摘要
本文采用三种面板数据方法(加总时序、混同和组平均)考察中国股票收益和交易量的动态关系,以期揭示由于系数异质性问题,建立在三种方法上的统计推断可能会产生矛盾。结果表明,混同和组平均都发现了收益和交易量间双向的格兰杰因果关系,交易量对未来收益的长期影响显著为负;这与投机泡沫理论的预示相符。而加总时序方法只能发现收益对交易量的单向引导。
This paper explores the dynamic relationship between Chinese stock returns and trading volume with three procedures （pooling, aggregating and averaging group estimates）which are widely used in panel data, expecting to show that conflicting inferences can rise due to slope heterogeneity problem. We document bi-directional causality between retum and volume when using pooling and averaging group procedures. Trading volume is negatively correlated with future retum. These facts are consistent with the predictions of speculative bubble theory. However, we only fred return leads volume when using aggregating procedure.
出处
《南方经济》
北大核心
2006年第1期94-104,共11页
South China Journal of Economics
基金
国家自然科学基金课题(批准项目号:70273060)的研究成果之一
关键词
收益
交易量
面板数据
因果关系
Return
Trading Volume
Panel Data
Granger Causality
