一种增强型指数追踪模型设计及应用

研究目标：构建了可以调节追踪误差和超额收益的增强型指数追踪模型，并给出了广义最小角度回归算法（GLARS），用以计算调节参数作用下模型解的折中路径。研究方法：通过模拟数据和五组世界主要股票市场指数的历史数据，对本文提出的模型和算法与同类模型和算法进行了性能比较；同时追踪上证50指数构建若干稀疏且稳定的资产组合模型，通过信息比率等指标对投资组合进行评价。研究发现：本文构建的模型可用以构造权衡追踪效果和超额收益，且稀疏的资产组合，GLARS算法相对传统预设参数的算法具有良好的求解能力和计算速度。研究创新：引入调节参数平衡追踪效果和超额收益，并针对中国股票市场的特点，在增强型指数追踪模型施加非负约束；GLARS算法可遍历所有折中意义下的最优解。研究价值：本文提出的增强型指数追踪模型在国内具有较强适用性，在保证资产稀疏性的前提下可以得到超额收益，同时丰富了目前投资组合中的方法论研究。

A New Enhanced Index Tracking Model and Its Application

Research Objectives: This article constructs an enhanced index tracking model which can control the trade-off between tracking errors and excess returns by tuning parameters, and proposes generalized least angle regression (GLARS) for computing the trade-off path of model coefficients estimator. Research Methods: This article compares the performance of the proposed model and algorithm against relative ones through simulated data and historical data of 5 major stock indices worldwide. Several sparse and stable portfolios are then constructed tracking SSE 50 Index and evaluated with criteria such as information ratio. Research Findings: The model constructed in this article can be applied in building up sparse portfolio trading off between tracking performance and excess return. GLARS is found to outperform regular preset parameter algorithms in solution finding and computation speed. Research Innovations: This article introduces tuning parameters for balancing tracking performance and excess return. A non-negative constraint is applied to the enhanced index tracking model in consideration of the features of China stock market. GLARS can traverse all potential optimums in the trade-off context. Research Value: The enhanced index tracking model proposed in this article is widely feasible in China stock market. It helps obtain excess return under portfolio sparsity, enriching the current methodology studies on portfolio building