【摘要】 在Kosowski等提出的Bootstrap方法的基础上，以非零的超额收益为假设前提，评价中国市场上基金的业绩表现并进行基金优选.基于2009年3月—2020年6月偏股混合型基金的数据，本文发现中国基金市场整体表现为负的超额收益率.在此基础上，以非零的超额收益为前提，发现Kosowski等提出的Bootstrap方法能够优化基金筛选.最后，在样本外区间、不同市场状态、与晨星基金评级比较以及考虑动态市场超额收益情况下，均发现在非零假设前提下能够挑选出更多显著具有能力的基金经理管理的基金，并且这些基金组合的表现更佳.该结果表明：对于掌握Bootstrap技术的机构投资者，可利用本文提供的方法实现基金优选并以此构建FOF组合为投资者带来更高的收益；对于不具备Bootstrap技术能力的散户投资者，通过聚焦普通回归方法下的超额收益的t值，并选取排序前5%的基金可以较为稳妥地实现基金优选并获得超额收益.更多还原

【Abstract】 This paper proposes a Bootstrap method to select mutual funds by extending the zero alpha assumption of Kosowski et al.(2006). Using a sample of hybrid equity funds from 2009 to 2020, this paper finds that the overall Chinese fund market shows negative abnormal returns. Thus, the traditional zero alpha assumption is not suitable for evaluating fund performance on the Chinese market. Further, this paper uses a non-zero alpha assumption and conducts the Bootstrap method to re-evaluate fund performance. The empirical results show that compared with the traditional Bootstrap method, under the non-zero alpha assumption, Bootstrap method picks out mutual funds that perform better. The above results are still robust in the out of sample test, after considering different market conditions, comparing with the Morning Star top rating funds and taking time-varying alpha into consideration. The results suggest that for institutional investors who can master the Bootstrap technique, the method proposed in this paper is helpful to construct Fund of Fund(FOF) portfolios that perform better. For retail investors who cannot conduct Bootstrap analysis, focusing on the top 5% t-statistic ranking funds from the normal regression analysis can also achieve higher abnormal returns.更多还原