Returns-Based Style Analysis (RBSA), introduced by William Sharpe (1992), estimates a portfolio's exposures to major asset classes or style factors by running a constrained regression of the portfolio's returns on a set of passive benchmark index returns:
R_portfolio ≈ Σ_k (w_k × R_factor_k) + ε
with constraints that weights w_k ≥ 0 and Σw_k = 1 (they sum to 1, consistent with a long-only allocation). The fitted weights represent the portfolio's inferred style exposures. The residual ε represents selection skill — return not explained by factor exposures alone.
Strengths and limitations
- Advantage — only requires a return history, not portfolio holdings data. Useful for analyzing funds where holdings are disclosed infrequently (e.g., quarterly).
- Limitation — style exposures are estimated over a rolling historical window and may lag rapid tactical shifts in actual positioning.
- Limitation — the non-negative constraint forces all exposures to be long-only. Strategies with significant short exposure require unconstrained RBSA variants that relax this constraint.