以开放式证券投资基金申购、赎回行为背后的行为金融机理为研究对象,通过分析相关变量作用于基金投资者心理的方式和路径,探讨其对基金申购、赎回所产生的影响。在理论分析的基础上,提出基金业绩、规模、存续时间、价位、分红等变量能够各自基于信号传递、心理账户、预期框定偏差、预期惯性、处置效应等多元化路径影响投资者心理,并作用于基金申购、赎回和基金流量,并据此提出了相关假设。进一步地,通过建立结构方程模型,以我国327只股票型开放式证券投资基金为样本,对2011至2013年三组年度截面数据进行了检验。实证结果表明信号传递、预期框定偏差路径始终稳定存在,预期惯性路径仅在2011、2013年实证检验中存在,处置效应路径只在2012年检验中存在、且呈现非对称性特征,而心理账户路径则无法得到证实,据此得出结论认为基金业绩对基金申赎影响最为显著但不具稳定性。最后,从投资者行为模式和基金市场发展环境等角度对实证结果进行了分析。
Focusing on the behavioral finance mechanism behind the purchase and redemption behavior of open- ended securities investment funds, the pattern and the path of related variables impacting fund investor psychology are studied in this paper to probe into their influence on the funds purchase and redemption. On the basis of theoretical analysis, the authors propose that the performance, scale, duration period, price and dividend level of funds variables may have influence on investors psychology and their funds purchase-redemption behaviors through diversification paths as signal transmission, mental account, the expectations framing bias, expectations inertia and disposition effect. The corresponding hypotheses are also proposed on it. With 527 open-ended stock investment funds selected in Chinese funds market as samples, annual cross -sectional data from 2011 to 2013 are tested through the establishment of structural equation model to verify the hypotheses. The empirical results show that the signal transmission and the expectations framing bias paths exist stably, while expectations inertia path exists only in 2011 and 2013. The disposition effect path is also found in 2012 and behaves with the charac- teristics of non-symmetry. As far as the mental account path is concerned, it has not been confirmed or denied in the empirical research. It is also found that the performance of the fund has most significant but unstable effect on fund purchase-redemption behaviors. Finally, the further analysis on the empirical results is also made from the perspective of investors' behavior and the environment of Chinese fund market.