Frontiers in Environmental Science (Oct 2023)

Application of gradient boosting model to forecast corporate green innovation performance

  • Jingyi Zhang,
  • Kedong Yin,
  • Kedong Yin

DOI
https://doi.org/10.3389/fenvs.2023.1252271
Journal volume & issue
Vol. 11

Abstract

Read online

Corporate green innovation performance can serve as a critical tool for policymakers to identify the best practice and provide support to micro-entities in need. Accurate forecasting of corporate green innovation performance plays a vital role in innovation incentives by simulating the effects of regulations and strategies. Based on the data of China’s A-share listed companies during 2010–2020, this paper elaborates the gradient boosting algorithm to predict the corporate green innovation performance and compares the prediction results of the gradient boosting model with the linear model, the decision tree model, and the random forest model. Subsequently, it examines the effectiveness of the influencing factors related to the enterprise’s internal driving mechanism and external policy pressure in promoting corporate green innovation performance. It finds that: 1) The gradient boosting model outperforms other methods in its predictive effect. 2) An enterprise’s resource base is a critical factor influencing its green innovation activities, and in particular, the influence of financial indicators on corporate green innovation performance has a significant incentive effect, indicating that the impetus from enterprises’ internal driving mechanism is crucial for enterprises’ green transformation. 3) The effect of secondary indicators is heterogeneous. In the command-based environmental regulation tools, the administrative penalties can activate enterprises’ green innovation better than the approvals of Environmental Impact Assessment (EIA) documents for construction projects do; as for the incentive-based environmental regulation, investment in pollution control projects has an apparent inducing effect on the corporate green innovation performance, while the environmental tax presents an inverted U-shape, implying that overly stringent taxation crowds out the corporate green innovation performance. 4) Similarly, in the operating capacity indicators, the increasing operating income growth rate can trigger the improvement of green innovation performance; nevertheless, the total asset turnover ratio shows a suppressing effect. The key to promoting corporate green innovation performance lies in effectively regulating the enterprises’ internal driving mechanism and the rational choice of external policy tools. This study helps to prospectively identify how corporate green innovation performance changes and provides theoretical guidance and micro evidence for the policymakers on choosing environmental regulation tools and for enterprises on adjusting the resource bases.

Keywords