Correlation regimes in international equity and bond returns

Measuring comovements across international financial markets is important for policy purposes and portfolio management. We develop a new approach to analyse such comovements in relation to key state variables, such as equity market volatility and short-term interest rates. These state variables can identify regimes in comovements through a fast, tractable threshold model. The advantage over existing methods is that our model can be easily estimated and does not have a serious bias, even when applied to large asset portfolios. Out-of-sample portfolio evaluation shows that our method outperforms the standard dynamic conditional correlation approach, especially during the recent global financial crisis when financial market comovements experienced substantial regime shifts. Overall, we contribute to the empirical literature by shedding new light on the fundamental drivers of comovements across international financial markets which can help guide analysis of systemic risk, asset allocation and hedging.