Multiple Alpha Sources and Active Management

Abstract

Active portfolio managers often see themselves as assemblers of efficient portfolios that maximize information ratios through the value-added of their proprietary investment insights (alphas), but approaches to reconciling often-contradicting investment insights differ widely. A linear combination of different investment insights can be used to support an analytical framework that maximizes the information from one single source. Simulations based on this framework reveal the impact of varying two key variables the cross-sectional correlation of alpha signals and the time series correlations of the information coefficients related to alpha signals. The latter plays a more important role, providing diversification benefit to strategy risk over time. This analytical foundation has applications in sensitivity analysis and improving active performance.