1996—2009年上海A股市场规模效应研究:基于收益的可预测性  

The Size Effect of Shanghai Stock Exchange during the Period of 1996-2009: on the Basis of Return Predictability

作　　者：韦倩[1] 

机构地区：[1]山东大学经济研究院,山东济南250100

出　　处：《管理工程学报》2014年第1期74-80,93,共8页Journal of Industrial Engineering and Engineering Management

基　　金：国家社科基金一般资助项目(12BJL014;10BJL008);教育部人文社科青年基金资助项目(10YJC790270);山东大学2011年自主创新基金人才引进与培养专项资助项目(2011TB006)

摘　　要：本文以1996—2009年期间上海A股市场为对象研究规模效应,发现规模较小组合获得更高收益的趋势比较明显。然而,某些年份也表现出反向的规模效应。与常规研究方法不同,本文还沿着Dimson和Marsh(1999)、Mills和Jordanov(2003)等人开创的方向,使用马尔科夫链方法从可预测性角度对规模效应进行研究。研究发现:与规模较小的组合相比,规模较大的组合容易拒绝随机游走假设,从而更具有可预测性。可见,规模效应虽然在市场中存在,却有着与传统上不同的另外一种表现。对于中国资本市场出现的这种规模效应及其反转现象,我们认为其根本原因在于机构投资者的操纵行为和缺乏做空机制。There are many anomalies in the capital market,such as January effect,weekend effect,P/E effect,and size effect.These anomalies enable investors to obtain more abnormal returns via a particular investment strategy than a marketing strategy.Size effect is one prevalent anomaly problem.It refers to the regular negative correlation between stock returns and firm size; that is,small company＇s stock has a high rate of return,but large company＇s stock has a lower yield.In order to understand the anomaly problem,this paper uses the database from Shanghai Stock Exchange and studies the size effect from three aspects：i） verify whether there exists the size effect; ii） inspect another expression form of size effect from the predictability aspect; and iii） explore the reason of size effect and its reversal phenomena.Unlike the previous research literature,this paper has many characteristics.First,the sample size and time series span are extended so that the research result has a stronger credibility.Second,different from most of the literature,the paper follows the direction created by Dimson and Marsh （1999）,and Mills and Jordanov （2003）,and investigates the size effect from the predictability aspect.Our findings show that large size portfolio is easy to reject the random walk hypothesis.Therefore,the predictable finding can help verify the size effect＇s performance is different from that of traditional forms.Third,we do not use the traditional autocorrelation test because the return series of financial assets generally don＇t meet linear and normal hypotheses.Instead,we use Markov chain methods that don＇t need those assumptions to avoid errors in judgment.In section 1,we investigate the size effect using the Shanghai Stock Exchange database for the period 1996-2009 and find that for both adjusted and unadjusted risks,small firms are prone to earning greater returns than large firms in the following years：1998,2000,2004,2007,2008 and 2009.However,some years,such as 1996 and 2003,also show the rev

关 键 词：规模效应 可预测性 随机游走假设 马尔科夫链 

分 类 号：F832.5[经济管理—金融学]