Can Automation Lead to Productivity Growth and Divergence among Firms:A Task-Based Perspective on Differential Substitution
In the development of modem industry,automation has been widely applied across industries and is becoming a core tool for firms to reshape their competitive advantages.Driven by the Fourth Industrial Revolution,traditional automation technologies are deeply integrated with artificial intelligence(AI),enhancing firm productivity and fueling modern economic growth.However,the productivity gains from automation are not evenly distributed across all firms and may alter the existing productivity distribution.This study explores the relationship between automation and firm productivity growth and divergence from both theoretical and empirical perspectives,highlighting the need for automation to more equitably and comprehensively enhance firm productivity to maintain market competition and prevent excessive productivity disparities between firms.This study innovatively constructs a theoretical framework from a task-based differential substitution perspective to analyze the impact of automation on firm productivity and its divergence.It explains how automation drives productivity growth through resource management,flexible production,and product innovation.Furthermore,it argues why high-productivity firms,with their significant advantages in production scale,operating profits,and data assets,are more capable of adopting automation technologies to substitute labor in routine and manual tasks,thus achieving greater productivity gains.In the empirical analysis,deviating from conventional automation indicators,this study constructs an automation dictionary based on patent information,identifying automation-related patent classification codes,and measuring firm-level automation.Utilizing data from Chinese listed manufacturing firms from 2016 to 2022,this study empirically examines the impact of automation on firm productivity and its divergence.The findings indicate that automation significantly enhances firm productivity.The productivity-enhancing effect of automation is stronger for high-productivity firms,thereby widening the productivity gap and leading to divergence among firms.These results hold robust across various robustness and endogeneity tests.The mechanism analysis reveals that the productivity divergence effect of automation is primarily achieved through labor substitution in routine and manual tasks.The substitution effect among routine tasks dominates in the conventional automation phase,while the substitution effect among manual tasks becomes more prominent in the intelligent automation phase.Further analysis indicates that the"superstar firm"effect exists.In the dynamic perspective,the positive impact of automation on high-productivity firms diminishes over time,while the negative impact on low-productivity firms becomes more pronounced.Additionally,non-automation patents contribute to productivity convergence among firms.The policy recommendations of this paper include preventing the excessive Matthew effect of productivity that suppresses market competition,promoting an equitable distribution of technological benefits across different labor task categories,and tailoring automation support strategies to different types of firms.This paper offers a significant contribution to existing literature by employing patent data to construct novel enterprise-level automation indicators,investigating the economic effects of automation through the lens of differential substitution across various work tasks,and providing an explanation for the persistence and expansion of productivity divergence in the context of Industry 4.0.
NETL
NSTL
万方数据
