Abstract: Based on the accurate definition of the statistical caliber of rural employed persons, this paper uses the 2015 Special Survey Data of Outflow Monitoring released by the National Health and Family Planning Commission of China to construct a multiclass prediction model based on the selection of rural labor employment types. Then using the supervised machine methods of AdaBoost algorithm, this paper estimates the absolute surplus labor force and relative surplus labor force in rural areas in 2010—2018. The results show that the absolute surplus labor force in rural areas in 2018 is 3692 million, 806% of which is the explicit surplus labor in the form of open unemployment; the relative surplus labor force in 2018 is 5369 million, and only 1051 million of them are the outside transferable surplus labor force. This paper judges that China has stridden across the first Lewis turning point, and will continue to face the dual dilemma of labor shortage in agricultural production and insufficient labor supply in nonagricultural sectors. In this regard, this paper puts forward policy recommendations such as promoting full employment of agricultural labor, improving the efficiency of the allocation of rural production factors, promoting the local transfer of rural labors and eliminating barriers to labor transfer across regions.

Key words:

摘要： 本文在准确界定乡村就业人员统计口径基础上，利用国家计生委2015年流出地监测专项调查数据，构建基于农村劳动力就业类型选择的多分类预测模型，采用有监督的机器学习的AdaBoost算法，对2010—2018年农村绝对剩余劳动力和相对剩余劳动力数量进行估算。研究结果发现，2018年农村绝对剩余劳动力数量为3 692万，其中806%是以公开失业为表现形式的显性剩余劳动力；2018年农村相对剩余劳动力数量为5 369万，其中能够向外出就业转移的相对剩余劳动力仅为1 051万。本文判断我国已经从总体上跨越刘易斯第一转折点，并且将持续面临农业生产劳动力紧张和非农部门劳动力供给不足的双重困境。对此本文提出促进农业劳动力充分就业、提升农村生产要素配置效率、推动农村劳动力就地转移以及消除劳动力跨区域转移障碍等政策建议。

关键词: