A Study on the Impact of Green Patent Data on ESG Environment Indicators

Abstract

Global investment and financial institutions such as BlackRock are exerting pressure by stating that they will no longer invest in companies that do not adhere to ESG management criteria while operating financial products related to ESG. As a result, ESG is recognized as a management activity that companies must necessarily perform. Various evaluation agencies publish ESG indicators to measure and manage such ESG activities. However, companies assessing ESG levels only disclose the comprehensive evaluation results encompassing the three areas of environment, society, and corporate governance, without revealing specific measurement methods. Detailed information is shared only when the evaluated company requests consulting, limiting the ability of the general public or unmeasured companies to confirm and compare ESG management levels. To address these limitations, this study aims to investigate the relationship between publicly available data and ESG indicators. The data used in the research are patent data representing a company’s technological development, secured from the Google Cloud Platform’s BigQuery. ESG indicator data utilized the Dow Jones Sustainability Indices (DJSI) from S&P Global, which received a high-quality assessment from the Sustainability Institute. The research model employed deep learning-based natural language processing technologies, utilizing Long-Short Term Memory (LSTM), Attention, and Transformer models.