基于参数二次优化的投资组合选择的再平衡策略之模型创建与实证检验  

Model construction and empirical test of rebalancing strategy for portfolio selection based on parametric quadratic programming

作　　者：齐岳[1,2] 黄佳宁 张喻姝 刘彤阳 QI Yue;HUANG Jianing;ZHANG Yushu;LIU Tongyang(Business school,Nankai University,Tianjin 300171,China;China Academy of Governance,Nankai University,Tianjin 300171,China;Economics school,Tianjin Normal University,Tianjin 300382,China;Accounting school,Tianjin University of Finance and Economics,Tianjin 300222,China)

机构地区：[1]南开大学商学院,天津300071 [2]中国公司治理研究院,天津300071 [3]天津师范大学经济学院,天津300382 [4]天津财经大学会计学院,天津300222

出　　处：《中国软科学》2024年第S1期244-262,共19页China Soft Science

基　　金：天津市高等学校研究生教育改革研究计划之重点项目“‘科教兴国战略’下基于金融科技改革财务金融课程群的教学与实践”(TJYGZ43)。

摘　　要：发展金融强国,实现中国式金融现代化,促进金融业加快发展新质生产力,需要提高金融的配置效率,增强金融风险管理水平。资产价格变动会导致投资权重偏离投资策略,增加金融风险,而再平衡可以将实际投资权重调整回既定策略,因此实施再平衡对实现金融风险管理具有重要意义。目前我国再平衡理论研究较少,学者多将均值方差模型作为再平衡的资产配置基础。传统的二次优化方法难以反映有效边界精确、完整的结构。参数二次优化作为唯一完整呈现有效边界的算法,具备优于传统方法的精准性。故创新地使用参数二次优化,针对有效边界提出3种再平衡策略,分别调仓至有效边界最小方差点、三分之一处、非劣点。这个策略能够精确获取组合再平衡后在有效边界上的位置及权重,进而准确计算再平衡绩效。同时构建3种传统再平衡策略(基于简单多样化、市值加权和价格加权的策略)作为绩效比较基准。全面选取我国主板2007—2021年的数据,计算在5、10、20、50、100、150、200、250只股票数量和1、3、6、12个月的调仓周期下6种再平衡策略的绩效并实证对比。研究首次发现,组合规模较小或调仓周期较短时,调仓至有效边界最小方差点的策略绩效相对其他策略更优,简单多样化策略绩效较为稳定。研究得到优于简单多样化策略和其他传统策略的再平衡策略,结论符合投资组合选择理论,也符合我国市场波动性较强的实际情况,首次为个人投资者和公募基金提供根据股票规模和调仓周期选取再平衡策略的建议,为保持实际投资风险可控提供参考。To support the construction of a strong financial industry through a Chinese path to modernization,and accelerate the development of new quality productive forces,we need to improve the efficiency of financial allocation and enhance the level of financial risk management.In practice,the real investment weight may deviate from the established strategy due to asset price changes,and investors need to adjust the actual investment weight back to the established strategy through portfolio rebalancing.At present,there are few studies on rebalancing in China.Most scholars use mean-variance model as the basis of asset allocation for rebalancing.The traditional optimization method is difficult to reflect the accurate and complete structure of the effective frontier.As the only method that presents the effective frontier completely,parametric quadratic programming has the accuracy which traditional methods do not.Parametric quadratic programming can calculate the complete structure of the efficient frontier,and the core advantage is that it can accurately and quickly calculate the specific weight of the kinks on the efficient frontier,as well as the function of each curve on the frontier.Therefore,the paper innovatively uses parametric quadratic programming to propose three rebalancing strategies for the effective frontier,which are adjusted to the minimum variance point,one-third of the effective frontier and the non-inferior point respectively.Different from the traditional methods,this paper can clearly and intuitively present where we need to adjust the portfolio through rebalancing.We are able to accurately calculate the changes in stock weights before and after rebalancing,and obtain accurate rebalancing results.Moreover,three traditional rebalancing strategies(based on naive diversification,market value weighted and price weighted strategies)are constructed as performance comparison benchmarks.This paper comprehensively selects all the relatively complete data available in China’s stock market,that is,1164 stocks

关 键 词：投资组合再平衡 参数二次优化 交易成本 有效边界 

分 类 号：F832.5[经济管理—金融学]