【作者】 徐杰；

【导师】 陆菁；

【作者基本信息】 浙江大学 ， 世界经济， 2025， 博士

【摘要】 新一轮科技革命和产业变革是智能革命,正在推动着世界工业前沿进入智能制造时代。世界主要工业国家竞相将智能制造视为振兴制造业和抢占先进制造业高地的重要抓手,纷纷出台智能制造发展战略。全球产业竞争格局正发生重大调整,中国现有的国际产业分工地位面临着来自发达国家和部分发展中国家的双向挤压与竞争。面对内外部环境发生的深刻变化,中国政府将智能制造作为建设制造强国的主攻方向,构建中国特色智能制造发展战略,与老牌工业强国一齐抢占智能制造竞争制高点。在中国政府的引导、支持和推动下,中国制造企业正有序地、梯度地向智能制造转型。在向智能制造转型的过程中,中国制造企业普遍面临着传统生产要素配置与新型生产方式不匹配的结构性挑战,而在生产全球化的趋势下,这种挑战进一步扩展至离岸生产要素配置层面。如何调整在岸与离岸生产要素配置,以在智能制造场景中实现成本最小化或利润最大化,是制造企业需要攻克的首要难题。然而,智能制造这一新兴研究领域目前并未得到经济学界的重视,呈现出一种“政府引导研究”的特殊现象。仅有少数研究对智能制造的经济效应进行了实证研究,且尚未有研究系统性地探讨智能制造如何影响中国制造企业在岸与离岸生产要素配置。在此背景下,本研究首先基于对世界主要工业国家和中国智能制造发展战略的横向对比与纵向分析,以中国智能制造发展战略为参照标准,从智能制造的定义、形态、模式及企业向智能制造转型的关键技术等方面全面地、系统地界定智能制造的内涵,为理论建模与实证检验提供概念支撑。在此基础上,本研究将智能制造建模为内含多重方向技术变革的生产方式,结合基于任务的模型设定,构建研究智能制造如何影响企业生产要素配置的任务模型,并对理论模型进行特定拓展,从而将智能制造如何影响企业在岸与离岸生产要素配置这一研究主题细分为三个子研究:第一,智能制造对企业劳动力需求的影响研究;第二,智能制造对企业技能相对需求的影响研究;第三,智能制造对企业离岸生产的影响研究。在理论研究部分,本研究主要推导和讨论智能制造的多重方向技术变革对企业在岸与离岸生产要素配置的联合作用机制。在实证检验部分,本研究分别从三个子研究入手,实证分析智能制造对中国制造企业劳动力需求、技能相对需求、离岸生产的影响效应和具体作用机制。本文的理论研究主要得出以下结论:第一,在研究智能制造影响企业劳动力需求的基准理论模型中,本研究将企业实现智能制造所需的技术组合分为三类不同方向的技术变革:以自动化为方向的技术变革、以创造新任务为方向的技术变革以及以资本深化为方向的技术变革,并分别从三个技术变革的方向分析智能制造对企业劳动力需求的影响,发现智能制造对劳动力需求的影响存在三种机制:一是“生产率效应”。任何提高生产率和扩大产出的技术变革方向都会增加对劳动力的需求;二是“替代效应”。在广延边际上,智能制造直接改变任务在生产要素之间的分配,推动资本完成原本分配给劳动力的部分任务,从而减少对劳动力的需求;三是“创造效应”。在广延边际上,智能制造直接扩展企业可行性任务集合,允许企业开发和创造出新的劳动密集型任务,从而增加对劳动力的需求。第二,在研究智能制造影响企业技能相对需求的扩展理论模型中,本研究仍然将企业实现智能制造所需的技术组合分为三类不同方向的技术变革:以自动化为方向的技术变革、以创造新任务为方向的技术变革以及以资本深化为方向的技术变革,并分别从三个技术变革的方向分析智能制造对企业技能相对需求的影响,发现智能制造在任务集合的广延边际上通过“替代效应”和“创造效应”提高企业对技能的相对需求,在集约边际上通过“资本深化效应”提高企业对技能的相对需求。第三,在研究智能制造影响企业离岸生产的扩展理论模型中,本研究进一步引入以提升离岸生产能力为方向的技术变革设定,分析智能制造对企业离岸生产的影响,发现智能制造对企业离岸生产的影响存在三种机制:一是“回流效应”。在任务集合的广延边际上,智能制造提升资本在特定任务中的生产率,推动资本替代原本分配给离岸产出的部分任务,导致这些任务回归国内生产,从而减少企业的离岸生产。二是“离岸效应”。在任务集合的广延边际上,智能制造提升企业离岸生产的能力,推动离岸产出替代原本分配给国内劳动力的部分任务,导致这些任务离岸生产。三是“生产率效应”。在任务集合的集约边际上,智能制造降低企业生产单位成本,提高企业整体产出,从而间接增加企业的离岸生产。本研究在实证研究中,基于中国制造业上市企业数据库,分别就智能制造对中国制造企业劳动力需求、技能相对需求以及离岸生产的影响及具体作用机制展开实证检验。本文的实证研究主要得出以下发现:第一,在智能制造对中国制造企业劳动力需求的影响及作用机制的实证检验中,本研究发现企业应用智能制造增加了企业劳动力需求。进一步考虑由变量测度偏误、双向因果以及遗漏变量偏误等情形导致的可能存在的内生性问题后,研究结论依然成立。机制检验结果表明,在智能制造对企业劳动力需求的直接影响机制中,智能制造的“创造效应”主导了“替代效应”,智能制造还通过“生产率效应”这一间接影响机制增加了企业对劳动力的雇佣。异质性分析结果表明,智能制造对资本密集度较低、所在地区人力资源供应条件较差和最低工资较低的企业的劳动力雇佣提升效应更为显著。进一步分析发现,智能制造显著提高中国制造企业的超额雇员率,表明当前企业引入智能制造很可能导致企业的实际雇员数超出其生产所需的最优要素配置。第二,在智能制造对中国制造企业技能相对需求的影响及作用机制的实证检验中,本研究发现智能制造提升了企业技能相对需求。在针对平行趋势假设、无预期效应假设、异质性处理效应、内生性问题、变量测度敏感性、样本选择敏感性等方面进行一系列检验后,这种因果关系仍然是稳健可信的。机制检验结果表明,在企业生产任务集合的广延边际上,智能制造通过“替代效应”和“创造效应”提高企业技能相对需求;而在集约边际上,智能制造通过“资本深化效应”提高企业技能相对需求。异质性分析结果表明,智能制造对非国有企业、非专利密集型企业和资本密集型企业技能需求的提升效应更为显著。此外,本研究还发现实施智能制造的中国国有企业可能承受着来自地方政府保护就业的刚性约束,致使其劳动力雇佣结构僵化,无法在国有企业样本中观测到智能制造对非技能劳动力的“替代效应”以及智能制造对企业技能相对需求的正向影响。第三,在智能制造对中国制造企业离岸生产的影响及作用机制的实证检验中,本研究发现智能制造对中国制造企业离岸生产存在显著且持久的积极影响。这一核心发现在一系列稳健性检验下变得更为可信。然而,智能制造的影响在各处理队列中存在显著的异质性,而这种异质性与企业资本密集度密切相关。机制检验结果表明,智能制造同时对企业离岸生产存在着回流的拉力和离岸的推力,对中国的企业而言,智能制造的“回流效应”很可能强于“离岸效应”,但与此同时智能制造的“生产率效应”非常强大以至于在总体结果上掩盖了“回流效应”的存在。异质性分析结果表明,智能制造的积极效应在中国的非国有企业和资本密集型企业中更为显著。进一步基于企业离岸的东道国的不同维度研究发现,智能制造同步增加了中国制造企业在发展中国家与发达国家的海外生产型子公司布局,并显著增加了企业向一带一路沿线国家的离岸生产布局。基于企业广义离岸活动的不同维度研究发现,智能制造在促进企业离岸贸易活动的同时还引发了企业离岸研发活动的逆势回流。更多还原

