Dynamic portfolio management with lag-optimised trading indicator using SeroFAM and genetic algorithm

Abstract

Moving average convergence divergence histogram (MACDH) is a common technical indicator used in trading. However, MACDH is a lagging indicator, affecting its effectiveness. This paper explores the use of forecasted data from the SeroFAM model along with Genetic Algorithm to create a novel optimized fMACDH indicator that shows reduced lag in predicting trend reversals. Two portfolio rebalancing strategies using the optimized fMACDH indicator, Tactical Buy and Hold Strategy and Rule-Based Business Cycle Portfolio Rebalancing, are then proposed and evaluated. An input depth and prediction depth of one each was found to be the optimal parameters for the SeroFAM model and fMACDH indicator as proposed by Tan and Quek. Genetic Algorithm was then used to optimize the parameters for reducing lag in the fMACDH indicator. The results obtained are promising, with simple trend reversal trading based on the optimized fMACDH indicator consistently outperforming buy and hold strategy across multiple market indexes. Tactical Buy and Hold strategy is a rule-based portfolio rebalancing strategy that takes advantage of the relative difference in risk levels to perform rebalancing during trend reversals. This strategy was evaluated using the optimized fMACDH indicator and was able to outperform simple buy and hold of the individual ETFs, despite commission rate loss. Rule-Based Business Cycle Portfolio Rebalancing is a rule-based portfolio rebalancing strategy that takes advantage of the offsets between the business cycles of different markets. This strategy was once again evaluated using the optimized fMACDH indicator. The results are highly encouraging, consistently outperforming simple Buy and Hold of each individual ETF. In addition, Rule-Based Business Cycle Portfolio Rebalancing seems to outperform Tactical Buy and Hold over the same time periods.