How Does the Conditional Quantile Dependence in the FOREX Market Change Depending on the Time of Day? Analysis Based on Ask and Bid Prices

Abstract

In the paper, we show that estimates of the strength of linkages between the exchange rates EUR/USD, AUD/USD, GBP/USD, and NZD/USD below small or above large quantiles de-pend on the time of day, at which the daily returns are calculated. We argue that this is due to the activity of traders in different parts of the world, and the impact of the information flow. We use daily returns calculated for each hour of the day. The analysis is performed separately for bid and ask prices. We model the conditional dependence by means of bivari-ate Markov-switching copula models. Then, quantile dependence probabilities are calculated for the fitted conditional copulas. The models include copulas that can capture different types of asymmetry and tail behavior. Our results show that the applied dynamic dependence measures change significantly, depending on the hour of the day. The rankings of the strength of the dependence below small or above large quantiles, which we obtain using the model confidence set methodology, can be useful in portfolio risk management.