TY - JOUR AU - Ludvigson,Sydney C. AU - Ng,Serena TI - The Empirical Risk-Return Relation: A Factor Analysis Approach JF - National Bureau of Economic Research Working Paper Series VL - No. 11477 PY - 2005 Y2 - July 2005 UR - http://www.nber.org/papers/w11477 L1 - http://www.nber.org/papers/w11477.pdf N1 - Author contact info: Sydney C. Ludvigson Department of Economics New York University 19 W. 4th Street, 6th Floor New York, NY 10002 Tel: 212/998-8927 Fax: 212/995-4186 E-Mail: sydney.ludvigson@nyu.edu Serena Ng Department of Economics Columbia University 440 W. 118 St. International Affairs Building, MC 3308 New York NY 10027 Tel: 212-854-5488 E-Mail: serena.ng@columbia.edu AB - A key criticism of the existing empirical literature on the risk-return relation relates to the relatively small amount of conditioning information used to model the conditional mean and conditional volatility of excess stock market returns. To the extent that financial market participants have information not reflected in the chosen conditioning variables, measures of conditional mean and conditional volatility--and ultimately the risk-return relation itself--will be misspecified and possibly highly misleading. We consider one remedy to these problems using the methodology of dynamic factor analysis for large datasets, whereby a large amount of economic information can be summarized by a few estimated factors. We find that three new factors, a "volatility," "risk premium," and "real" factor, contain important information about one-quarter ahead excess returns and volatility that is not contained in commonly used predictor variables. Moreover, the factor-augmented specifications we examine predict an unusual 16-20 percent of the one-quarter ahead variation in excess stock market returns, and exhibit remarkably stable and strongly statistically significant out-of-sample forecasting power. Finally, in contrast to several pre-existing studies that rely on a small number of conditioning variables, we find a positive conditional correlation between risk and return that is strongly statistically significant, whereas the unconditional correlation is weakly negative and statistically insignificant. ER -