Performance Attribution and Investment Alpha › Summary

This chapter developed two complementary frameworks for understanding the sources of portfolio performance:

Brinson attribution decomposes active returns into allocation, selection, and interaction effects based on portfolio holdings and sector weights. Allocation effect measures value added from over- or underweighting sectors, while selection effect captures the contribution from security picks within each sector. The interaction effect accounts for the combined benefit of both overweighting a sector and outperforming within it.

Factor-based attribution uses regression analysis to separate returns into systematic factor exposures (beta) and residual alpha. This approach uses the factor models developed earlier in Part IV to determine how much return comes from market exposure, size tilts, value tilts, and other systematic factors versus stock selection skill.

The distinction between alpha and beta lies at the heart of investment management. True alpha, returns uncorrelated with known factors, represents investment skill and justifies active management fees. However, alpha is model-dependent, and statistical significance requires long track records. What appears as alpha under simple benchmarks often becomes beta when analyzed against richer factor models.

Practical applications include manager evaluation, investment process improvement, risk attribution, and fee negotiations. Challenges include data quality, benchmark selection, factor model specification, and the fundamental difficulty of distinguishing skill from luck given the noise in financial returns.

Performance attribution has driven significant industry changes, including the growth of factor investing and increased scrutiny of active management fees. While backward-looking and subject to model specification choices, attribution remains essential for understanding investment performance and making informed allocation decisions.