Modeling price dynamics and risk Forecasting in Tehran Stock Exchange: Conditional Variance Heteroscedasticity Hidden Markov Models

Volatility and risk measurement are essential parameters in risk management programs that can affect economic activities and public confidence in the stock market. Also, these two are the keys in the studies that connect the stock market, economic growth, and other financial factors. In recent years, due to the instability in the Tehran Stock Exchange, controlling the adverse effects caused by the volatility of stock prices, predicting and modeling price dynamics, and measuring risk have become necessary for the participants in this market. In the present research, the class of hidden Markovian index models of conditional variance Heteroskedasticity (HM-GARCH) is used to predict the volatility of stock prices and accounts of the Tehran Stock Exchange. For a comprehensive review, the models are selected to include the characteristics of volatility clustering, asymmetry in volatility (leverage effect), and heavy tail of stock returns (with t-student distribution). Based on RMSE and AME criteria, the HM-EGARCH-Normal Exponential GARCH model with normal distribution is more effective than other models in predicting stock market volatility. Therefore, leverage is necessary to analyze stock market risks using hidden Markov models, but heavy tail distribution is unnecessary. The results indicate that the HM-EGARCH-Normal model appropriately assesses volatility and improves market transparency and risk management forecasts. Also, the VaR and CVaR market risk assessment post-tests using Kupiec and DQ tests do not show evidence of overestimation or underestimation.