Correlation measures how often two assets move together, on a scale from minus one to plus one, and crypto's correlation with conventional assets has been unstable. The International Monetary Fund's January 2023 Global Financial Stability Report documented the Bitcoin-S&P 500 correlation rising roughly fourfold after 2020, from near zero in 2017-2019 to multiyear highs around 0.4 during 2020-2022.
Horison publishes information, not investment advice, and what any correlation history means for a portfolio depends on individual circumstances this site cannot know. Correlation is a descriptive statistic, not a forecast, and past figures are never projected forward on this site. Crypto assets can lose most or all of their value quickly, a risk that no correlation reading reduces.
What Does a Correlation Number Actually Measure?
Correlation is the tendency of two return series to move in the same direction, conventionally measured as a Pearson coefficient between minus one and plus one. A reading near plus one means the assets rise and fall together; near minus one, they move oppositely; near zero, direction shows little relationship. Readings depend heavily on choices the analyst makes: return frequency, window length, and whether the series is rolling or full-period. Two honest reports can therefore show different numbers for the same asset pair.
Three practical cautions follow. First, correlation says nothing about magnitude: bitcoin can be near-uncorrelated with equities in direction yet swing several times harder in size. Second, correlations are time-varying, and the documented record shows regime shifts big enough to flip a diversification argument within months. Third, correlation measured in calm periods often breaks in stressed ones, a pattern documented across asset classes long before crypto existed.
How Has Bitcoin's Correlation With Stocks Behaved?
The documented record divides into eras. From 2017 through 2019, bitcoin's correlation with the S&P 500 averaged near zero, and the IMF described the period as crypto trading largely independent of equities. From 2020 through 2022, the relationship strengthened as institutional participation grew, and by the 2022 drawdown bitcoin was falling alongside technology-heavy indexes on shared macro drivers, with the IMF documenting the fourfold rise in correlation. In 2023-2025, readings fluctuated at moderate levels rather than settling at either extreme.
| Period | Documented behavior | Source basis |
|---|---|---|
| 2017-2019 | Bitcoin-S&P 500 correlation near zero (~0.01) | IMF GFSR, January 2023 |
| 2020-2022 | Correlation about 0.4, a roughly fourfold rise | IMF GFSR, January 2023 |
| March 2020 | Brief convergence in the crash; bitcoin fell about 50% in days | Exchange market data |
| 2023-2025 | Moderate, fluctuating readings; no settled regime | Market data providers |
The March 2020 episode is the sharpest documented illustration. During the pandemic crash, bitcoin fell roughly half in two days while U.S. equities hit repeated circuit breakers, a reminder that liquidity shocks hit simultaneously across markets previously assumed independent. Convergence under stress repeated in 2022, when rate-driven repricing pulled both technology stocks and crypto downward together.
How Does Ethereum Compare With Bitcoin?
Ethereum's equity correlations track bitcoin's eras closely, because the two largest crypto assets correlate strongly with each other. Data providers regularly record bitcoin-ethereum daily-return correlations in the 0.7-0.9 range across the 2020s, meaning Ethereum has rarely offered diversification within crypto itself. Relative to equities, Ethereum has tended to show slightly higher sensitivity in both directions, closer to a higher-beta version of the same exposure than to a separate asset class. Documented exceptions cluster around crypto-specific events, such as protocol upgrades, that moved Ethereum independently of macro markets.
What About Bonds?
Bitcoin's measured correlation with high-grade government bonds has hovered near zero to slightly negative across most documented windows, including the IMF's 2023 analysis. In direction terms, that reading resembles the one equities showed with bonds during the low-inflation 2010s. It does not mean bonds and crypto hedge each other, because the size of their moves differs by an order of magnitude in volatility terms.
The 2022 experience added a documented caution about regime dependence. That year, rising rates pushed stocks and long bonds down together, their worst joint year in decades, and crypto fell harder than both. An asset with near-zero average correlation to bonds still declined in the same macro episode, which is the pattern portfolio research calls correlation breakdown. Directional independence measured over years did not survive a single inflation shock.
A second caveat concerns what the bond comparison excludes. Most published figures measure direction against a benchmark index of long-dated government debt, not against the credit, duration, and liquidity factors that make up fixed-income portfolios. A near-zero reading against one Treasury index does not describe behavior against corporate bonds or against the cash-like instruments that many allocation frameworks treat as the true low-risk anchor. The documented bond-correlation record is therefore narrower than the equity record and should be read that way.
What Does This Mean for Diversification?
The documented implication is that correlation is the smaller half of the crypto portfolio question, and volatility is the larger half. Realized volatility for bitcoin ran roughly four times equity-market volatility across 2020-2024, per Coin Metrics and Bloomberg data, so position size dominates the effect of any correlation reading. An illustrative calculation makes the point, with stated assumptions: constant volatilities of 70 percent for bitcoin and 16 percent for equities, and zero correlation.
Illustrative, assumptions in text: at 70% annualized volatility, a 2% bitcoin sleeve contributes about as much return variance as a 9% allocation to a 16%-volatility equity index. (0.02 × 0.70 ≈ 0.014, and 0.0875 × 0.16 ≈ 0.014.)
The example is arithmetic, not advice, and its assumptions are static while markets are not. It shows why professional practice sizes volatile positions by their risk contribution rather than by capital weight, a framework documented in risk-budgeting literature. Whether any allocation to crypto fits a given investor remains a matter of individual circumstances, time horizon, and risk capacity.
What Are the Documented Limitations of Correlation Analysis?
Four limitations recur in the research. Window sensitivity: 90-day and multiyear correlations have told different stories simultaneously. Regime shifts: institutional adoption, the 2024 listing of spot funds, and dollar-token growth each changed market structure enough to question older data. Data quality: early crypto price series come from venues with thin books and gaps. Nonlinearity: correlation captures average co-movement, not tail behavior, and crypto's documented tail events exceed what average readings imply.
These limits do not make the measurements useless; they make them historical. The IMF's fourfold rise, the 2020 and 2022 stress convergences, and the near-zero bond readings are facts about specific periods, properly sourced and dated. Treating any of them as a permanent property of the asset would extrapolate past data forward, which this site does not do.
For more context, read Crypto Drawdown History: What Volatility Means.
For more context, read What Dollar-Cost Averaging Means for Crypto Investors.
For more context, read What Bitcoin Dominance Means for Allocations.
