Research Repository

Hedge fund return predictability; To combine forecasts or combine information?

Panopoulou, E and Vrontos, S (2015) 'Hedge fund return predictability; To combine forecasts or combine information?' Journal of Banking and Finance, 56. 103 - 122. ISSN 0378-4266

[img]
Preview
Text
1-s2.0-S0378426615000655-main.pdf - Accepted Version

Download (749kB) | Preview

Abstract

© 2015 Elsevier B.V. While the majority of the predictability literature has been devoted to the predictability of traditional asset classes, the literature on the predictability of hedge fund returns is quite scanty. We focus on assessing the out-of-sample predictability of hedge fund strategies by employing an extensive list of predictors. Aiming at reducing uncertainty risk associated with a single predictor model, we first engage into combining the individual forecasts. We consider various combining methods ranging from simple averaging schemes to more sophisticated ones, such as discounting forecast errors, cluster combining and principal components combining. Our second approach combines information of the predictors and applies kitchen sink, bootstrap aggregating (bagging), lasso, ridge and elastic net specifications. Our statistical and economic evaluation findings point to the superiority of simple combination methods. We also provide evidence on the use of hedge fund return forecasts for hedge fund risk measurement and portfolio allocation. Dynamically constructing portfolios based on the combination forecasts of hedge funds returns leads to considerably improved portfolio performance.

Item Type: Article
Subjects: H Social Sciences > HG Finance
Q Science > QA Mathematics
Divisions: Faculty of Science and Health > Mathematical Sciences, Department of
Depositing User: Jim Jamieson
Date Deposited: 14 Apr 2015 10:03
Last Modified: 30 Jan 2019 16:19
URI: http://repository.essex.ac.uk/id/eprint/13515

Actions (login required)

View Item View Item