Evaluating Investment Performance: The p-index and Empirical Efficient Frontier

Battle Count: 74

The paper introduces a novel risk metric (p-index) derived from option pricing theory that can be used for portfolio construction, risk-adjusted performance measurement (p-ratio), and as an additional factor in multi-factor models. The EEF provides a practical framework for portfolio selection. However, the extremely high reported returns for SSE strategies raise practical concerns about implementability. The p-index as a sixth factor shows incremental information for NYSE but not SSE, limiting its universal applicability. The weekly rebalancing strategies could inform short-term trading systems, but transaction costs are not considered.

The core methodology involves: (1) estimating binomial parameters from weekly high/low prices, (2) computing put option prices via risk-neutral valuation, (3) calculating p-index and p-ratio, (4) solving linear programming problems for EEF construction, and (5) running cross-sectional/time-series regressions. The linear programming and regression components are standard. The main complexity lies in the binomial parameter estimation from observed prices, handling edge cases (pi outside [0,1]), and the price limit adjustments for SSE stocks. No code or repository is provided.

The paper provides detailed formulas for p-index construction, EEF linear programming models, and regression specifications. Data sources are identified (CSMAR for SSE, Investing.com/Yahoo Finance for NYSE, Kenneth French's data library for US factors). However, the exact stock selection criteria, handling of delisted stocks, and full code are not provided. The binomial model parameter estimation from weekly high/low prices is described but some edge-case handling (e.g., when pi is outside [0,1]) is only briefly mentioned.