Description
TitleEssays on jump risk factors in financial forecasting
Date Created2020
Other Date2020-05 (degree)
Extent1 online resource (x, 102 pages) : illustrations
DescriptionThis dissertation consists of two essays that explore issues in empirical asset pricing and portfolio management using high-frequency financial econometrics techniques. The first essay investigates the cross-sectional return predictability of various jump risk factors. The second essay develops sparse portfolio variance forecast models that incorporate informative realized jump risk factors.
In Chapter 2, we study the cross-sectional relationship between (small and large) jump variation measures and future stock returns, based on portfolio sorts and Fama-MacBeth type regressions. We document that a new risk factor, signed small jump variation (i.e., the difference between upside and downside small jump variation measures), strongly predicts the cross-sectional variation in future returns. Constructed based on a data-driven threshold, signed small jump variation has stronger predictive power for future returns than other realized risk measures, in the cross-section. We also conduct various experiments (e.g., event studies, etc.) to further explore the linkages between different jump risk measures and economic factors relating to news in the markets. We show that large jumps are closely associated with ``big'' news. While such news related information is embedded in large jump variation, the information is generally short-lived, and dissipates too quickly to provide marginal predictive content for subsequent weekly returns. By contrast, we find that small jumps are more likely to be diversified away than large jumps, thus tend to be more closely associated with idiosyncratic risks, and are therefore more likely to be driven by liquidity conditions and trading activity.
In Chapter 3, we investigate whether the decomposition of realized covariance matrices of portfolios of asset returns into components based on both the signs and magnitudes of the underlying high-frequency returns is useful for forecasting. In particular, our decomposition separates realized covariation into components based on signs (positive and negative) and magnitudes (continuous, small jump, and large jump). Sparse portfolio variance forecast models, which are constructed by utilizing the most informative covariance components, produce significant improvements in predictive accuracy. We show that such predictive gains can be traced to the identification of short-lived pricing signals associated with co-jumps.
NotePh.D.
NoteIncludes bibliographical references
Genretheses, ETD doctoral
LanguageEnglish
CollectionSchool of Graduate Studies Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.