EconPapers    
Economics at your fingertips  
 

How Regimes Affect Asset Allocation

Andrew Ang and Geert Bekaert

Financial Analysts Journal, 2004, vol. 60, issue 2, 86-99

Abstract: International equity returns are characterized by episodes of high volatility and unusually high correlations coinciding with bear markets. This article provides models of asset returns that match these patterns and illustrates their use in asset allocation. The presence of regimes with different correlations and expected returns is difficult to exploit within a framework focused on global equities. Nevertheless, for global all-equity portfolios, the regime-switching strategy dominated static strategies in an out-of-sample test. In addition, substantial value was added when an investor switched between domestic cash, bonds, and equity investments. In a persistent high-volatility market, the model told the investor to switch primarily to cash. Large market-timing benefits are possible because high-volatility regimes tend to coincide with periods of relatively high interest rates. International equity returns are more highly correlated with each other in high-volatility bear markets than in normal times. Regime-switching (RS) models perform well at replicating the degree of asymmetric correlations observed in the data because they draw data from a normal regime most of the time but transition to a bear market regime when asset returns are, on average, lower and much more volatile than in normal times. The regimes are persistent, and in the bear markets, asset correlations are higher than in the normal regime.We show that the presence of regimes in international returns is exploitable in active asset allocation programs. We illustrate how the presence of regimes can be incorporated into two asset allocation programs—a global equity allocation setting (with six equity markets) and a market-timing setting for U.S. cash, bonds, and equity. For global portfolios, the optimal equity portfolio in the high-volatility bear market is very different from the optimal portfolio in the normal regime; for example, it is more home biased in bear markets. For a domestic U.S. portfolio, optimally exploiting regime switches implies portfolio shifts into bonds or cash when a high-volatility bear market regime is expected.To build a quantitative model for the international asset classes, we incorporated two regimes in the basic capital asset pricing model. Conditional means, volatilities, and correlations in this model depend on which regime prevails at each time. The RS model can produce rich patterns of stochastic volatility and time-varying correlations. The regimes are identified endogenously through the estimation procedure, which provides an easy way for an investor to determine which regime is prevailing at a given time.The regime-dependent strategies have the potential to outperform static investment strategies because they set up a defensive portfolio in the bear market regime that hedges against high correlations and low returns. Theoretically, the presence of two regimes implies two mean–variance frontiers, one for each regime. The presence of two regimes and two frontiers means that the regime-switching investment opportunity set dominates the investment opportunity set offered by one unconditional frontier. For example, in the global asset allocation setting in the normal regime, the unconditional tangency portfolio yielded a Sharpe ratio of 0.619. The investor could improve this trade-off to 0.871 by holding the risk-free asset and the optimal tangency portfolio. Similarly, in the bear market regime, the unconditional tangency portfolio had a Sharpe ratio of only 0.129, but it could be improved to 0.268 by holding the optimal regime-dependent tangency portfolio.To illustrate the practical implementation of the regime-dependent strategies, we used an out-of-sample analysis starting in 1985 and ex post Sharpe ratios as a performance criterion. For the global asset allocation example, the regime-switching strategy's Sharpe ratio was more than double the world market portfolio's Sharpe ratio.In an out-of-sample test of the market-timing model for U.S. equities, bonds, and cash, we found that substantial value could be added when an investor moved assets among cash, bonds, and equity investments. When a persistent high-volatility market hit, the investor switched primarily to cash. Market-timing benefits were large because high-volatility markets tend to coincide with periods of relatively high interest rates.The results reported here provide a clear demonstration of how active managers can incorporate regime-switching strategies to enhance returns in market-timing models. Our results lead to two robust conclusions. First, one can add value by considering regime switches in global all-equity portfolios; the presence of a bear market regime does not negate the benefits of international diversification. Although portfolios in the high-correlation regime are more home biased, they still involve significant international exposure. Second, an even more valuable situation in which to consider regime-switching models is tactical asset allocation programs that allow switching to a risk-free asset.

Date: 2004
References: Add references at CitEc
Citations: View citations in EconPapers (8)

Downloads: (external link)
http://hdl.handle.net/10.2469/faj.v60.n2.2612 (text/html)
Access to full text is restricted to subscribers.

Related works:
Working Paper: How do Regimes Affect Asset Allocation? (2003) Downloads
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:ufajxx:v:60:y:2004:i:2:p:86-99

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/ufaj20

DOI: 10.2469/faj.v60.n2.2612

Access Statistics for this article

Financial Analysts Journal is currently edited by Maryann Dupes

More articles in Financial Analysts Journal from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().

 
Page updated 2025-03-20
Handle: RePEc:taf:ufajxx:v:60:y:2004:i:2:p:86-99