加载中...
Fund Performance Attribution and Investor Behavior
开放式基金是全球最受欢迎的资产管理工具之一,过去十年中国开放式基金也蓬勃发展。截至2016年底,中国开放式基金净值增至8.886899万亿元,占整个基金市场的97.74%。因此,中国投资者在选择开放式基金时最关注什么是一个值得探讨的话题。关于投资者如何做出投资决策的文献主要集中在美国市场。一些研究发现,CAPM alpha比更复杂模型的alpha更能解释美国共同基金资金流。由于中国基金市场仍处于新兴阶段,且微观结构与美国不同,我们感兴趣的是中国投资者如何识别基金经理能力。他们在申购或赎回时最看重基金业绩的哪方面?是风险暴露还是alpha?以及哪种alpha?此外,机构投资者被认为比个人投资者具有更强的专业能力和信息优势,因此在决策时更不易出现非理性问题。那么他们在基金选择时是否与个人投资者行为不同?本文的主要贡献在于试图识别投资者如何在长期内区分基金经理的主动管理能力与基金风险暴露。我们比较了不同风险因子调整后的alpha对基金净现金流的影响以及风险暴露的影响。本文还从新视角为中国基金市场的赎回异象文献提供了补充。我们发现机构与个人投资者在风险调整基准和基金经理能力的评价上存在差异。在实证部分,我们聚焦于2006年1月1日至2016年12月31日期间的64只开放式基金,所有数据来自CSMAR。我们发现当前和滞后的原始超额收益、CAPM alpha、Fama-French三因子alpha和Fama-French五因子alpha对基金净流量均具有较强的解释力。当前alpha与净资金流正相关,而滞后alpha与净资金流负相关,表明滞后alpha对申购决策影响更大,而同期alpha更能解释赎回行为。投资者倾向于在同期基金alpha较高时赎回份额,以实现即时收益。这种行为对应于赎回异象,可能与处置效应有关。总体而言,我们发现原始超额收益和CAPM alpha等简单模型能更好地解释基金净流量变化,这可能是因为中国市场缺乏足够的投资工具。在子样本检验中,我们发现个人投资者在评价基金经理能力时更依赖简单调整,而机构投资者则区分更严格。我们还发现原始超额收益和CAPM alpha的优势主要集中在市场波动较低和投资者情绪较高的时期。此外,我们发现基金流量对风险暴露的敏感度远高于对基金经理能力alpha的敏感度,这与美国市场的发现相反。
Abstract
Open-end funds are one of the most popular asset management tools worldwide and the past decade has also witnessed their blossoming in China. By the end of 2016 the net asset value of open-end funds in China had increased to 8,886.899 billion RMB, gaining a market share of 97.74% in the whole fund market. Therefore, what most influences Chinese investors when selecting open-end funds is a topic worth exploring. The literature on how investors make investment decisions is mainly focused on the U.S. market. Some studies find that CAPM alpha explains U.S. mutual fund flows better than the alphas of more sophisticated models. As the Chinese fund market is still emerging and differs from that of the U.S. in terms of microstructure, we are interested in how Chinese investors identify manager ability. Which aspect of fund performance do they value most when making decisions on buying or redeeming? The risk exposure or the alpha? And which alpha? In addition, institutional investors are believed to have greater professional skill and information advantage compared with individual investors, and thus to be less prone to irrational issues when making decisions. Do they therefore behave differently from individual investors during fund selection? The main contribution of this paper lies in seeking to identify how investors distinguish the fund manager's active management ability from fund risk exposure over a long period. We compare the impact of alphas adjusted by different risk factors on the net cash flow of funds and the influence of risk exposure. Our paper also supplements the literature on the redemption anomaly in China's fund market. We find differences between institutional and individual investors in the evaluation of risk adjustment benchmarks and fund manager ability. In the empirical section, we focus on 64 open-end funds lasting from January 1, 2006 to December 31, 2016, drawing all data from CSMAR. We find that the current and lagged raw excess return, CAPM alpha, Fama-French 3-factor alpha and Fama-French 5-factor alpha all have strong explanatory power for net fund flows. The current alphas are positively correlated with net capital flow, whereas the lagged alphas are negatively correlated, indicating that the lagged alpha has a greater effect on buying decisions whereas the contemporaneous alpha is more responsible for redeeming behavior. Investors tend to redeem their shares when the contemporaneous fund alpha is high, perhaps to realize immediate gains. Such behavior corresponds to the redemption anomaly, which may be related to the disposal effect. Overall, we discover that simple models such as the raw excess return and CAPM alpha can better explain changes in net fund flows, which may be due to a lack of sufficient investment tools in the Chinese market. In subsample tests, we find that individual investors rely more on simple adjustments in evaluating a fund manager's ability while institutional investors make stricter distinctions. We also find that the advantages of raw excess return and CAPM alpha are mainly concentrated in periods of lower market volatility and higher investor sentiment. In addition, we find that fund flows are far more sensitive to risk exposure than to manager ability, contrary to the findings in the U.S. market.
当前alpha与净资金流正相关,滞后alpha负相关
基于2006-2016年64只开放式基金数据,实证发现当前CAPM alpha、FF3 alpha和FF5 alpha与基金净流量正相关,而滞后一期alpha与净流量负相关。表明投资者申购时更关注历史alpha,赎回时更关注当前alpha。
简单模型(CAPM alpha)比复杂模型更能解释资金流
比较原始超额收益、CAPM alpha、FF3 alpha和FF5 alpha对基金净流量的解释力,发现原始超额收益和CAPM alpha的回归系数更大、拟合优度更高。说明中国投资者在评价基金经理能力时更依赖简单风险调整模型。
基金流量对风险暴露的敏感度远高于对alpha的敏感度
回归结果显示,基金净流量对市场风险暴露(beta)的敏感度显著高于对alpha的敏感度。投资者更关注基金风格和风险暴露而非基金经理的主动管理能力,这一发现与美国市场相反。
个人投资者与机构投资者在评价基金经理能力上存在差异
子样本检验发现,个人投资者更依赖简单模型(如原始超额收益和CAPM alpha)来评价基金经理能力,而机构投资者则能更严格地区分不同风险调整基准的alpha,表现出更强的专业能力。
简单模型的优势集中于低波动、高情绪时期
分时期检验表明,原始超额收益和CAPM alpha对基金净流量的解释优势主要集中在市场波动较低和投资者情绪较高的时期。在高波动或低情绪时期,复杂模型alpha的解释力相对提升。
核心解释变量
基金业绩指标:原始超额收益、CAPM alpha、Fama-French三因子alpha、Fama-French五因子alpha(当前和滞后一期)
被解释变量
基金净流量(flow),定义为基金净值增长率调整后的资金净流入
样本与数据
2006年1月1日至2016年12月31日期间64只中国开放式基金,共8448个基金-月度观测样本,数据来源于CSMAR数据库
识别方法 / 模型设定
采用Fama-MacBeth横截面回归和面板固定效应模型,分析不同风险调整alpha对基金净流量的影响,并引入风险暴露、基金特征等控制变量
内生性及稳健性检验
使用不同子样本(个人与机构投资者、高/低波动时期、高/低情绪时期)进行稳健性检验;采用Fama-MacBeth回归控制时间效应;替换alpha估计窗口和基准模型。
更多相关数据正在补充