Developing the stress testing algorithms for companies: A metal industry case study

Abstract

Subject. The article addresses the development of stress testing algorithms for companies operating in the real economy. Objectives. The study generalizes stress-testing algorithms and assesses the hypothesis about the resistance of steel companies to external shocks. Methods. I employ general scientific principles and methods of research, like abstraction, generalization of approaches of domestic and foreign authors to stress testing and industry analysis. Results. I developed methods of aggregated direct single-factor historical stress testing based on the top-down approach. The methodology includes three stages. First, I identify stress-testing parameters: stress factors, variables and constant model parameters. Revenue, which is influenced by market risk, is defined as a stress factor. At the second stage, I substantiate the algorithms for predicting stress-testing parameters, then I calculate the predictive values of variable and permanent parameters of the model, assess the volatility of the stress factor, calculate confidence intervals and justify stress-testing scenarios. At the final stage, the results of stress testing are analyzed. The methodology was tested on the investigation of data of 454 metallurgical enterprises. Conclusions. The companies assigned to the first group (20% of companies that account for 80% of revenue) have a high degree of resilience in the short term, owing to their operational efficiency, which ensures continuity of operations. In the long term, iron and steel companies have insufficient resilience due to the high loan debt burden offsetting the aggressive dividend policy, and inadequate investment in production capacity and safety.