Investigating Corporate Environmental Risk Disclosure Using Machine Learning Algorithm

The volume of the environmental risk disclosure in the annual reports of firms in the pharmaceutical and chemical, tannery, telecommunications, and paper and printing industries listed on the Dhaka Stock Exchange (DSE) in Bangladesh was analyzed in this paper. The research used a content analysis of the annual reports of 43 companies that represented four DSE sectors. To quantify the level of environmental risk disclosure reporting practiced by corporations in their annual reports, the authors established the ERDIPCI for the pharmaceutical and chemical industry, the ERDITI for the tannery industry, the ERDITeI for the telecommunications industry, and the ERDIPPI for the paper and printing industry. Similarly, the machine learning clustering algorithm, k-means clustering, is used to cluster the companies based on the completion of different environmental indices. It is observed that from four sectors, the highest number of companies from the pharmaceutical and chemical industry disclosed environmental risk disclosures, and the lowest number of companies was from the tannery industry, followed by the telecommunications and the paper and printing industries. The enterprises differ significantly in their environmental risk disclosures, and the overall scenarios of the environmental reporting practices by companies in Bangladesh are quite poor. It also shows that among the 43 companies, a limited number of enterprises are placed first. The majority of the businesses are in the midst of a cluster that reflects the increasing order of indices fulfillment. This paper provided a few specific proposals to the relevant authorities in order to establish a regularity framework in which all the firms listed on the DSE in Bangladesh will be expected to address environmental risk disclosures and conservation actions in their annual reports towards adaptation to climate change and achieving environmental sustainability.