Modeling the Persistency and Reversibility of Accounting Earnings Using Markov Chains

Objective: One of the most important accounting information is the earnings and its components, which is probably part of the information for investors' decision making. Changes in the way earnings are reported in the evolution of accounting show that earnings and their components have been important to the financial reporting audience, and there is little doubt that earnings are an important output of the accounting system. One of the most important features of accounting earnings is its persistency, which is considered by users when making economic decisions. Earnings persistency means the continuation of past earnings in future periods. By knowing the nature of accounting earnings persistency, we can better predict the financial aspects of the company that are associated with this figure. Thus, the purpose of this study is to investigate the persistency and reversibility of accounting earnings status using Markov processes. Methods: In order to investigate the persistency and reversibility of the accounting earnings situation, the mathematical methods of Markov processes have been used using the data of 308 companies listed on the stock exchange in the period 1393 to 1397. In this research, in order to test the hypotheses, first, the transfer probability matrix is formed using the items of net income to sales ratio and operating profit to sales ratio, and the hypotheses are tested with the help of concepts in the field of Markov chains. For this purpose, the data is entered into Excel software and then mathematical calculations are performed through MATLAB software. Results: The research findings show that current year earning is related to next year earning and this issue provides the basis for using Markov process methods to study the characteristic of earning persistency. The results show that since Markov chains related to transfer probability matrices are irreducible for all years, all earnings states in the Markov chain will be reversible. The reversibility of the earnings states means that any earnings state will be repeated in the future. The results also show that the transfer probability matrices calculated based on the net income to sales index and the operating profit to sales index are regular for all years and the sum of years. Therefore, it can be concluded that the effect of the initial situation of accounting earnings on the long-term forecast of accounting earnings persistency is decreasing. In other words, the effect of the initial accounting earnings status on the long-term forecast of accounting earnings decreases over time. Other findings show that the reversibility of accounting earning status based on operating earning to sales ratio and net income to sales indicators are not significantly different. Conclusion: Using current year earnings to predict future earnings has informative content. However, when the longer time horizon is considered, the information content of current year earnings decreases over time, and in the long run there will be virtually no difference between using operating profit and net profit as indicators to assess the persistrncy and reversibility of accounting earnings