During periods of market stress, realized volatility exhibits explosive self-reinforcing feedback: rising volatility triggers further selling, which amplifies volatility further. We show that this process is exactly captured by sigmadot = sigma exp (lambda sigma), where sigma denotes normalized volatility (VIX/100) and lambda > 0 encodes feedback strength. This is precisely the Fₗambda class introduced in a previous work, which is shown there to be the unique family of the form udot = g (u) Phi (lambda u) whose associated function space is stable under both natural derivation operators — and the unique such family admitting a transcendental non-elementary linearizer, namely the exponential integral E1. Exploiting this exact linearization, we derive: a closed-form time to volatility explosion T*crash = E1 (lambda sigma0), a universal scaling law T* ~ -ln lambda as lambda -> 0 independent of initial conditions, an exact analytical boundary in (lambda, sigma0) space separating stable and crash regimes, calibration on historical VIX crises (2008, 2020, 2018) yielding lambda values consistent with market stress levels, a PDE model of spatial cross-asset contagion, (Pₗambda^+), with exact solution via Gaussian convolution and a global existence theorem, and a computable real-time early warning indicator T*ᵣemaining (t) = E1 (lambda sigma (t) ). The canonicity of Fₗambda — forced by algebraic structure, not chosen by convention — provides a principled foundation for this framework that is absent from standard GARCH and stochastic volatility models.
Judicael Brindel (Fri,) studied this question.