Turbulence Index
A Mahalanobis-distance regime indicator: how far today's cross-asset return vector is from its trailing covariance-implied normal.
The turbulence index measures the multivariate unusualness of today's cross-asset returns. Mathematically it is a Mahalanobis distance:
T_t = (r_t − μ)ᵀ Σ⁻¹ (r_t − μ)
where r_t is today's vector of factor returns, μ is the trailing 252-day mean vector, and Σ is the trailing 252-day covariance matrix. Low values mean today looks like a normal day in recent history. High values mean today is unusual relative to the joint distribution — the move might be small per-asset, but the combination is anomalous.
Why it complements VIX
The VIX measures a single-asset volatility expectation. It misses regime change that shows up in cross-asset correlations — for example, equity–bond co-movement flipping sign, or dollar strengthening on the same day stocks fall. Turbulence captures these multivariate shifts. Empirically, turbulence spikes have led VIX spikes by days to weeks at major regime changes.
How MarketGrep displays it
The Risk Radar tab plots the turbulence index over selectable look-back windows (30D / 1Y / 3Y / max) with a regime-band overlay. The current reading is annotated with the level (calm / elevated / stressed) and the day-on-day change.
Source
Concept due to Kritzman & Li, "Skulls, Financial Turbulence, and Risk Management," Financial Analysts Journal, 2010. The MarketGrep implementation uses a rolling 252-day window over a curated factor basket (US large-cap, small-cap, Treasuries, USD, gold, oil).
Related
MarketGrep is a market-environment dashboard. Every signal on this page is descriptive, not prescriptive — there is no buy or sell recommendation anywhere on the site. Data is for informational purposes only and may be delayed, inaccurate, or revised.