Knowing when diversification stops working

Diversification assumes you can combine assets whose behaviour differs, so that the whole moves less than the parts. It only works where one tends not to fall while the other does. Inside crypto, that assumption holds poorly. Assets with different purposes and different designs have repeatedly been observed rising and falling at much the same moments.

Part of the reason is shared plumbing: the same investor base, the same exchanges, the same flows of money. When interest in the sector rises, buying spreads across it; when appetite for risk retreats, selling does the same. Short-term prices are dominated more by money entering and leaving the sector than by any individual project's progress. Splitting into ten names is no defence against that single factor.

Worse, the strength of that linkage is not stable. Prices that looked reasonably independent in calm conditions have been noted to converge on the same direction during sharp declines. In other words, diversification weakens precisely where you most wanted it. Concluding from quiet-period data that a pair 'does not move together' says little about how they behave under stress.

Only a narrow conclusion follows. Diversification does help dilute problems specific to a single asset — abandoned development, misconduct by an issuer, a rule aimed at that asset. It does not prevent losses when the market as a whole falls. Blurring that distinction into 'I am diversified, so I am fine' is how expectations and reality diverge during an actual decline. This site recommends no combination and offers no view on future correlations.