Daily reading lists:

• Stock return prediction, Stacking a variety of models
• EPU spillovers and stock return predictability: A cross-country study

Stock return prediction, Stacking a variety of models

• 将机器学习中的 stacking 算法用于股权溢价交易，
• 通过一组底层模型来训练一个高层模型，即元模型，从而产生更精细化的预测。

EPU spillovers and stock return predictability: A cross-country study

• 本文考察了经济政策不确定性（EPU）的传递（即EPU溢出效应） 在多大程度上推动了全球股市的收益（因此可以预测）。
• 文章亮点：

  • 应用高维 factor-copula model 来估计 EPU spillovers (EPU 溢出效应) as expected probability in distress (EPD) (作为预期陷入困境的概率).
  • 本文研究了 EPU 溢出效应对全球 23 个市场的股票收益的影响。
  • This paper finds EPD significantly and negatively predicts country-level returns in global markets. (本文发现 EPD 显着地负面预测了全球市场的国家级回报)。
  • Return predictability through the spillover channel is highly significant and dominates that through the domestic channel. (通过溢出渠道的回报可预测性非常重要，并且在国内渠道中占主导地位。)
  • The results are robust to various model specifications and alternative measures of EPU spillovers. (结果对各种模型规范和 EPU 溢出的替代措施是稳健的。)