How do I do factor-based performance attribution?

Question

AcadiFi · Accepted Answer

Factor-based attribution decomposes excess return into systematic factor contributions and residual (stock selection/alpha). For each factor k: contribution_k = β_k × factor_return_k. Sum the contributions and the residual equals total active return.

Example: Juniper Core Fund returned 14.20% over a year, benchmark 11.50%, so excess = 2.70%. Regressing excess returns monthly on FF5+UMD:

| Factor | Loading β | Factor Return | Contribution |
|---|---|---|---|
| Market | 0.05 | 9.8% | 0.49% |
| SMB | 0.22 | 2.1% | 0.46% |
| HML | -0.15 | -1.8% | 0.27% |
| RMW | 0.31 | 3.4% | 1.05% |
| CMA | 0.09 | 0.8% | 0.07% |
| UMD | 0.18 | 4.2% | 0.76% |
| Sum factors | | | 3.10% |
| Residual (α) | | | -0.40% |

Factor loadings contributed 3.10% while stock selection subtracted 0.40%. The manager's outperformance came almost entirely from being tilted toward profitability, momentum, and small-caps — not security picking. This is crucial for manager evaluation: you might be paying active fees for factor exposures available cheaply via smart beta ETFs.

Attribution statements are reported monthly to institutional clients. Persistent positive residual over 3+ years with t-stat above 2 is the bar for skill.

How do I do factor-based performance attribution?

Master Level III with our CFA Course

Related Questions

Practice Questions