【Abstract】 The new wave of technological and industrial revolution is an intelligent revolution,propelling the global industrial frontier into the era of intelligent manufacturing.Major industrial nations are vying to establish intelligent manufacturing as a pivotal strategy for revitalizing their manufacturing sectors and seizing the high ground in advanced manufacturing,leading to the widespread adoption of national intelligent manufacturing development strategies.The global competitive landscape of industries is undergoing significant realignment,with China’s current position in the international division facing dual pressures and competition from both developed countries and certain developing nations.In response to these profound internal and external changes,the Chinese government has identified intelligent manufacturing as the central focus of its strategy to build a manufacturing powerhouse,developing an intelligent manufacturing development framework with Chinese characteristics to compete with established industrial powers for supremacy in intelligent manufacturing.Under the guidance,support,and promotion of the Chinese government,Chinese manufacturing enterprises are undergoing a structured and tiered transition towards intelligent manufacturing.During this transformation,these enterprises commonly encounter structural challenges stemming from a misalignment between traditional factor allocation and new production paradigms—a challenge further compounded at the offshore factor allocation level by the trend of production globalization.Determining how to adjust onshore and offshore factor allocations to minimize costs or maximize profits within intelligent manufacturing scenarios represents a primary obstacle that manufacturing firms must overcome.However,this emerging field of research has yet to gain significant traction within the economics discipline,presenting a peculiar phenomenon of“government-guided research”.Only a limited number of studies have empirically examined the economic effects of intelligent manufacturing,and none have systematically investigated how intelligent manufacturing influences the onshore and offshore factor allocations of Chinese manufacturing enterprises.Against this backdrop,this study first conducts the horizontal comparative analysis and vertical chronological analysis of intelligent manufacturing strategies across major industrial nations and China.Using China’s intelligent manufacturing development strategy as a benchmark,it comprehensively and systematically defines the conceptual framework of intelligent manufacturing—encompassing its definition,forms,models,and key technologies for enterprise transformation—thereby establishing the conceptual foundation for theoretical modeling and empirical testing.Building upon this framework,the study models intelligent manufacturing as a production paradigm embedded with multi-directional technological change.By incorporating a task-based model framework,it constructs a theoretical task model to analyze how intelligent manufacturing influences enterprise factor allocation.Specific extensions of this model break down the central research theme—how intelligent manufacturing affects firms’onshore and offshore factor allocation—into three subsidiary studies:first,its impact on enterprise labor demand;second,its impact on the relative demand for skills;and third,its impact on offshoring production.The theoretical component primarily deduces and discusses the joint mechanisms through which intelligent manufacturing’s multi-directional technological change affects onshore and offshore factor allocation.The empirical segment separately analyzes the effects and specific mechanisms of intelligent manufacturing on Chinese manufacturing firms’labor demand,relative demand for skills,and offshoring