An Analysis of Return States in Iran Stock Market: Hidden Semi-Markov Model Approach

Analyzing the behavior of Tehran Stock Market, based on the daily asset return for the duration between 1387 and 1397 has been the main aim of this research.

Tehran Stock Market daily asset return can be considered as a time-series and therefore all existing models can be applied to it. Considering the distributional and temporal properties of such series, it can be shown that the series is stationary. Hence, Hidden Semi-Markov Model, which is widely used for analyzing time series, could be employed to analyze this series.

Based on Kolmogorov-Smirnov test and Akaike and Bayesian indices, the best density function for the series is a three parameter Gussian Mixture. Moreover, employing three-state Hidden Semi-Markov Model would be the suitable method for modeling. In addition, it was found that Tehran Stock Market followed three states namely bull, bear, and sidewalk and the definitions for such states have been given, while the probability of being in each state has also been provided.

Tehran Stock Market was in sidewalk state around half of the analyzed duration and the luckiest state after both bear and bull states was sidewalk. The market almost never came to bull state after the bear state. Moreover, the chance of getting into bear state from sidewalk was three times more than the chance of getting into the bull market.