China's Rural Labor Migration and Its Economic Development. Xiaoguang Liu

development. Therefore, both at the industrial level and at the spatial level, the allocation efficiency of China’s labor resources continues to improve.

      The continuous large-scale transfer of agricultural labor has not only completely changed the fundamental characteristics and the efficiency of the configuration of China’s labor market but also profoundly affected China’s investment, savings, technological progress, urban–rural income distribution and macroeconomic fluctuations, and more importantly, it has played a vital role in the rapid development of urban sectors. Research shows that the transfer of agricultural labor is closely associated with the improvement of China’s total factor productivity,⁴ the rapid development and exportation of the manufacturing industry,⁵ the high savings rate and the high rate of investment,⁶ the change in the pattern of income distribution⁷ and other important macroeconomic characteristic phenomena. Du Yang et al. discovered through their latest research that the flow of the labor force from rural areas to urban areas is conducive to expanding the size of the labor market and improving the total factor productivity of the urban economy. The net benefits brought by the flow of labor are still considerable, despite the negative impacts on the capital–output ratio and working hours.⁹

      Figure 2.2. The transfer of China’s agricultural labor (1985–2014).

      Source: The data for the period 2008–2014 were extracted from the National Monitoring Survey Report on Migrant Workers by the National Bureau of Statistics over the years; for the data regarding the period 1985–2007, see Lu Feng: Reflection on the Economic Catch-up by a Large Country — An Understanding of China’s Open Macro Economy (2003–2013) (Volume 1), Peking University Press, 2014, Edition 1.⁸

      

      In the future, there is still a lot of room for the further transfer of agricultural labor in China. Similar to the extensive international experience, the proportion of China’s agricultural labor force has fallen since the reform and opening-up. Over the past 30 years and more, the proportion of the agricultural labor force has decreased by an average of more than 1 percentage point annually, from over 70% on the eve of the reform to about 31.4% in 2013. The international experience reveals that the proportion of the agricultural labor force in developed countries usually drops to less than 10% (see Figure 2.3) with the improvement of the level of economic development. Over the past century, for instance, the average proportion of the agricultural labor force in Organisation for Economic Co-operation and Development (OECD) countries has declined from 53% to about 10% today. Lu Feng and Yang Yewei speculate that the proportion of China’s agricultural labor force will decline to 13.6% in 2030, indicating the great potential for the transfer of China’s agricultural labor force in the next 20 years.¹⁰ Du Yang et al. also call for comprehensively intensifying the reform of the household registration system, to further promote labor mobility and “sustain the China Miracle by reaping the dividends from Hukou reforms”.¹¹

      Figure 2.3. The level of economic development of each country and the proportion of agricultural labor force.

      Source: World Bank WDI Database.

II. Driving Factors for the Transfer of Agricultural Labor

      Fresh and valuable materials have been provided by the practical experience of economic development and labor transfer during the period of China’s reform and opening-up to understand the relationship between economic transformation and labor transfer discovered by developmental economics. The group of the large-scale transfer of laborers constitutes an important supporting force for China’s economic growth, which helps expand the size of the labor market and increase the total factor productivity of the urban economy. The key to the long-term rapid growth of the Chinese economy will still lie in the continued transfer of surplus labor, which will bring about obvious benefits to China’s economic development in the next few years.¹²

      

      After more than 30 years of rapid economic growth, China’s economy has entered a critical stage of optimization of the economic growth rate and structural adjustment. On the one hand, China still has a lot of agricultural labor force to be transferred because the existing agricultural labor force accounts for more than 30%. On the other hand, a “labor short-age” has occurred frequently in the developed provinces along the east coast in recent years. Many enterprises face a labor shortage, despite the rapid rise in labor wages. Labor shortage has been an important factor in the declining international competitiveness of Chinese enterprises. In this context, how to promote the continued transfer of the surplus agricultural labor force is an important issue to be addressed urgently.

      To solve this practical problem, it is theoretically necessary to further clarify the determinants of labor transfer. In the past 30 years or so, why has the rural labor force shifted on a large scale and shown the extremely obvious characteristics of spatial agglomeration? What are the major factors that have affected this process and made it a turning point? How does the difference between urban and rural productivity form and act on the transfer of labor? In addition to the differences in productivity between urban and rural sectors, are there more specific factors that affect the transfer of labor in China? Only when these general and special factors are clarified can we truly grasp the internal laws of the historical process of the transfer of labor in China to bring forward pertinent suggestions and measures for the further promotion of the transfer of labor and for the reform of the labor factor market and economic growth in the next stage.

      

      Therefore, this section focuses on the driving factors of the transfer of agricultural labor by means of a theoretical and empirical analysis. Due to limitations of space, the theoretical analysis is limited and can be found in the “Appendix B: Driving Factors and Spillover Effects of the Transfer of Agricultural Labor”. The processes and results of the empirical analysis are described in the following sections.

1. Econometric model, regression variables and data description

      This section shows the empirical analysis of the determinants of the transfer of agricultural labor by the use of the panel data of China’s provinces and regions. Before the regression analysis, it is necessary to briefly introduce the methods of measuring to be used. In view of the fact that the core variables examined are likely to be affected by spatial correlation, the spatial econometric model is used for the regression analysis. The research by Xu Haiping and Wang Yuelong reveals that a significant autocorrelation exists in the urban–rural income gap in China’s provinces, municipalities and autonomous regions in terms of spatial distribution, and the studies by Luo Yongmin and Zhang Guangnan also indicate that infrastructure has spatial spillover effects.¹³ A spatial correlation may be derived from the system of economic variables under consideration or from the spatial correlation of the terms of error. Therefore, depending on the source of the effect of spatial correlation, the model of spatial measurement can be divided into the spatial autoregressive model (SAR) and the spatial error model (SEM). By reference to the practices in the above literature, both SAR and SEM models are used for analysis in this section to overcome the influence of potential spatial correlation and carry out maximum likelihood estimation. The same is also true for the setting of a spatial weight matrix. The weight coefficient of adjacent provinces is set as 1, and the weight coefficient of non-adjacent provinces is set as 0.¹⁴ The weight matrix is standardized in the specific measurement estimation. For purposes of comparison, the regression results under the two models are reported symmetrically in each table.

      

      The following sample interval analyzed