Markowitz vs. 1/N: Portfolio Performance, Estimation Errors, and Subjectivity

Abstract

This paper reconciles the ‘mean-variance (MV) vs. equal-weighting (1/N)’ debate using an unbiased performance-to-error map and delineates thresholds where subjective interventions become essential to surpass MV optimisation. Traditional evaluations often yield inconclusive or contradictory results due to a ‘joint-test’ problem entangled with alpha and risk models. This paper introduces a methodology that isolates portfolio selection, enabling an impartial assessment of performance sensitivity to estimation errors, identification of when MV becomes suboptimal, and exploration of why appropriate subjectivity can enhance performance beyond MV. These findings highlight the necessity of structurally integrating subjectivity and rational investor behaviour into portfolio theory and practice.

Keywords: alpha model, factor risk model, joint tests, Markowitz, Monte Carlo Simulation, investor subjectivity

JEL Classification: C61, C63, D81, G11, G41