ABSTRACT
The research focuses on the value-relevance problem and the study of the relationship between information from financial reports and the value of companies. The research, which is the continuation of previous work, is based on the audited annual financial statements of companies listed on the Warsaw Stock Exchange (WSE). In the first study by Kubik-Kwiatkowska, the analysis was performed by means of the panel regression model. In the current study, the authors verify previous results based on machine-learning techniques and financial reports from 2011 to 2020. Both studies identified a positive relationship between the prices of shares and accounting information, such as: equity, profit, tax, the size of the company, and the industry (sector of activity). In addition, both studies produced a novel observation, where the tax amount is related to the value of companies. Although the models considered only financial statements, i.e., past information about the companies’ performance, it was possible to create a theoretical investment strategy, which led to abnormal returns. Similar results of the correlation between model estimates and actual values of companies were obtained during the second study. This evidence shows that value-relevance models can be considered as one of the tools in investment strategy building.
