气候风险与企业智能制造转型

——来自大语言模型的经验证据

王乾坤　任慧　陈明生

（中国政法大学）

摘要：在全球气候风险加剧与“双碳”目标深入推进的背景下，向智能化、绿色化、融合化方向转型应对气候风险，正成为推动经济社会发展全面绿色转型、建设美丽中国的关键所在。基于2011—2023年沪深A股上市公司样本，本文使用大语言模型构建了企业层面气候风险指标，详细考察了气候风险与企业智能制造转型之间的因果关系。研究结果显示，气候风险驱动了企业智能制造转型。机制分析表明，气候风险通过推升环境成本催生转型动机、推升用工成本锁定转型方向以及促进资产脱虚向实提供转型保障三条路径驱动企业智能制造转型。异质性分析表明，气候风险对企业智能制造转型的驱动效应在沿海地区、资源型城市以及制造业行业中更为明显。本文的研究为理解气候风险背景下企业通过智能制造转型增强自身风险抵御能力与经济发展韧性提供了新的微观证据，对于协同推进降碳、减污、扩绿、增长具有重要的现实意义。

关键词：气候风险；智能制造；大语言模型；环境成本；用工成本；资产脱虚向实

作者简介：王乾坤，中国政法大学商学院讲师，北京，100088；任慧，中国政法大学商学院博士研究生，通信作者；陈明生，中国政法大学商学院教授、博士生导师。

基金项目：教育部人文社会科学研究规划基金项目“人工智能发展、机器人税与智能经济和传统经济的协调发展：以出租车行业为例”（25YJA790004）；中央高校基本科研业务费专项资金中国政法大学青年教师科研启动项目“智能经济新形态：代理式AI影响我国收入分配差距的效应、机制与优化路径研究”（10826310）

引用格式：王乾坤，任慧，陈明生. 气候风险与企业智能制造转型：来自大语言模型的经验证据[J]. 经济与管理研究，2026，47（9）：3-17.

Climate Risk and Corporate Intelligent Manufacturing Transformation

—Evidence from a Large Language Model

WANG Qiankun, REN Hui, CHEN Mingsheng

(China University of Political Science and Law, Beijing 100088)

Abstract: Against the backdrop of intensifying global climate risk and the deepening advancement of China's "dual carbon" goals, how to address climate risk by pursuing intelligent, green, and integrated development has become a pressing issue in promoting the comprehensive green transformation of the economy and society and in building a Beautiful China. Intelligent manufacturing (IM), with its unique strengths in green empowerment, becomes a new engine for firms to build competitive advantages in the context of creating new forms of intelligent economy. This paper innovatively introduces the semantic understanding capability of the FinBERT large language model into the field of textual analysis, constructs a firm-level climate risk index, and investigates the causal relationship between climate risk and IM transformation.

Based on panel data from A-share listed manufacturing firms in China from 2011 to 2023, the empirical findings indicate that climate risk can drive corporate IM transformation. Mechanism analysis reveals three key channels: increasing environmental costs to incentivize transformation, raising labor costs to define the transformation direction, and reallocating financial resources from virtual to real sectors to safeguard transformation. Heterogeneity analysis further shows that the driving effect is more pronounced in coastal areas, resource-based cities, and manufacturing industries. These conclusions provide new evidence on how firms enhance their risk resilience through IM transformation in the context of climate risk, and hold practical implications for the coordinated advancement of carbon emission reduction, pollution reduction, green development expansion, and economic growth.

The marginal contributions are threefold. First, this paper employs a large language model to construct a firm-level climate risk measure and innovatively integrates climate risk with corporate IM transformation within a unified framework, revealing firms' strategic responses to climate risk pressure through IM. Second, it systematically unpacks the internal logic through which climate risk drives corporate IM transformation, deepening conventional transmission channels such as environmental and labor costs, and uncovering a distinctive micro-level mechanism of the reallocation of financial resources from virtual to real sectors under climate risk conditions. Third, it comprehensively explores heterogeneous effects across spatial locations and industry attributes, expanding the research boundaries.

Based on these findings, this paper offers three policy recommendations. First, it is essential to improve the climate risk governance system and build an institutional framework for climate risk-driven IM transformation. Second, it is crucial to focus on the transmission channels through which climate risk drives IM transformation and implement coordinated policies on cost mitigation and transformation support. Third, efforts should be made to take full account of the heterogeneous characteristics of the climate risk-IM transformation relationship and to formulate targeted policies tailored to different regions and industries.

Keywords: climate risk; intelligent manufacturing; large language model; environmental cost; labor cost; financial resource from virtual to real sectors

下载全文