Revisiting the bull and bear markets notions in the Tunisian stock market: New evidence from multi-state duration-dependence Markov-switching models

This paper attempts to better apprehend the bull and bear markets notions by extending the Markov-switching model of Maheu and McCurdy (2000) for the multi-state case. By accounting for the duration dependence in conditional mean return, volatility, risk-return trade-off and transition probabilities, our four-state model with regimes characterized as boom, crash, bull and bear states enables us to define the bull and bear markets according the trend-based schemes. Finally, we establish a market state indicator which can detect the market cycle's inflexions and highlights the deterioration of the market conditions during the post-revolution period. (C) 2016 Elsevier B.V. All rights reserved.