Concept · The RiskModels engine (ERM3)
Vasicek shrinkage
L2/L3 betas shrunk toward the log-mcap-weighted peer mean within a 4-digit industry cohort; short-history and small-caps are pulled more. Market beta (L1) is not shrunk.
In depth
L2/L3 betas are shrunk toward the log-mcap-weighted peer mean within a 4-digit industry cohort; short-history and small-caps are pulled more, mega-caps barely. Market beta (L1) is NOT shrunk — shrinking it hurt residual variance (41% win rate).
Formula
β_pub = (1−α)·β_Huber + α·β_peer, α = σ²ᵢ / (σ²ᵢ + τ²)
In the methodology
Beta estimation: Huber + Vasicek peer adjustment →Referenced by (2)
- The Industry Beneath the Index
Vasicek peer-β cross-sections expose what sector ETFs paper over
- Cascade Hedging and the Cost of Interpretability
Subsector ETF value, joint optimization, and executable hedge layers across 9,074 US mutual funds