exist that reduce the benefits of working in higher-productivity (higher-paying) nonagricultural jobs, but that education mutes the strength of these factors. We have in mind a number of such factors: (a) the costs of collecting/processing information about potential jobs, (b) the cost of dealing with migration regulations1 or, more generally, moving in search of a job, and (c) individual preferences for working in certain types of jobs (e.g., those that require “brawn” more than “brains”). Note that we will not attempt to measure education's effect through each individual factor but will instead quantify education's impact on economic growth though facilitating labor reallocation more generally. To provide a metric for assessing the quantitative importance of reallocation between sectors, we will also measure education's contribution to growth through increasing human capital within sectors.2

We build a model in which individuals choose their level of education and sector of employment based on their observable characteristics and unobservable sector-specific ability. We assume education increases workers’ human capital, possibly at a different rate in each sector, and also affects sector choice through non-income factors. Individuals make their choices to maximize the expected present discounted value of their lifetime income, net of education costs but inclusive of non-income considerations. In our model, an individual's sector-specific income depends on his unobservable ability as well as observable characteristics standard to Mincerian regressions, such as age, race and gender. We follow structural labor models of sector choice (e.g., Lee and Wolpin, 2006) by including non-income factors that depend on observable characteristics but, conditional on observables, are uncorrelated with earnings. Finally, education costs also depend on observable individual characteristics including school accessibility at the time an individual makes his schooling choice. We estimate the model using simulated method of moments and multiple waves of the China Health and Nutrition Survey, a panel survey containing data on the occupation, education, and income of individuals.3

We find that for nonagricultural workers, the completion of middle school translates into 12% more income, whereas for agricultural workers, the increase in income is around 8%. For the median worker, completing middle school increases the likelihood of working in the nonagricultural sector by 42%. Moreover, even after controlling for worker heterogeneity, a large income difference still exists between agricultural and nonagricultural jobs. For example, the average agricultural worker could increase her income by roughly 2 times by moving to the nonagricultural sector. This is a sizable gain, but smaller than the income difference of 2.75 that results from simply comparing average incomes across sectors. This finding hints at the importance of incorporating heterogeneity in the empirical analysis, which has previously been emphasized in a general context by Heckman (2001) and for economic growth by Banerjee and Duflo (2005).

With our micro estimates in hand, we then perform a growth accounting exercise that decomposes education's effect on aggregate growth into a within-sector channel – raising human capital – and a between-sector channel – reallocating workers from agriculture to non-agriculture. Education's within-sector contribution to growth is calculated from sector-specific returns to education estimated from the micro data. Education's contribution to growth through labor reallocation is a function of both the increase in nonagricultural workers that is attributable to increased educational attainment and the nonagricultural-agricultural productivity differential for these workers. Our main findings are that roughly 11% of the aggregate growth in output per worker in China from 1978 to 2004 is accounted for by increased education, with 9% coming through the labor-reallocation channel. The finding that the labor-reallocation channel is relatively more important is robust to several alternative specifications of our empirical model.

Our paper is related to several recent studies that attempt to identify the driving forces behind labor reallocation in developing economies. For China, Brandt et al. (2008) and Dekle and Vandenbroucke (2012) have calibrated two-sector general equilibrium models and conduct counterfactual simulations to gauge the importance of various driving forces, such as changes in sectoral TFP, the rate of fixed investment, and labor market frictions. Song et al. (2011) discuss how financial frictions sustain large productivity differences across the state and non-state sectors in China and how reducing these frictions could have led to reallocation-driven growth. The quantitative models in these studies are calibrated to match aggregated time series. Our approach differs by using household-level data to estimate education's effect on sector choice while controlling for a rich amount of heterogeneity. The use of micro data makes our approach more closely related to the work of Midrigan and Xu (2010), who explore the role of financial frictions for the allocation of resources between manufacturing plants using establishment-level data for Korea and Columbia.

In addition to the macro-development literature, this paper contributes to two strands of the applied micro literature. The first measures returns to education. Returns to education in China have generally been found to be lower than those in comparable developing countries.4 Our findings suggest that education returns in China are largely attributable to education increasing the likelihood of working in non-agriculture rather than increasing within-sector human capital. Note that many studies in the literature do not account for sector choice and estimate returns only for the nonagricultural sector, which may explain the low returns to education.

The second literature examines the determinants of and potential income gains associated with migration in China.5 Our finding that, all else equal, more-educated individuals are more likely to leave agriculture is consistent with evidence that relaxed migration regulations reduced educational attainment (e.g., de Brauw and Giles, 2008). Furthermore, because nonagricultural jobs are more abundant in urban areas, our estimates of the income gains from reallocating to non-agriculture can be interpreted as part of the potential gains from migration.

The rest of the paper is organized as follows. In Section 2, we motivate our investigation of the roles of education and labor reallocation in accounting for recent Chinese economic growth. Section 3 lays out a two-sector growth-accounting framework and identifies the variables we need to estimate from micro-level data to account for education's contribution. Section 4 describes our empirical model of education and sector choice, and Section 5 describes our data and reports estimates of education's impact on individual outcomes. In Section 6, we then present measures of education's effect on aggregate output growth. Section 7 concludes.

