Characteristics Are Covariances: A Unified Model of Risk and Return

Abstract

Louis Federal Reserve, and Yale for helpful comments. We are grateful to Andreas Neuhierl for generously sharing data with us. A previous draft was circulated under the name "Some Characteristics Are Risk Exposures, and the Rest Are Irrelevant." AQR Capital Management is a global investment management firm, which may or may not apply similar investment techniques or methods of analysis as described herein. The views expressed here are those of the authors and not necessarily those of AQR or the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.

Conditional Dynamics and the Multi-Horizon Risk-Return Trade-Off

“…If Equation (11) holds, α = 0. Ferson and Harvey (1999) and Kelly, Pruitt, and Su (2019) are prominent examples of tests of Equation 12with β t posited to be an affine function of conditioning variables z t .…”

“…There are many papers that test conditional versions of factor models. For instance, Boguth, Carlson, Fisher, and Simutin (2011), Ferson and Harvey (1999), Farnsworth, Ferson, Jackson, and Todd (2002), Jagannathan and Wang (1996), Kelly, Pruitt, and Su (2019), Lettau and Ludvigson (2001), Lewellen and Nagel (2006), and Moreira and Muir (2017). Our contribution relative to this literature is to show that MHR in asset pricing tests effectively serve as conditioning variables endogenous to the model and that, empirically, multi-horizon factor returns indeed are informative in terms of uncovering novel conditional dynamics of prominent factor models.…”

A Dynamic Network of Arbitrage Characteristics

“…Some recent papers such as Kim, Korajczyk and Neuhierl (2019) [20] and Kelly, Pruitt and Su (2019) [19] analyzed a similar model to ours, which assumes that both h(X i ) and g j (X i )…”

“…And then they showed that their arbitrage portfolios are statistically and economically significant. However, Kelly, Pruitt and Su (2019) [19] applied instrumented principal component analysis (IPCA) to the entire time span from 1965 to 2014, and came to a different result, with no evidence to reject the null hypothesis H 0 : h(X i ) = X i B = 0. The introduction of IPCA can be found at Kelly, Pruitt and Su (2017) [18].…”

Portfolio Efficiency with High-Dimensional Data As Conditioning Information

“…Several studies in macroeconomics made use of principal components to condense information (see Stock and Watson (2002a,b)). Kelly et al (2019) proposed a dimension reduction based on PCA to explain the cross-section of average returns. Gu et al (2018) made use of PCR, among other machine learning techniques, for measuring asset risk premia.…”