Designing a Rule-Based Algorithm to Identify Bull and Bear Markets Regimes

To develop the financial literature related to the identification of financial market regimes, the present research proposes a new rule-based method for selecting turning points in business cycles, which eliminates subjectivity in the process of classifying market regimes. The proposed algorithm has a heuristic approach. No conditions are imposed to determine the duration of the regimes or the amplitude of their returns. Also, for the comparison between the proposed algorithm and the algorithms of Pagan, Lunde, White's bootstrap test has been used in various assets, including the Tehran Stock Exchange index, copper and gold metals, and oil commodities, and the Sharpe ratio has been used as a performance measure in out-of-sample data. The results show that the proposed algorithm has better or equal performance with other identification methods in identifying out-of-sample regimes, especially in time series that are different from capital market index data. The obtained results show the success of the proposed algorithm compared to other algorithms. The straightforward structure of the proposed algorithm avoids the potential fluctuations that occur in parameter optimization by providing a specific method and is useful and practical in identifying different regimes in various sets of time series without changing the parameters.