Residual-Free Attribution
针对透明因子策略的需求，本文证明约束多因子均值方差优化的解可完全分解为底层约束因子组合，实现理论上合理的无残差业绩归因，并揭示投资约束对最优组合的影响。
In response to the need for transparent factor-based, long-only strategies, investors have devised distinct residual-free attributions designed to leave little active return unexplained. The lack of information on how most of these operate is a problem. In this article, the author demonstrates that the solution of the constrained multifactor mean–variance optimization can be fully decomposed into a combination of underlying constrained factor portfolios that allows for a theoretically sound residual-free attribution of performance. The proposed decomposition also provides a unique perspective of the impact of investment constraints on the optimal portfolio.