Volatility Connectedness Across Oil-Dependent Currencies : A Multivariate Har with Measurement Errors Approach

This paper proposes a convenient approach to measure the time-varying volatility connectedness indexes by developing the connectedness measures within a multivariate Heterogeneous Autoregressive model with measurement errors (VHAR-Q). We utilize this framework to examine the dynamic volatility connectedness between major oil-exporting and oil-importing currencies. We find pronounced volatility connectedness within each group of currencies. Meanwhile, the cross-group connectedness is relatively low and concentrated in the short run. In the time-varying cross-group connectedness, the oil-exporting group is consistently the net volatility transmitter. More importantly, this cross-group volatility connectedness is conditional on oil volatility’s regime and significantly increases when oil volatility switches from a low to a high state. We illustrate the financial implication of our findings by proposing a dynamic oil’s regime-dependent trading strategy that outperforms the buy-and-hold strategy in terms of risk-adjusted return