Utilization of multidimensional methods for corporate sustainability

Abstract

The article is focused on the use of multivariate methods for creating the Corporate Sustainability Index (CSI) predictive model for measuring sustainability of industrial companies according to CZ-NACE, and on comparing these methods. The goal of this article is to propose a suitable CSI predictive model and to determine which financial and non-financial indicators can most influence a company tending to sustainability. To determine the CSILA predictive model, the Logistic Regression is used, and to determine the CSIMDA model, the method of Multiple Discriminant Analysis. However, based on the theoretical analysis of each method it is necessary to state that the methods cannot be unequivocally compared, even though each of these methods identified similar significant financial and non-financial indicators, coefficients and tests, which are interpreted analogically in the methods and are created in different ways. The results of the comparison of the methods for determining the predictive model CSI show that the logistic regression seems to be the best, which has a high percentage of correctly classified companies based on the calculated probability; in this case, the Gini index is also highest. The resulting classification of companies into different groups in comparing these two methods underwent significant changes as opposed to the classification of the companies according to the Data Envelopment Analysis method. The conclusions of the research of measuring the sustainability of the company show that currently-in addition to financial indicators-also non-financial indicators must be included in predictive models, namely the environmental indicator, the social indicator and the corporate governance indicator. It means that for the companies it has become a necessity to build a unified system for measuring sustainability of a company; this requirement has been confirmed also by managers of companies.