VIX - Volatility Index - emerged as an alternative calculation of
implied volatility in order to mitigate some problems encountered in models
of the Black-Scholes. This kind of volatility is seen as the best predictor
of future volatility, given that option traders' expectations are embedded
in their values. In this paper we test whether the VIX has more predictive
power for future volatility and contains relevant information not found in
time series models time for non-negative variables, treated by
multiplicative error model. The results indicate that the VIX has greater
predictive power in periods of economic stability, but does not contain
relevant information to the realized volatility which here is considered as
the "true volatility". In periods of economic crisis the result changes,
with the VIX presenting the same explanatory power, but contains relevant
information in the short term.