Scholarlink

<p>This paper introduces a novel voting-based approach to forecasting the equity premium, emphasizing directional consistency and robustness to structural change. Using a comprehensive dataset of 155 macroeconomic, financial, and technical predictors spanning January 1960 to December 2022, we develop a two-step framework that combines statistical screening with voting-based model weighting. Empirical results show that our method consistently outperforms traditional forecast combination techniques as well as several sophisticated alternatives-including LASSO, Elastic Net, and dynamic factor models. It nearly doubles out-of-sample forecast accuracy relative to standard benchmarks and delivers substantial economic gains for mean-variance investors, as measured by improvements in Certainty Equivalent Returns. Notably, the method excels during recessionary periods by adaptively emphasizing predictors-such as interest rates, volatility, and labor market indicators-whose importance rises in turbulent conditions. These results underscore the method’s advantages in high-noise, high-uncertainty environments, making it a valuable tool for asset allocation, risk management, and policy analysis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Out-of-sample equity premium prediction: A voting approach to forecast combination

  • 期刊论文

  • 发表日期：

    2025年7月28日

查看全文
  • Hyder Ali,
  • Salma Naz

摘要

This paper introduces a novel voting-based approach to forecasting the equity premium, emphasizing directional consistency and robustness to structural change. Using a comprehensive dataset of 155 macroeconomic, financial, and technical predictors spanning January 1960 to December 2022, we develop a two-step framework that combines statistical screening with voting-based model weighting. Empirical results show that our method consistently outperforms traditional forecast combination techniques as well as several sophisticated alternatives-including LASSO, Elastic Net, and dynamic factor models. It nearly doubles out-of-sample forecast accuracy relative to standard benchmarks and delivers substantial economic gains for mean-variance investors, as measured by improvements in Certainty Equivalent Returns. Notably, the method excels during recessionary periods by adaptively emphasizing predictors-such as interest rates, volatility, and labor market indicators-whose importance rises in turbulent conditions. These results underscore the method’s advantages in high-noise, high-uncertainty environments, making it a valuable tool for asset allocation, risk management, and policy analysis.