Abstract
Theory predicts that aggregate volatility ought to be a priced risk factor. In an influential study with more than 1000 citations on Google Scholar, Profs Andrew Ang, Bob Hodrick, Yuhang Xing and Xiaoyan Zhang propose an ex post factor, FV IX, intended as a proxy for aggregate volatility risk. Their test validating FV IX is an ordinary least squares (OLS) regression of portfolio excess returns on FV IX and other independent variables over the data period February 1986 – January 2001. October 1987 is an outlier, in which FV IX exhibits a 26 sigma deviation. The inclusion of this outlier results in a spurious reduction of the regression standard error by more than a factor of two, creating the appearance of statistical significance when none is present. The speaker explains how standard statistics can be used to assess the suitability of a dataset for OLS regression.
About the speaker
Prof Robert Anderson received his PhD in Mathematics from Yale University in 1977. He was McMaster Fellow at McMaster University from 1977 to 1978. He went to Princeton University as Assistant Professor of Economics and of Mathematics from 1978 to 1982, and Associate Professor of Economics from 1982 to 1983. He joined the University of California at Berkeley in 1983, and is currently Professor of Economics and of Mathematics, Coleman Fung Professor of Risk Management, and Director of the Coleman Fung Risk Management Research Center.
Prof Anderson’s research ranges from the intersection between probability theory and logic, to general equilibrium theory, to mathematical finance. His current research focuses on the determination of portfolio returns.
Prof Anderson has taken on numerous assignments for the University of California system Academic Senate, including Vice Chair and Chair of the UC Academic Senate and Faculty Representative to the Board of Regents in 2010-12. He was named an Alfred P. Sloan Research Fellow in 1982 and a Fellow of the Econometric Society in 1988.
|
