Title: Enhancing trading strategies under regime shifts
Authors: Christina Erlwein-Sayer - OptiRisk Systems (United Kingdom) [presenting]
Tilman Sayer - Advanced Logic Analytics (United Kingdom)
Gautam Mitra - OptiRisk Systems Ltd (United Kingdom)
Abstract: A Markovian regime switching model is applied to identify market regimes in equity markets. Hidden regimes are commonly modelled through Hidden Markov Models. To detect non-observable market regimes, we apply filtering techniques to filter these possible market states out of the observation process. We use the detected regime to choose suitable long and short selling ratios within the portfolio for a daily trading strategy, which is based on a Second Order Stochastic Dominance criterion. We analyse major indices, find most probable state sequences through the Viterbi algorithm and find suitable trading strategies for each regime. The performances of the trading strategies are computed using well established static and dynamic measures like Sharpe and Sortino ratios as well as max drawdowns and days to recovery. We find that the consideration of regime shifts improves the performances of our trading strategies.