Style based portfolios with regime shifting

Abstract

Style investing (or factor investing) is one of the most wildly discussed topics in the modern financial world. It is considered as a paradigm which provides new vision of objects and mechanisms of investing. The purpose of style investing is to capture risk premia, which are associated with some specific risks which shows up with the diversifying of systematic risks of a classical asset portfolio. In this thesis, six different style, such as Value, Size, Momentum, Quality, High dividend yield, and Volatility, were used for asset allocations. The broad academic research of factors features shows high diversification benefits from applying style instead of classical asset classes.

The asset allocation approach also plays a critical role in constructing an effective portfolio. Five different risk-based allocation methods, such as the inverse volatility, the equally risk contributed, the alpha risk parity, the maximum diversification, and the diversified risk parity, are applied to construct style portfolios.

The study also highlights the existence of risk-on and risk-off regimes in assets returns. The regime shifting statistical model can be used for showing the comovements over time between the portfolio returns and stages of the economic cycle or some financial policies. The first insight in the regime-switching approaches was done by Hamilton (1989). He asserted that the model is governed by an unobservable regime variable which regulates the regime switching mechanism and follows Markov process, which based on a one-step autoregression without assuming how the previous state was achieved.

The results of the research supported the idea that the portfolio with regime switching mechanism outperforms the best risk-based portfolio and the benchmark. However, this result is found just for the portfolio in a long-run.