Econ Papers Daily / 期刊论文
人工智能试点区与货币政策不确定性下的企业投资效率：来自中国的证据
AI pilot zones and corporate investment efficiency under monetary policy uncertainty: evidence from China
摘要
This paper examines whether a place-based artificial intelligence policy improves corporate investment efficiency and whether its effect varies with monetary policy uncertainty and firms’ financial vulnerability. Exploiting the phased rollout of China’s National New Generation Artificial Intelligence Innovation and Development Pilot Zones, we employ a staggered difference-in-differences approach for a panel of Chinese A-share listed firms from 2015 to 2023. We find that the AI pilot-zone policy significantly reduces investment inefficiency, and the reduction is more pronounced among firms with greater short-term refinancing exposure or tighter financing constraints when monetary policy uncertainty is elevated. Using 6.93 million online job postings classified with a fine-tuned MacBERT model, we further show that the policy shifts the composition of AI recruitment towards finance-related vacancies, while overall postings decline. These findings extend evidence on place-based AI policy to corporate capital allocation and highlight the role of financial information-processing capacity in shaping investment efficiency under monetary policy uncertainty.
中文摘要
本文考察了基于地点的人工智能政策是否能够提升企业投资效率，以及其效果是否随货币政策不确定性和企业财务脆弱性而变化。本文利用中国国家新一代人工智能创新发展试验区的分阶段推广，采用交错双重差分方法，对2015年至2023年中国A股上市公司面板数据进行了分析。研究发现，人工智能试验区政策显著降低了投资非效率，且在货币政策不确定性上升时，对于短期再融资敞口更大或融资约束更紧的企业，这种降低效应更为明显。本文进一步使用经微调的MacBERT模型对693万条在线招聘信息进行分类，表明该政策使人工智能招聘结构转向与金融相关的岗位，而总体招聘数量下降。这些发现将基于地点的人工智能政策的证据拓展至企业资本配置领域，并凸显了金融信息处理能力在货币政策不确定性下塑造投资效率中的作用。