Correlation Is Not Constant: Crypto Co-Movement Across Regimes

Portfolio construction leans heavily on correlation. If two assets do not move together, holding both reduces risk. In crypto, the largest assets are usually positively correlated, which leads many allocators to treat the complex as a single exposure. That is a reasonable first approximation — but the strength of co-movement changes meaningfully through time, and those changes are exactly what risk models need to capture.

Why co-movement rises in stress

During broad deleveraging, participants sell what they can, not only what they want to. Liquidations on leveraged venues cascade across assets, market makers reduce inventory across the board, and idiosyncratic stories matter less than the common factor of risk appetite. The result is that correlations tend to rise precisely when diversification would be most valuable.

Measurement choices

• Use log returns sampled at the same timestamps for every asset to avoid spurious decorrelation.
• Daily closes are robust; intraday sampling reveals more but introduces asynchronous-trading bias.
• Report the window length alongside every correlation figure — it changes the answer.
• Correlation captures linear co-movement only; tail dependence can be stronger than correlation suggests.

When periods of idiosyncratic news dominate — a protocol upgrade, a network outage, an ecosystem-specific flow — correlations can fall for a time. Tracking a rolling measure alongside the long-run average makes those windows visible, and helps separate genuine diversification from a temporary break in the common factor.

This publication is provided for informational purposes only and does not constitute investment, legal, or tax advice, or an offer or solicitation to buy or sell any asset. Live figures are computed from third-party public market data and may be delayed, incomplete, or inaccurate.