Studies in Science of Science ›› 2025, Vol. 43 ›› Issue (4): 864-875.

Previous Articles     Next Articles

Antitrust and the entry of technology start-ups: Causal inference based on double machine learning

  

  • Received:2024-02-05 Revised:2024-05-13 Online:2025-04-15 Published:2025-04-15

反垄断法与科技企业进入———基于双重机器学习的因果推断

蒋墨冰,李莹,徐晓慧   

  1. 浙江理工大学
  • 通讯作者: 徐晓慧
  • 基金资助:
    国家社会科学基金

Abstract: Technological innovation is a crucial support for achieving high-quality economic development, with technology enterprises being the main agents of such innovation. Encouraging the entry of technology start-ups is a vital measure to stimulate technological innovation and complete the transformation of the economic dynamics from old to new energy. Therefore, in the current situation where external technological blockades are becoming increasingly severe, how to promote the entry of technology start-ups has become key to China’s breakthrough in overcoming the critical core technology “bottleneck” dilemma. Based on the matching data of the national industrial and commercial enterprise registration information database and the city level data from 2004 to 2020, this paper takes the implementation of Antitrust Law in 2008 as a quasi-natural experiment, and uses the double machine learning model to identify the impact of antitrust on the entry of technology start-ups and its internal mechanism. This paper finds that the implementation of Antitrust Law significantly promotes the entry of technology start-ups, and the above promotion effect is more significant in east cities, developed cities and cities with better transportation infrastructure. Mechanism analyses show that Antitrust Law promotes the entry of technology start-ups by improving the innovation ecosystem, optimizing the business environment and boosting the level of venture capital. Further analyses show that the entry of technology start-ups brought about by the implementation of Antitrust Law can drive high-quality economic development in cities. The above conclusions can provide enlightenment for policy guidance on cultivating technology start-ups and achieving high-quality economic development. The marginal contributions of this paper are as follows: First, this study enriches the research on the microeconomic consequences of the Antitrust Law. The economic consequences of competition policy have always been a hot topic in academia. Existing literature has mostly focused on the impact of the Antitrust Law on internal business decisions of enterprises, while research on enterprise dynamics such as the entry of technology start-ups is relatively scarce. Second, this paper employs cutting-edge methods to enhance the effectiveness of policy evaluation. Existing literature often uses parametric methods to assess policy effects, inevitably facing the “curse of dimensionality” and model specification bias issues. This paper leverages the advantages of machine learning algorithms in high-dimensional, non-parametric prediction, using a double machine learning method for causal inference. This approach not only better mitigates endogeneity issues but also overcomes the regularization bias of machine learning methods, thereby more accurately assessing the microeconomic effects of the Antitrust Law. Third, this paper deeply analyzes the impact mechanism of the Antitrust Law on the entry of technology start-ups. Specifically, this paper reveals the internal mechanisms by which the Antitrust Law affects the entry of technology start-ups from three aspects: improving the innovation ecosystem, optimizing the business environment, and enhancing the level of venture capital, which helps to deeply understand the channels through which the Antitrust Law affects the dynamics of technology start-up entry.

摘要: 科技企业是经济高质量发展的重要推动力。本文基于2004-2020年全国工商企业注册信息数据库与城市层面匹配数据,以2008年《反垄断法》实施为准自然实验,使用双重机器学习模型识别反垄断对科技企业进入的影响及内在机制。研究发现:《反垄断法》的实施显著促进了城市科技企业进入,且上述促进效应在城市等级高、东部以及交通基础设施水平高的样本中更为显著。机制分析表明,《反垄断法》通过改善创新生态、优化营商环境以及提升风险投资水平促进科技企业进入。进一步分析发现,《反垄断法》实施带来的科技企业进入能够推动城市经济高质量发展。上述结论可为城市科技企业培引和经济高质量发展的政策指导提供启示。