인과 포레스트를 이용한 벤처기업 사회적 활동의 매출액에 대한 효과 분석

Abstract

The impact of corporate's social activities on sustainable management is increasing, and this is not just for large companies, but also expanding to small and medium-sized companies and venture companies. Accordingly, in this paper, the effect of the social activities of venture companies on sales was empirically analyzed through the Causal Forest method. Causal Forest is a machine learning model that allows the effect (treatment effect) of the variable of interest to be tested through causal reasoning after controlling several other characteristic variables that affect the objective variable through random forest modeling. As a result of empirical analysis using detailed survey data of venture companies from 2015 to 2020, it was confirmed that the social activities of venture companies had a significant average treatment effect (ATE) on sales. In addition, it was confirmed that the difference in the individual conditional average treatment effect (CATE) on the sales of social activities was significant depending on the size of the company's debt and the degree of investment activation. At this time, the treatment effect of companies with low total debt and low investment activation was widely distributed, which can be attributed to the characteristics of venture companies that focus on investment and technological innovation and show various corporate forms.