The Frisch-Waugh-Lovell theorem is the algebraic basis for residualizing one factor against prior factors before estimating its incremental contribution.
References · literature map
Literature behind hierarchical attribution
A selective bibliography for the RiskModels.org research program. The emphasis is not citation volume; it is the lineage behind orthogonal decomposition, factor exposure measurement, residual risk, and model-governance discipline.
Orthogonal Decomposition
Foundational tools for separating explanatory layers without double-counting shared covariance.
Completes the modern FWL framing used for partial regressions and orthogonalized explanatory variables.
Klein, L. R., & Chow, G. C. (1953). The use of econometric models in economic control. Econometrica, 21(2), 201–215.
Early applied econometric control logic that motivates decomposing system behavior into interpretable model components.
Factor Models
Canonical equity factor literature behind systematic exposure measurement and its limits.
Introduces the market beta concept that remains useful but insufficient for total risk attribution.
Establishes the empirical multi-factor lens that motivates decomposing returns beyond the market factor.
Adds momentum to the factor toolkit and demonstrates how apparent skill can be absorbed by systematic exposures.
Idiosyncratic Risk and Active Share
Literature on the residual and active components that beta-only reporting can obscure.
Active Share is a central complement to tracking-error analysis when evaluating whether reported active management is economically meaningful.
Shows that idiosyncratic volatility is empirically important and not safely ignored in expected-return or risk analysis.
Highlights how much firm-level variation remains unexplained by common factors, directly motivating residual-risk diagnostics.
Factor Zoo and Model Governance
Warnings about overfitting, factor proliferation, and the need for disciplined model interpretation.
Frames the factor-zoo problem and the need to distinguish economically durable structure from statistical artifacts.
Raises the multiple-testing bar for claimed factors and reinforces the need for parsimonious, governed attribution systems.
Demonstrates how many proposed anomalies weaken under replication, arguing for transparent factor selection and validation.