SMU SOE Seminar Series (April 13, 2026): An Anatomy of Asset Returns

TOPIC:

ABSTRACT

We propose a new high-frequency econometric framework to decompose asset returns using realized autocovariances of power transformed semi-returns. The nonlinear transformation reshapes the interaction among volatility, drift, leverage, and jumps, yielding moment conditions that separate continuous and jump components without truncation. The framework further delivers separate identification of drift and leverage effects, providing a unified framework to isolating the core drivers of price dynamics. Using equity index futures and the cross section of U.S. stocks, we reveal two mechanisms through which overnight returns impact intraday returns: a momentum channel driven by continuous returns and a reversal channel driven by jumps. This explains why aggregate overnight-to-intraday predictability may appear weak without return decomposition and why reversal patterns emerge primarily in extreme cross sectional deciles.