Deep Learning, Predictability, and Optimal Portfolio Returns
Preprint
- 7 September 2020
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
We study optimal dynamic portfolio choice of a long-horizon investor who uses deep learning methods to predict equity returns when forming optimal portfolios. The results show statistically and economically significant out-of-sample portfolio benefits of deep learning as measured by high certainty equivalent returns and Sharpe ratios. Return predictability via deep learning generates substantially improved portfolio performance across different subsamples, particularly the recession periods. These gains are robust to including transaction costs, short-selling and borrowing constraints.Keywords
All Related Versions
This publication has 61 references indexed in Scilit:
- Time-varying short-horizon predictability☆Journal of Financial Economics, 2011
- Out-of-Sample Equity Premium Prediction: Combination Forecasts and Links to the Real EconomyThe Review of Financial Studies, 2009
- Equilibrium Portfolio Strategies in the Presence of Sentiment Risk and Excess VolatilityThe Journal of Finance, 2009
- Predicting Excess Stock Returns Out of Sample: Can Anything Beat the Historical Average?The Review of Financial Studies, 2007
- The Dog That Did Not Bark: A Defense of Return PredictabilityThe Review of Financial Studies, 2007
- Efficient tests of stock return predictability☆Journal of Financial Economics, 2006
- Spurious Regressions in Financial Economics?The Journal of Finance, 2003
- Investing for the Long Run when Returns Are PredictableThe Journal of Finance, 2000
- Dividend yields and expected stock returnsJournal of Financial Economics, 1988
- Stock returns and the term structureJournal of Financial Economics, 1987