Investors' Individual Herd Behavior:Distribution and Degree——A Matrix Model Based on Partitioning Clustering Algorithm

Li Xuefeng,Li Jiaming
2011-01-01
Abstract:In order to give a precise description of the individual herd behavior,we develop a matrix model based on partitioning clustering algorithm by use of LSV models.Focusing on institutional investors,we use this new model to examine the herd behavior of open-ended funds in our country and find that,the institutional investors' individual herd behavior index show a leptokurtic and negatively skewed distribution;institutional investors prefer to herd in large-cap stocks;the trend of market's general herd behavior is upward;and institutional investors' herd behavior increases both gains and risks.
What problem does this paper attempt to address?