ESG 성과가 기업가치에 미치는 영향의 장기 추세 분석 및 머신러닝 모델의 예측 성능 비교

Through the analysis of time series data, this study examines how Environmental, Social, and Governance (ESG) performance affects corporate value. It also evaluates the predictive power of machine learning models. In order to do this, financial and ESG performance data from 2011 to 2023 were gathered, and regression analysis was carried out using a three-year rolling window dataset. The impact of environmental performance on corporate value was initially statistically insignificant. There is a negative correlation between environmental performance and corporate value because, during the COVID-19 pandemic, investments in environmental initiatives were perceived as harmful to profitability. On the other hand, before the pandemic, social performance had a positive impact on corporate value, though this effect progressively diminished over time. However, the importance of social responsibility was reaffirmed after the pandemic, which resulted in a brief revival of the influence of social performance. Additionally, governance’s impact on corporate value was initially negligible. However, as the pandemic spread, governance’s role as a risk management tool gained more recognition, which increased its impact on corporate value. Furthermore, the study discovered that CatBoost model outperformed conventional linear regression models in terms of predicting corporate value based on ESG performance, and even higher predictive efficacy was noted when using the three-year rolling window data. By clarifying the dynamic relationships between ESG factors and corporate value and the predictive capabilities of machine learning methodologies, this research aids in the development of ESG strategies and improves corporate value forecasting.