Crypto Correlation: Why a Diversified Portfolio Can Still Move Together

Crypto correlation explains why a diversified portfolio can still fall as one. Learn to calculate it from weekly returns, read the numbers, and see what helps.

You can own eight coins from eight different projects and still watch every one of them drop on the same day. That isn’t bad luck. It is correlation, and it decides how much protection a longer list of holdings really gives you.

What does correlation measure?

Correlation measures how consistently two assets’ returns move together, on a scale from −1 to +1:

• +1: they always move in the same direction, in proportion.
• 0: knowing one tells you nothing about the other.
• −1: they always move in opposite directions.

Two details trip people up. Correlation is calculated from returns, not price levels, because two prices that both trend upward will look related even if their weekly moves aren’t. And it measures direction and consistency, not size: two coins can be highly correlated while one swings twice as far as the other.

FINRA lists correlated holdings as one way concentration risk creeps in unnoticed, because investments that share a driver can fall together even when they look different on paper.

Worked example: calculating correlation from weekly returns

Suppose two hypothetical coins post these weekly returns:

The steps, which a spreadsheet’s CORREL function does in one go:

1. Average each column: Coin A = +0.375%, Coin B = +0.125%.
2. For each week, multiply Coin A’s distance from its average by Coin B’s distance from its average. These products sum to 196.1.
3. Divide by the number of weeks minus one to get the covariance: 196.1 ÷ 7 = 28.0.
4. Divide the covariance by both standard deviations (5.84 and 6.23): 28.0 ÷ (5.84 × 6.23) = 0.77.

The coins moved in the same direction in only 5 of the 8 weeks, yet the correlation is 0.77. Weeks 4 and 6, the two biggest moves, supply 77% of the sum in step 2. Leave them out and the correlation of the remaining six weeks drops to 0.45. Correlation is dominated by large moves, which are exactly the moves that decide your drawdowns.

Correlation vs beta: direction vs size

Correlation says how reliably two coins move together, not how far. For size, use beta: the covariance divided by the variance of the reference coin. Here, Coin A’s variance is 5.84² = 34.1, so Coin B’s beta to Coin A is 28.0 ÷ 34.1 = 0.82. In this sample, Coin B moved about 0.82% for every 1% move in Coin A. Two coins can share a high correlation while one has a beta of 2 and swings twice as hard, so check both numbers before treating a pair as interchangeable.

Why does correlation limit diversification?

For two holdings with weights w₁ and w₂, volatilities σ₁ and σ₂ and correlation ρ:

Portfolio volatility = √(w₁²σ₁² + w₂²σ₂² + 2·w₁·w₂·ρ·σ₁·σ₂)

Take a 50/50 mix of a coin with 60% annual volatility and one with 90%. At a correlation of 1, the portfolio’s volatility is the simple weighted average, 75%. Anything below 1 pulls it down:

At 0.9, pairing the two coins removes less than 2 points of volatility. At 0.3 it removes almost 14. This is the same effect that makes the number of coins you hold matter less than you’d expect, and it is why measuring concentration with the HHI can flatter a portfolio: the HHI treats every holding as a separate bet.

How correlated are crypto assets?

The U.S. Financial Stability Oversight Council’s 2022 report found that crypto-asset prices “have tended to be widely correlated with each other, exposing crypto-asset market participants to largely non-diversifiable risk inside the crypto-asset ecosystem.”

Correlations also shift over time, including against other markets. The IMF found that the correlation between daily moves in Bitcoin and the S&P 500 was 0.01 in 2017 to 2019 and jumped to 0.36 in 2020 to 2021. A figure estimated from one period can be a poor guide to the next, which is why a single number is never enough.

How to check your own portfolio’s correlation

1. Download weekly closing prices for each holding over the past 52 weeks.
2. Convert prices to weekly percentage returns.
3. Use CORREL on each pair of return columns to build a small correlation table.
4. Repeat with only the last 26 weeks and compare the two tables. A large gap means the relationship is shifting, so plan around the higher of the two readings.
5. Flag any pair above 0.8 in either window as, in practice, one position.

As a rough reading guide, 0.8 to 1.0 means the pair behaves almost like one holding, 0.5 to 0.8 means mostly together, and below 0.5 means independent enough to matter. When you size positions, count a highly correlated pair as a single bet; position weighting covers how to set those sizes.

What actually diversifies a crypto portfolio?

If most of your coins share the same driver, adding another similar coin does little. What changes total risk is weight, not variety: a smaller share of the most volatile holdings, or a cash-like sleeve that barely moves. Stablecoin allocation shows how a sleeve scales losses down in proportion to its size. Beyond that, the question becomes how crypto fits alongside everything else you own, which is a whole-portfolio decision rather than a coin-picking one.

Common correlation mistakes

• Using prices instead of returns. Two rising price lines tend to look related whether or not their moves are.
• Trusting a short window. Eight weeks, as in the example, is fine for learning the method but far too few for a decision.
• Assuming a low correlation will last. The IMF’s figures show how quickly a relationship can change.
• Counting related tokens as separate bets. Tokens from one ecosystem can share the same driver.

Put your holdings into our portfolio health check for concentration and custody readings, then adjust your view for any highly correlated pairs. The vital signs of a healthy crypto portfolio shows where correlation fits among the other checks.

The bottom line

Correlation is why a portfolio of many coins can behave like a portfolio of one. Calculate it from returns, check it over more than one window, and treat highly correlated holdings as a single position when you judge concentration. Diversification only works to the extent your holdings actually move differently.

This content is for education only and is not financial, investment, tax or legal advice. Crypto assets are volatile and you can lose money. Examples use hypothetical numbers. See our disclaimer and editorial policy.