Dynamic Correlations, Estimation Risk, and Porfolio Management During the Financial Crisis

We evaluate alternative multivariate models of dynamic correlations in terms of realized out-of-sample Sharpe ratios for an active portfolio manager who rebalances a portfolio of international equities on a daily basis. The evaluation period covers the recent financial crisis which was marked by increased volatility and correlations across international stock markets. Our results show that international correlations fluctuate considerably from day to day, but we find no evidence of decoupling between emerging and developed stock markets. We also find that the recursively updated dynamic correlation models display remarkably stable parameter estimates over time, but that none yields statistically better portfolio performances than the naive diversification benchmark strategy. The results clearly show the erosive effects of model estimation risk and transactions costs, the benefits of limiting short sales, and the far greater importance of including a risk-free security in the asset mix whether or not market turbulence is high.