Marcos Aurélio Rodrigues

Spot and future BRL/USD exchange rates were used to estimate the optimal hedge ratio and hedging effectiveness. The tests show evidence of ARCH effects, non-Gaussian distributions, a cointegrating vector equal to the “basis”, and non-constant conditional correlations. Thus the series were modeled using multivariate GARCH with error-correction term equal to the lagged “basis” and a Studentized distribution. The results point out superior hedge effectiveness, in sample (48%) and out of sample (66%), for the DCC GJR model under a t distribution.