We use the information in intraday data to forecast the volatility of crude oil at a horizon of 1 to 66 days using a variety of models relying on the decomposition of realized variance in its positive or negative (semivariances) part and its continuous or discontinuous part (jumps). We show the importance of these decompositions in predictive regressions using a number of specifications. Nevertheless, an important empirical finding comes from an out-of-sample analysis which unambiguously shows the limited interest of considering these components. Overall, our results indicates that a simple autoregressive specification mimicking long memory and using past realized variances as predictors does not perform significantly worse than more sophisticated models which include the various components of realized variance.
* We are indebted to Lutz Kilian for his many comments that led to improve significantly the presentation of the paper, and Paolo Pasquariello for making useful remarks about an earlier version. We thank Serena Ng for making available her Matlab package for large approximate factor models. We also thank Pierre Perron and Yohei Yamamoto for making available their program for tests of a model with structural breaks.
This paper investigates the relationship between trading volume and price volatility in the crude oil and natural gas futures markets when using high-frequency data. By regressing various realized volatility measures (with/without jumps) on trading volume and trading frequency, our results feature a contemporaneous and largely positive relationship. Furthermore, we test whether the volatility-volume relationship is symmetric for energy futures by considering positive and negative realized semivariance. We show that (i) an asymmetric volatility-volume relationship indeed exists, (ii) trading volume and trading frequency significantly affect negative and positive realized semivariance, and (iii) the information content of negative realized semivariance is higher than for positive realized semivariance.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.