The VIX Puzzle

Abstract

The CBOE volatility index, VIX, represents the market’s expected volatility over the next 30 days, and is supposed to increase when uncertainty rises. Over the last years, this index has remained at historically low levels, despite a considerable increase in geopolitical tension. This phenomenon is referred to as the VIX Puzzle and is the basis of this work. We examine how realized volatility and stock index correlation on the S&P 500 affect the VIX, and to what extent they can contribute in explaining the depressed levels. This is done by using ordinary least squares (OLS) with Newey West robust standard errors. According to our results, realized volatility exhibits a positive, but weak effect. We conclude that low realized volatility can explain a low VIX to some degree. Further, we observe a positive relation between stock index correlation and movements in the VIX. This result is not statistically significant, and its contribution to the puzzle is weak. In our analyses we also examine and discuss the relation between economic policy uncertainty and the VIX. We find a positive, but small, relationship and confirm the VIX puzzle in our data. As a second contribution to this thesis, we question how well the VIX suits its purpose. By running a simple regression we suggest that the VIX is a biased estimator of future realized volatility. In light of existing literature, we discuss whether the role of the index has changed, as it now may seem like the emergence of VIX derivatives drives the market. In conclusion we question how well the index truly reflects future volatility. In particular, we stress that the trading of VIX derivatives may reduce the role of the index as a reliable measure of future volatility. We consider this an important aspect of the VIX, seeing that it is widely recognized as precisely this.