Title: The information content of option implied moments and co-moments
Authors: Julian Williams - Durham University (United Kingdom) [presenting]
Yang Zhang - Durham University (United Kingdom)
Abstract: The aim is to introduce a new tensor derived computation of higher dimensional average implied correlations determined from high frequency options panels. Using supersymmetric identities to decrease we can determine average moments and co-moments for equities and incorporate these forward looking measures to a standard asset pricing framework. To compute the metrics we use both a broad market index (S\&P 500) and nine sector indices versus the underlying components. Our empirical analysis shows that the cross section stock returns have substantial exposure to risk captured by the market average correlation factor. The results are robust to various permutations of the estimation procedure. The risk premium of the market average correlation factor is statistically and economically significant when controlling the other common market risk factors or firm characteristics. Finally, we test the higher-order CAPM with the ex ante market beta, gamma, and theta approximated by the option- implied average correlations. In line with the evidence documented in previous literature, we find positive significant risk premiums for the ex ante market beta and theta but mixed results for the ex ante market gamma.