题 目：Raising the Bar——Iterated Combination Approaches for Commodity Price Forecasting
地 点：腾讯会议 ID：672 723 991
内 容：This paper introduces a novel iterated combination approach to enhance commodity price predictability. We conduct a comprehensive comparison of in- and out-of-sample predictive accuracy between classic combination methods and their iterated counterparts, demonstrating that the proposed approach significantly outperforms alternative methods. Specifically, the implementation of iterated combinations leads to the average increase in in-sample R2 of 8.69% and out-of-sample R2 of 8.43% across all commodity types. We further investigate the economic gains of commodity price predictability using the iterated combination approach, showing that a mean-variance investor can realize substantial economic gains by employing these forecasts. Our results remain robust across various settings, encompassing alternative variable sets, forecasting windows, combination methods, benchmark evaluation models, estimation regression in iterated combinations, and estimation windows, showcasing the superiority of the proposed approach from both statistical and economic perspectives.
报告人简介：马勇，湖南大学金融与统计学院教授、博士生导师。主要研究领域为金融工程与风险管理、信息与金融市场等。在Journal of Futures Markets, Quantitative Finance, International Review of Finance，North American Journal of Economics and Finance, B.E. Journal of Economic Analysis & Policy和管理科学学报、中国管理科学、管理工程学报、管理科学等国内外重要学术期刊发表论文30余篇。