Modeling of the corporate capital structure

Abstract

The capital structure is determined by the ratio of the company’s own and borrowed funds. It has a direct impact on the financial stability of the company, profit and return on equity. Thus a reasonable calculation of the value of the specified ratio is an urgent problem for any large enterprise. The correct ratio of borrowed and own capital in the structure of the aggregate corporate capital will allow to provide the specified return from the use of own capital in changing market conditions. The methodology of optimizing the structure of capital proposed by the authors is based on a formula for increasing return on equity (ROE) as a function of three variables — return on sales (ROS), resource productivity (Y) and equity multiplier M. The interpretation of the DuPont model as a model with lagged variables is original, which allows to predict the change in the equity multiplier at the expense of possible increase in other variables: resource productivity and return on sales, provided that the return on equity will reach the target level. There have been obtained formulas which allow modeling the structure of capital that guarantees a given profitability. The heuristic principle based on the three-factor model of DuPont ROE = ROS×Y×M, was that keeping the return on equity at the same level (as compared with the reference period) should entail compensation for the multidirectional dynamics of return on sales, resource productivity and equity multiplier. The paper considers three variants of the dependence of changes in the ratio of equity and borrowed capital of the company: when resource productivity is unchanged (ΔY = 0), when return on sales is unchanged (ΔROS = 0) and when both indicators change simultaneously (ΔY ≠ 0 and ΔROS ≠ 0). The methodology presented by the authors also allows to calculate the structure of capital which would provide a given level of return on equity ROE based on the forecast of return on assets ROS and return on sales Y with reference to the values of ROS and Y calculated in the reference period. We present the testing of this algorithm (in all three variants) on the basis of the reporting data of one of the metallurgical companies of Russia for the period 2017—2019.