InsidersTradesSigma
Performance & Risk Metrics
The statistical variance in a security's systematic risk coefficient (beta) measured across successive rolling time windows, signaling regime shifts or structural breaks in market sensitivity that undermine predictive signal reliability.
Rolling beta instability arises when a stock's beta swings materially from one estimation window to the next (typically 30, 60, or 90 trading days). When the relationship between a stock and the index keeps breaking down, factor-neutral positioning becomes unreliable and outperformance is hard to attribute to genuine alpha rather than a passing regime. Thinly traded or tightly held names are most prone to it.
It is usually quantified as the coefficient of variation (standard deviation over mean) of the rolling beta estimates, or by testing for structural breaks with a Chow test. A high score warns that any model relying on a stable beta, factor-neutral hedges, sector-momentum overlays, will be working off a moving target and should re-estimate more often.