Stocks, Gold and Crude Oil: How Valuable are Volatility and Correlation Timing?

We study the sources of portfolio returns under parameter uncertainty arising from allocation strategies based on holdings of the NASDAQ index, WTI crude oil and gold. We investigate the contribution of volatility and correlation forecasts for dynamic portfolio allocations in the model of Brandt, Goyal, Clara, and Stroud (2005). We estimate alternative forecasting models using Bayesian methods, and evaluate investor’s utility under the Bayesian predictive density. We then compare the performance of different portfolio strategies through both Sharpe ratios and utility-based metrics. We show that dynamic strategies based on timing volatilities and correlations can add positively to the economic gains generated by non-diversified portfolios involving holdings of either crude oil or gold only. Hence, the economic benefits generated by holding crude oil and gold in asset allocations with stocks arise from the predictability of their volatilities and correlations.