Title：Granular Information and Sectoral Movements

　　Schedule: Nov 18, TH 12:00-1:00 PM

　　Location: Rm 310, Chengming Hall (诚明楼) & 腾讯会议 928 292 828

　　Abstract: This paper shows a strong link between granular information contained in individual stock prices and sectoral movements. Using machine learning algorithms, we find that a predictor that aggregates the price movements of a broad cross-section of individual stocks predicts sector ETF returns out-of-sample. When we combine the structural information of economic links among firms with machine learning, the resulting information signals have even stronger return predictability. These results support the hypothesis of granular origins of aggregate shocks, and illustrate the advantages of structural machine learning.

　　

题目：粒状的信息和行业的变动

摘要：这篇文章展示个别股票包含的粒状信息和行业的变动有很强的关联。使用机器学习演算法，我们发现从个股价格总和而成的预测变量可以在样本外预测指数基金的报酬。当我们将公司间的经济连接和机器学习结合起来，我们可以得到更强的预测力。这些结果支持了总和冲击的粒状来源假说。

Introduction：袁佩轩, 2021年罗格斯大学金融学博士，中国人民大学财政金融学院助理教授。个人信息详见：http://sf.ruc.edu.cn/info/1436/9692.htm