5. Conclusions
This study shows effectiveness of the utilized copula-EGARCH-DCC model to reduce variance of portfolios of foreign currencies of the Australian dollar, Canadian dollar, euro, British pound and Japanese yen. For the portfolio hedging purposes, it is recognized efficiency of the estimated bivariate model to account for the evolution of the dynamic conditional correlation between the spot and futures markets. However, the measures of the hedging performance show that the estimated unconditional OLS model's ability to reduce variance of a portfolio is generally larger compared to the other models. The only exception is the dynamic conditional correlation model estimated for the currency markets, i.e. the copula-EGARCH-DCC model with the external realized volatility estimators included into the variance equation of the model. This can be seen as efficiency of the model to account for the clustered nature of the data variance. The in-sample hedging effectiveness in this study examined, suggests that the conditional hedge outperforms the traditional unconditional hedging strategy. As the estimation results show, the conditional correlation model with included external realized variance estimators is superior in portfolio variance reduction. Also, the estimation results of the longer time period in this research applied confirm the findings. In effect, the external realized variance estimator included into the variance equations of the model improves the model ability to fit into the data of the currency market returns estimated. The outcome of the superiority is a result from the information content of the realized variance estimates that improves ability of the model to estimate the conditional variance of the market data in low and high volatility periods. In addition, it is observed that the constant correlation models hedging performance is weak, suggesting that the model is inadequate as used to minimize variance of a portfolio.
