Seeing Risk Before It Spikes: Volatility Regime Change and the Governance Challenge for Investment Committees

Volatility changes character before it changes magnitude. In this article, we argue that standard risk measures such as levels, thresholds, and scenarios miss the diagnostic signals that precede regime change: rising return autocorrelation, increasing volatility of volatility, unstable cross-asset correlations, and asymmetric market responses. We propose a character-versus-level decomposition that ranks these indicators by informativeness and apply it to the 2008 financial crisis and the 2015 Chinese devaluation scare. The character indicators distinguished the genuine regime change from the false alarm. We address the central objection—that false alarms make early action prohibitively costly—through a graduated-response framework illustrated with a stylized 2007 example. The framework highlights how investment committees can incorporate early diagnostic signals into governance and risk-management processes without overreacting to temporary market volatility. A mathematical appendix formalizes the signal detection problem and shows that the optimal response scales continuously with signal strength.