Diversify your crypto portfolio: what automated indices change

Coin Encyclopedia Improve your knowledge of cryptos

This article contains affiliate links allowing you to support the daily work of the Journal Du Coin teams.

We rarely enter crypto with an allocation plan. We buy bitcoin, then a little ether, sometimes a fashionable asset.

The market itself is pushing in this direction. As of October 16, 2026, bitcoin represented nearly 58% of a total crypto capitalization of approximately $2,860 billion (CoinGecko).

Since Markowitz (1952), we have known that the risk of a portfolio is not limited to the average risk of its components: it also depends on the correlations between them. Combining assets that do not move together reduces overall volatility.

The nuance is important. Diversification reduces the specific risk, that of a project that fails, a hacked protocol or a token that loses its liquidity. It does not reduce market risk. In a crash, correlations between cryptoassets rise sharply: almost everything falls at the same time. Holding ten cryptos instead of two therefore protects little.

To really reduce overall risk, you have to get out of crypto: gold, major stock indices or certain technological stocks do not always follow the same cycle.

Choosing your assets is only half the battle. You then have to decide on their weight. Three approaches dominate.

Academic research invites us to remain humble in the face of sophisticated models. DeMiguel, Garlappi and Uppal (2009, Review of Financial Studies) showed that none of the 14 optimization models tested on 7 datasets consistently beat the naive 1/N rule. Plyakha, Uppal, and Vilkov (EDHEC, 2012) observed that, over four decades of U.S. stocks, an equal-weighted portfolio rebalanced monthly beat cap-weighting in return and Sharpe ratio, albeit with higher risk.

This work focuses on stocks, not crypto. They only remind us that a simple rule, applied rigorously, is often better than fragile optimization.

Machine-translated from French