Document Type : Research Paper
Authors
1
Ph.D. Candidate of Accounting, Faculty of Management & Accounting, Shahid Beheshti University, Tehran, Iran
2
ssistant Professor of Accounting, Faculty of Management & Accounting, Shahid Beheshti University, Tehran, Iran
3
Associate Professor of Accounting, Faculty of Management & Accounting, Shahid Beheshti University, Tehran, Iran
4
Assistant Professor of Industrial Management & Information Technology, Faculty of Management & Accounting, Shahid Beheshti University, Tehran, Iran
10.22103/jak.2024.23666.4071
Abstract
Objective: The objective of this research is to design a model for evaluating the risk disclosure quality based on the alignment probability between risk disclosure and the actual risk state of the firm.
Methods: Hidden Markov Chain method is used to analyze and predict risk disclosure patterns and the firm’s risk state based on firm risk indicators. The Markov Chain Monte Carlo algorithm is utilized, based on annual data from 150 listed and active firms on the Tehran Stock Exchange, covering the period from 2013 to 2019. This data is used to train and estimate the model, and data from 2020 to 2022 is used to evaluate the model’s prediction accuracy.
Results: The risk indicators such as beta, stock return volatility, export revenue ratio, and operating cash flow volatility significantly influence the likelihood of alignment between risk disclosure and the firm’s risk state. Conversely, liquidity, operating leverage, financial leverage, and firm size reduce the risk disclosure quality under specific risk conditions. The evaluation of risk disclosure quality for the sampled firms, based on the research model, predicts an average alignment probability of 43% between the level of risk disclosure and the firm’s risk state. Furthermore, the prediction error test indicates that the model has a high prediction accuracy.
Conclusion: The model can provide valuable insights into analyzing and predicting the risk disclosure quality using quantitative information from financial statements. Unlike previous approaches, this model is not limited to counting interpretive disclosures and reduces the subjectivity involved in the evaluation process.
Keywords
Main Subjects
)