人工智能会改变制造企业的成本粘性吗?

Willartificial intelligence change the cost stickiness of manufacturing companies?

以2013—2019年沪深A股上市制造企业为研究样本,探讨人工智能对制造企业成本粘性的影响及作用机制。研究发现,人工智能对成本粘性有降低效应,在资产专用性较高、环境不确定性较大的制造企业中降低效应更为显著;人工智能通过调整成本路径、管理层乐观预期路径降低成本粘性,而未通过代理冲突路径降低成本粘性;在资产专用性弱、环境不确定性低的制造企业中,人工智能提高了成本粘性,但此情形只出现在内部控制质量较低的样本中;人工智能对成本粘性的降低效应具有一定的持续性,对物质资源成本粘性的降低效应更显著。提出提高企业管理者对成本粘性的认识、制造企业要正确选择和应用人工智能、政府应为人工智能应用提供必要的支持等建议,以期为制造企业发展、政策制定及相关研究提供参考。

Will artificial intelligence change the cost stickiness of manufacturing companies?This question is much relevantagainst the background where artificial intelligence has become a key element and an important driver of high-quality development of manufacturing firms.This paper explores the impact of artificial intelligence on the cost stickiness of manufacturing firms and its mechanism of action,using the A-shares listed manufacturing firms in Shanghai and Shenzhen from 2013 to 2019 as samples.The study found that artificial intelligence has a reducing effect on cost stickiness,and the reducing effect plays a more significant role in manufacturing firms with higher asset specificity and environmental uncertainty.Artificial intelligencereduces cost stickiness by adjusting the cost path and the management’s optimistic expectation path,but not by the agency conflict path.In manufacturing firms with weak asset specificity and low environmental uncertainty,artificial intelligence improves cost stickiness,but this only occurs in the sample with low quality of internal control.The reducing effect of artificial intelligenceon cost stickiness features a certain level of persistence,with stronger impact on material resource cost stickiness in particular.The paper provides suggestions at the end regarding improved awareness of cost stickiness among enterprise managers,manufacturing enterprises’decisionto select and apply rightartificial intelligence,and necessary government support for artificial intelligence applications,with the aim to enhance growth of manufacturing enterprises,policy formulation,and relevant research.