production,corresponding to the three subsidiary studies.The theoretical component of this study yields the following principal conclusions:First,within the benchmark theoretical model examining the impact of intelligent manufacturing on firm’s labor demand,the requisite technological portfolio for its implementation is categorized into three distinct types of directional technological change:automation-oriented,new task creation-oriented,and capital deepening-oriented.Analyzing the influence of intelligent manufacturing on labor demand through these lenses reveals three operative mechanisms:(1)The“Productivity Effect”:any technological change that enhances productivity and expands output subsequently increases the demand for labor;(2)The“Displacement Effect”:on the extensive margin,intelligent manufacturing directly alters the assignment of tasks between production factors,enabling capital to perform tasks previously allocated to labor,thereby reducing labor demand;(3)The“Creation Effect”:on the extensive margin,intelligent manufacturing directly expands the firm’s feasible set of tasks,allowing for the development and creation of new labor-intensive tasks,thereby increasing labor demand.Second,within the extended theoretical model investigating the effect of intelligent manufacturing on the relative demand for skills,the technological portfolio is again decomposed into the same three types of directional change.The analysis demonstrates that intelligent manufacturing elevates the relative demand for skills through two channels on the extensive margin—the“Displacement Effect”and the“Creation Effect”—and through a“Capital Deepening Effect”on the intensive margin.Third,within the extended theoretical model analyzing the impact of intelligent manufacturing on offshoring production,a further dimension of technological change—aimed at enhancing offshoring production capabilities—is introduced.The investigation identifies three mechanisms:(1)The“Reshoring Effect”:on the extensive margin,intelligent manufacturing increases the productivity of capital in specific tasks,prompting the reallocation of tasks previously assigned to offshoring production back to domestic capital,thereby reducing offshoring activity;(2)The“Offshoring Effect”:on the extensive margin,intelligent manufacturing enhances the firm’s capability for offshoring production,enabling offshore output to substitute for tasks previously performed by domestic labor,thereby promoting offshoring;(3)The“Productivity Effect”:on the intensive margin,intelligent manufacturing lowers the firm’s unit production cost,increases its overall output,and thus indirectly boosts its offshoring production.In the empirical segment of this research,utilizing a database of Chinese manufacturing listed enterprises,this study conducts empirical tests on the influence of intelligent manufacturing on firm’s labor demand,relative demand for skills,and offshoring production within these firms,alongside investigating the specific underlying mechanisms.The primary empirical findings are as follows:First,regarding the effect and mechanisms of intelligent manufacturing on firm’s labor demand,the study finds that the adoption of intelligent manufacturing increases enterprise labor demand.This conclusion remains robust after addressing potential endogeneity concerns arising from measurement error,reverse causality,and omitted variable bias.Mechanism tests indicate that the direct