Section snippets

Motivation

Two facts indicate that cross-sector labor reallocation may have been an important contributor to China's economic growth over the past three decades. First, the employment share of the nonagricultural sector increased from 30% in 1978 to 53% in 2004 (China Statistical Yearbooks). Second, a significant gap between per-worker output in the two sectors persisted throughout this period. Fig. 1 shows that nonagricultural output per worker was 4–6 times larger than agricultural output per worker

Growth-accounting framework

We consider a two-sector growth-accounting framework that decomposes the growth rate of aggregate output per worker into three terms: growth of output per worker in the agricultural sector, growth of output per worker in the nonagricultural sector, and growth related to the increased employment share of the nonagricultural sector (i.e., labor reallocation). Our focus will be on education's contribution to each of these three components, but especially the last.

Empirical model

In this section, we describe the empirical model we use to estimate education's effect on sector choice and labor productivity.

Data

Estimating our model requires micro-level data with detailed information about workers’ income, education, and sector of employment. We use data from the China Health and Nutrition Survey (CHNS),11

Education's contribution to economic growth

We now assess the role of increased educational attainment in accounting for growth in Chinese output per worker from 1978 to 2004. As discussed in Section 3, education's contribution to output per worker came through two channels: increasing within-sector human capital and facilitating labor reallocation from agricultural jobs to higher-productivity nonagricultural jobs. We find that the latter channel has been more important.

Table 6 displays our estimates of education's contribution to growth

Conclusion

This paper examines education's role in accounting for recent Chinese economic growth through two channels: facilitating labor reallocation from the agricultural sector to the nonagricultural sector and increasing human capital within sectors. We find that a large fraction of education's contribution comes through facilitating labor reallocation, a channel that had not previously been studied. Moreover, we show that incorporating individual heterogeneity in the analysis is essential for not

References (49)

  • D. Benjamin et al.

    Markets, Human Capital, and Inequality: Evidence from Rural China

  • M. Bils et al.

    Does schooling cause growth

    American Economic Review

    (2000)
  • L. Brandt et al.

    Accounting for China's Growth

  • L. Brandt et al.

    Growth and Structural Transformation in China

  • R.P. Byron et al.

    Returns to education in China

    Economic Development and Cultural Change

    (1990)
  • P. Cahuc et al.

    Labor Economics

    (2004)
  • F. Caselli et al.

    The U.S. structural transformation and regional convergence: a reinterpretation

    Journal of Political Economy

    (2001)
  • K.W. Chan et al.

    Is China abolishing the Hukou system?

    China Quarterly

    (2008)
  • K.W. Chan et al.

    The Hukou system and rural–urban migration in China: processes and changes

    China Quarterly

    (1999)
  • Giles A.J. de Brauw

    Migrant Opportunity and the Educational Attainment of Youth in Rural China

  • A. de Brauw et al.

    Reconciling the returns to education in off-farm wage employment in rural China

    Review of Development Economics

    (2008)
  • S. de Mel et al.

    Measuring Microenterprise Profits: Don't Ask How the Sausage is Made

  • Feder, G., Lau, L. J., 1991. “Economies of Scale in Chinese Agriculture: Evidence from Microeconomic Data,” Unpublished...
  • E.A. Hanushek et al.

    The role of cognitive skills in economic development

    Journal of Economic Literature

    (2008)
  • Cited by (18)

    • Effects of payments for ecosystem services programs in China on rural household labor allocation and land use: Identifying complex pathways

      2020, Land Use Policy
      Citation Excerpt :

      Second, females may face more barriers than males in searching for jobs in off-farm labor markets, and are socially obligated to take care of children, the elderly, and disabled members in the household (Hajjar et al., 2020). Finally, a lower education limits a person’s access to opportunities in local off-farm work or for out-migration for employment in cities (Lee and Malin, 2013), so the less educated are more likely to engage in on-farm work (Wei et al., 2016). In some cases, eductaion is found to have a nonlinear relationship with migration decisions and/or city destinations (Hao and Tang, 2018; Xiao and Zhao, 2018).

    • Performance evaluation of China's Hi-tech zones in the post financial crisis era - Analysis based on the dynamic network SBM model

      2015, China Economic Review
      Citation Excerpt :

      At present, there are invisible barriers hindering the flow of resources that affect the technological innovation flow among different administrative regions in China, such as local protectionism, scientific research system, and household registration system. Therefore, the R&D TEs in HTZs depend on a relative abundance of local scientific and technical resources as well as an institutional environment for innovation (Lee & Malin, 2013). For example, Beijing, Shanghai, and Shaanxi are provinces with well-developed science and education resources, and hence their number of scientific research institutions and key universities far surpasses other provinces.

    • Indian Agriculture vis-à-vis Youth: Evidence from Tamil Nadu

      2022, Indian Journal of Economics and Development
    View all citing articles on Scopus

    An earlier version of this paper circulated under the title “The Role of Education in Economic Growth through the Sectoral Reallocation of Labor.” We have benefitted from discussions with Ginger Jin, Pete Klenow, John Pencavel and seminar participants at Aarhus, BLS, Federal Reserve Board, Free University of Berlin, George Washington, Stanford, 2006 SED, 2007 PACDEV, 2010 Seoul Summer Economics Conference, and the International Symposium on Econometric Theory in Honor of Takeshi Amemiya's Contribution to Econometrics. We thank Dongfang Shao for providing access to a digital version of the CSY data sets and Shufa Du for answering numerous questions regarding the CHNS. Lee acknowledges the Taube Fellowship for financial support. The views expressed here are those of the authors and do not necessarily reflect the views of the Federal Reserve System.

    View full text