Machine-Learning-Based Return Predictors and the Spanning Controversy in Macro-Finance

The authors are grateful to Haoxiang Zhu (finance department editor), an associate editor, and an anonymous referee for their extensive and detailed comments that helped to improve the article substantially. The authors also thank Yakov Amihud, Gurdip Bakshi, Charles Cao, Long Chen, Mike Chernov, Ed Coulson, Heber Farnsworth, Peter Feldhütter (American Finance Association Conference discussant), Laura Field, Itay Goldstein, Olesya Grishchenko, Raymond Kan, Anh Le, Hong Liu, Stefan Nagel, Matt Pritsker (Darla Moore Fixed Income Conference discussant), Marco Rossi, Ilona Shiller (Financial Management Association Conference discussant), Andrea Tamoni, Dan Thornton, Joel Vanden, Hong Yan (Summer Institute of Finance Conference discussant), Weina Zhang (China International Conference in Finance discussant), Wei Zhong, and Hao Zhou; seminar participants at Fordham, National University of Singapore, Penn State Smeal College, Penn State Mathematics Department, Federal Reserve Bank of Philadelphia, Singapore Management University, Federal Reserve Bank of St. Louis, Temple University, University of Rhode Island, University of Waterloo, and University of Wisconsin-Milwaukee; and participants at the 20th Federal Deposit Insurance Corporation Derivatives Securities and Risk Management Conference, the 2010 China International Conference in Finance, the 2010 Financial Management Association, the 2010 Summer Institute of Finance Conference, the 2011 American Finance Association, and the 2012 Darla Moore Fixed Income Conference for valuable comments and suggestions. The authors also thank Aaron Henrichsen and Terrence O’Brien for editorial assistance. This paper is a new incarnation of the previous work circulated under the titles “Determinants of Bond Risk Premia” and “Determinants of Bond Risk Premia: A Machine-Learning-Based Resolution of the Spanning Controversy.”