Academic journal article The Journal of Business and Economic Studies

Strategic Asset Allocation and Markov Regime Switch with GARCH

Academic journal article The Journal of Business and Economic Studies

Strategic Asset Allocation and Markov Regime Switch with GARCH

Article excerpt


During the financial crisis of 2008, the S&P 500 Implied Volatility Index (VIX), known as the "fear gauged," jumped to 80% of the highest level it has ever reached. Portfolio managers faced tremendous pressures in these environments of such high levels market volatility. Because it is well known that asset allocation dominates portfolio performances, this paper focuses on asset allocation strategies. It develops a strategic asset allocation solution for portfolio management under all conditions and at all levels of market volatility. The approach is to derive a dynamic optimal portfolio that is based on the well-known asset allocation Black-Litterman (1991) framework. In addition, this paper proposes a methodology that considers the features of volatility regime-switching over time. This new strategic framework allows portfolio managers to derive a systematically optimal portfolio in a timely, accurate fashion.

Keywords: Asset allocation, volatility, regime switching, GARCH, portfolio optimization.

JEL Classification: Gl 1, G12

(ProQuest: ... denotes formulae omitted.)


In the last few decades, strategic asset allocation has become an intensely interesting topic for institutional investors and portfolio managers. The efficiency and timely accuracy of asset allocation can significantly improve a portfolio's performance. Research literature shows that accurate asset allocation can contribute to about 90% of a portfolio's return. This paper strives to achieve two goals:

1. Allocate assets within a portfolio based on time-varying asset-return volatilities.

2. Derive an optimal portfolio that provides superior returns.

Our approach to this paper provides for robust and sophisticated methodologies for investors in managing their portfolio's returns and risks. The model for this paper is incorporated in the Black-Litterman model with Markov Regime-Switching parameters in GARCH process (MRS-GARCH). The model is designed in such a way so that the results from MRS-GARCH model are represented as an investor's views in the Black-Litterman model. We tested the new model (MRS-GARCH with Black-Litterman) with data from 1969 to 2011. The new model yields some remarkable results. Especially during the periods of the latest financial crisis in 2008, the performance of the tested portfolio, based on the new model, are promising.

Asset returns show time-varying volatility characteristics. This phenomenon has been researched and documented in research literature for decades. Figure 1 shows historical 10-year rolling volatilities in the United States. These data are taken from the MSCI World Index by country (e.g., United States). In this figure, during the period from 1982 to 1984, the annual volatility level was around 16%. At year 1986, the volatility level dropped to 14% per annum. However, during the period from 1988 to 1992, the volatility level was back to 16% and 17% per annum. At year 1998, the volatility level dropped to 13% per annum - the lowest point in history. During the collapse of Lehman Brother Holdings in 2008, the rolling volatilities reached the highest level in history at around 17% per annum. The data from the figure clearly showed that volatility levels change over time.

Another phenomenon is that asset return volatilities are inclined to be sticky when return volatilities are at higher levels or lower levels. This is often referred to as "volatility cluster." Over the long run, volatilities of asset returns will decrease if they reach too high and increase if they reach too low. This is referred to as "volatility mean reversion" in financial literature. For decades there has been research literature focused on the phenomenon of both "volatility cluster" and "mean reversion."

To accurately forecast time-varying volatilities is not only a benefit for risk management of an asset or a portfolio of assets but, it is also critical in forecasting risk-adjusted asset returns and determining asset allocations in order to obtain an optimal portfolio. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.