The Empirical Signatures of Crisis: Spectral and Directed Diagnostics of a: Recovered Dependence Operator

When an economy tips from calm into crisis, what changes? This paper argues that what changes is the web of statistical dependence among its assets: their movements bunch together onto a single dominant pattern, transmission between them turns more one-way, and the system as a whole becomes harder to predict. Working from the joint day-to-day movement of returns, never from raw price levels, the paper recovers a map of that dependence and grades three signatures of crisis by how firmly the data pin each one down. Of the three, the concentration of movement is the one the data pin down firmly. The one-way flow shows up only as a broad contrast between calm and crisis, and the paper declines to name the direction of any single link, because the data cannot settle it. These signatures diagnose a crisis as it happens rather than forecast it: knowing today's pattern of joint movement does not reduce the forecasting error that no method can escape. A worked example across a calm period, the 2008 crisis, and the pandemic shows the method in action.