impact is dominated by the“creation effect”of intelligent manufacturing over its“displacement effect”.Additionally,intelligent manufacturing indirectly boosts labor hiring through the“productivity effect”.Heterogeneity analysis reveals that the labor-enhancing effect of intelligent manufacturing is more pronounced in firms with lower capital intensity,those located in regions with poorer human resource supply conditions,and those facing lower minimum wages.Further analysis shows that intelligent manufacturing significantly increases firms’rate of excess employment,suggesting that its current adoption likely leads firms to employ more workers than the optimal level required for production under efficient factor allocation.Second,concerning the impact on the relative demand for skills,the study finds that intelligent manufacturing elevates the relative demand for skills.This causal relationship proves robust across a series of tests addressing parallel trends,the absence of anticipatory effects,heterogeneous treatment effects,endogeneity,measurement sensitivity,and sample selection sensitivity.Mechanism tests confirm that on the extensive margin of the task set,intelligent manufacturing raises the relative skill demand through both the“displacement effect”and the“creation effect”,while on the intensive margin,it does so via the“capital deepening effect”.Heterogeneity analysis indicates that the positive effect on skill demand is more substantial in non-state-owned enterprises(non-SOEs),non-patent-intensive firms,and capital-intensive enterprises.Furthermore,the study finds that Chinese state-owned enterprises(SOEs)implementing intelligent manufacturing likely face rigid employment protection constraints imposed by local governments,resulting in an inflexible employment structure that prevents the observation of the“displacement effect”on unskilled labor and the overall positive impact on the relative demand for skills within SOEs sample.Third,regarding the impact on offshoring production,the study identifies a significant and persistent positive effect of intelligent manufacturing on offshoring production activities of Chinese manufacturing firms.This core finding gains greater credibility through a battery of robustness checks.However,the impact exhibits significant heterogeneity across treatment cohorts,closely correlated with firm capital intensity.Mechanism tests reveal that intelligent manufacturing simultaneously exerts a pulling force(reshoring effect)and a pushing force(offshoring effect)on offshore decisions.For Chinese firms,the“reshoring effect”is likely stronger than the“offshoring effect”,yet the powerful“productivity effect”masks the presence of the“reshoring effect”in the aggregate outcome.Heterogeneity analysis shows that the positive effect of intelligent manufacturing is more significant in non-SOEs and capital-intensive enterprises.Further investigation based on host country dimensions demonstrates that intelligent manufacturing synchronously increases the establishment of overseas production subsidiaries in both developing and developed countries,with a notable increase in offshore production layout directed towards countries along the Belt and Road initiative.Analysis based on different dimensions of generalized offshore activities reveals that while intelligent manufacturing promotes offshoring trade activities,it also triggers a counter-trend reshoring of offshoring R&D activities.更多还原