cross-asset · regime-classification · python
RoRo — Risk-Regime Switching Model
Cross-sectional risk-on/risk-off classifier from the realized return-vs-volatility slope across 64 country index series — segmented into 10 cuts and validated against FRED proxies.
Overview
A daily, reproducible, academically-anchored classifier of the global risk-on / risk-off (RoRo) state. Instead of reading a single proxy like the VIX or HY OAS, RoRo fits a weighted regression of trailing realized return on trailing realized volatility across the entire investable cross-section of 64 country indices, and interprets the slope as the market-wide price of risk.
What it does
- Cross-sectional realized SML slope estimated daily under cap-weighted and equal-weighted schemes.
- Ten economically meaningful segments: global, DM, EM, Equity, FI, DM/EM × Eq/FI, LatAm bloc.
- Tercile classification against the trailing 5-year distribution → Risk-on / Transitional / Risk-off.
- Parallel correlation-structure signal (PC1 variance share, average pairwise correlation) with disagreement-event flagging à la Beber, Brandt & Kavajecz (2013).
- External validation: rolling 60-day correlation vs five FRED proxies (VIX, BBB OAS, EM Corp OAS, HY OAS, 2s10s) with degradation alerts.
- Deterministic functional pipeline, byte-identical CSV artifacts, interactive HTML report.
Stack
Python · pandas · NumPy · FRED API · Plotly.
Status
v1.0 — diagnostic only (no predictive layer). Open source; proprietary panel git-ignored, free FRED proxies vendored.