自动化资本偏向型技术进步与就业的非线性关系—基于中国制造业上市公司的证据
The Non-linear Relationship between Automated Capital Biased Technical Change and Employment
-
摘要: 基于2012—2019年制造业上市公司数据,可以理论分析和实证检验自动化资本偏向型技术进步对就业总量及其结构性影响和内在机制。研究发现:(1)制造业自动化资本偏向型技术进步与就业总量呈现“U型”的非线性关系,对高、低技能劳动力就业的影响分别为“倒U型”“U型”。(2)企业生产率和产出规模在自动化资本偏向型技术进步对就业总量以及技能结构的影响中发挥了非线性效应。对于研究的样本而言,自动化资本偏向型技术进步通过生产率提升效应和规模降低效应对高技能劳动力就业产生了创造效应;而生产率提升效应和规模降低效应对就业总量和低技能劳动力产生了替代效应。因此,提升低技能劳动力的技能水平,大力培育高技能劳动力,强化自动化资本偏向型技术进步与劳动力的技能匹配,深入挖掘生产率和规模效应在自动化技术深化过程中对就业的积极作用,是适应“机器换人”,实现稳就业、促进高质量就业的重要途径。Abstract: Based on data of the listed manufacturing companies from 2012 to 2019, this paper theoretically analyzes and empirically tests the impact and internal mechanism of automated capital-biased technical change on the total employment and structure. The research found: (1) There is a U-shaped non-linear relationship between the automated capital-biased technical change and total employment, and the impact of the technical change on the employment of high- and low-skilled labor is inverted U-shape and U-shape respectively. (2) Enterprise productivity and output scale play a non-linear effect on the impact of automated capital-biased technical change on the total employment and labor skill structure. For the sample in this paper, automated capital-biased technical change has produced a creative effect on the employment of high-skilled labor through the productivity improvement effect and the scale reduction effect, while which reduce total employment and the low-skilled labor force. Therefore, improving the skill level of low-skilled labor, vigorously cultivating high-skilled labor, strengthening the skill matching of automation capital-biased technical change with labor force, and exploring in-depth the positive role of productivity and scale effects for development of automation technology are important measures to achieve stable employment and promote high-quality employment.
-
Key words:
- automated capital /
- biased technical change /
- employment
-
表 1 变量说明及描述性统计
变量类型 变量名称 变量符号 观测值 均值 标准差 最小值 最大值 被解释变量 就业总量对数值 lnl 2898 7.5125 1.1735 5.1180 10.6705 低技能劳动力就业数量对数值 lnL 2898 7.4531 1.1820 4.9488 10.5725 高技能劳动力就业数量对数值 lnZ 2898 5.7247 1.1927 3.0910 8.9882 核心解释变量 自动化资本偏向型技术进步指数 bias 2898 0.1698 0.1935 0.0547 1.5574 控制变量 总资产对数值 lnasset 2898 21.4999 1.3550 15.3539 27.3493 研发支出对数值 lnrd 2898 13.9736 1.7590 1.0276 22.4736 对外贸易规模对数值 lnex 2898 18.9047 2.2391 6.0768 25.4970 企业年龄对数值 lnage 2898 2.6786 0.4792 0 4.1589 员工平均工资对数值 lnw 2898 11.4639 0.4470 5.0024 14.6230 低技能劳动力平均工资对数值 lnwl 2579 11.3103 0.3706 9.2808 14.5969 高技能劳动力平均工资对数值 lnwz 2593 12.0318 1.0073 4.0420 16.9170 表 2 基准回归结果
变量 (1) (2) (3) lnl lnL lnZ bias −1.4049*** −1.7627*** 0.8172*** (−6.9676) (−8.4457) (3.6161) bias2 0.7269*** 0.8960*** −0.3965*** (6.5339) (7.8940) (−3.1793) lnasset 0.4659*** 0.4618*** 0.4841*** (14.8894) (12.9980) (12.0796) lnrd 0.1747*** 0.1454*** 0.2681*** (7.9052) (6.2958) (9.2889) lnex 0.0575*** 0.0567*** 0.0498*** (6.5991) (5.6252) (4.7187) lnage 0.2162 0.3078* 0.3931* (1.4663) (1.7456) (1.8095) lnw −0.3261*** (−7.1697) lnwl −0.6323*** (−12.1646) lnwz −0.1108*** (−6.7830) 企业固定效应 是 是 是 时间固定效应 是 是 是 样本量 2677 2393 2389 调整R2 0.9650 0.9643 0.9454 注:***、**、*分别代表1%、5%和10%的显著性水平;括号内为t值;下同。 表 3 稳健性和内生性检验
变量 (1) 替换代理变量 (2) GMM估计 (3) 固定效应交乘项 (4) 两阶段最小二乘法 lnl lnl lnl lnl bias −1.4057*** −1.4080*** −1.0299* (−6.9677) (−7.0141) (−1.9113) bias2 0.7271*** 0.7188*** 0.7307** (6.5290) (6.6581) (2.1892) bias1 −3.6546*** (−5.0627) bias12 4.1651*** (4.0293) 控制变量 是 是 是 是 企业固定效应 是 是 是 是 时间固定效应 是 是 是 是 行业×时间固定效应 是 不可识别检验 P值 0.0000 Anderson canon. corr. LM检验 P值 0.0000 Sargan检验 P值 0.3700 样本量 3224 2677 2649 1993 调整R2 0.9612 0.9650 0.9670 0.4111 表 4 生产率效应机制检验
变量 (1) (2) (3) (4) tfp_op lnl lnL lnZ bias 1.2258***
(5.7580)−1.3969***
(−6.6403)−2.1041***
(−8.8663)1.2109***
(5.1371)bias2 −0.6346***
(−5.5478)0.7473***
(6.4294)1.0876***
(8.4360)−0.5930***
(−4.5210)tfp_op 0.5034***
(2.2147)0.5585***
(2.3828)0.3908
(1.6124)tfp_op2 −0.0846***
(−2.6339)−0.0869***
(−2.6491)−0.0811**
(−2.3240)控制变量 是 是 是 是 企业固定效应 是 是 是 是 时间固定效应 是 是 是 是 样本量 2480 2480 2480 2480 调整R2 0.8766 0.9670 0.9632 0.9392 表 5 产出规模效应机制检验
变量 (1) (2) (3) (4) lnY lnl lnL lnZ bias −0.9613***
(−4.1447)−1.3037***
(−6.8099)−1.9663***
(−9.1925)1.1032***
(4.8348)bias2 0.4828***
(3.9239)0.6826***
(6.4555)0.9984***
(8.5376)−0.5447***
(−4.3883)lnY 1.4778***
(5.4713)1.4078***
(4.5786)1.1136***
(2.8167)lnY2 −0.0258***
(−4.1494)−0.0238***
(−3.3573)−0.0197**
(−2.1354)控制变量 是 是 是 是 企业固定效应 是 是 是 是 时间固定效应 是 是 是 是 样本量 2677 2677 2677 2677 调整R2 0.9618 0.9703 0.9670 0.9374