Abstract:
This study tests both the in-sample and out-of-sample predictive value of oil tail risk for the tail risk of US Dollar exchange rates (USD/CAD, USD/GBP and USD/JPY), where the conditional autoregressive value at risk (CAViaR) of the Engle and Manganelli (2004) is used to estimate the tail risks under 1% and 5% VaRs. Thereafter, we construct a predictive model using the best fit tail risks while the predictive value of the oil tail risk is evaluated for both the in-sample and out-of-sample forecasts. We find evidence of a positive association between the oil tail risk and the USD tail risks when the USD/CAD, USD/GBP are considered, where downtowns in the oil markets are capable of causing instabilities in the U.S. foreign exchange market while it is negative for USD/JPY albeit at 5% VaR, suggesting the safe haven property of the latter during oil crisis. Accounting for the dynamics of oil tail risk in the predictive model of the tail risks of USD exchange rates improves both the in-sample and out-of-sample forecasts and the outcome leading to these conclusions is insensitive to the choice of oil price proxy and the magnitude of VaR.
