Cross-Asset Volatility Spillover

Asset classes don't panic in isolation. A sharp volatility spike in one market — say a sudden equity selloff — tends to be followed, often within hours or days, by rising volatility in bond yields, currencies, and sometimes commodities too, even when there's no obvious fundamental reason those markets should be linked at that moment. This tendency is called volatility spillover, and it's driven by mechanisms like shared investors deleveraging across asset classes at once, margin calls forcing sales of unrelated positions to raise cash, and algorithmic strategies that react to volatility in one market by cutting risk in others.

The practical consequence is that correlations between assets are not stable — they tend to rise sharply during periods of market stress, precisely when investors most need diversification to work. A portfolio built on historical correlations measured during calm periods can understate how much its components will move together once volatility spikes, because the spillover channel activates specifically under stress and is largely invisible in quiet markets.

Researchers measure spillover using tools like variance decompositions in vector autoregressions, tracking how much of one market's volatility forecast error is explained by shocks originating in another market. For a risk manager, the simpler takeaway is to stress-test a portfolio using volatility and correlation assumptions drawn from past crisis periods, not calm-period averages, since spillover systematically breaks diversification exactly when it's needed most.