The explanatory power of signed jumps for the risk-return tradeoff
Abstract
Patton and Sheppard (2011) develop the concept of signed jumps as the difference between positive and negative realized positive semivariances. This quantity is well-suited for gauging the risk-return trade-off at high-frequency as it is well-defined each day and, contrary to the squared jump contribution following Barndorff-Nielsen and Shephard (2004, 2006) which is dedicated to rare jumps, it is signed. We show that signed jumps only occasionally help in explaining future returns, at least when the horizon of interest is one-day ahead as in Bali and Peng (2006).