EcoSta 2021: Start Registration
View Submission - EcoSta2021
A0303
Title: Stock return prediction: Stacking a variety of models Authors:  Tingting Cheng - Nankai University (China) [presenting]
Bo Zhao - Nankai University (China)
Abstract: An ensembling machine learning approach, Stacking, is employed to refine and combine a variety of linear and nonlinear individual stock return prediction models. In an application of forecasting U.S. market excess return, Stacking can outperform the traditional historical mean benchmark in terms of both in-sample and out-of-sample performances. Moreover, Stacking performs better than a simple combination forecast and the C-Enet forecast in terms of out-of-sample forecasting consistently over time. More importantly, we find that the out-of-sample gains of Stacking are especially evident during extreme downside market movements. Overall, Stacking can generate substantive improvements in out-of-sample market excess return predictability.