Is Bitcoin Correlated With Stocks? What IMF Data Shows

For years the pitch for Bitcoin included diversification: it was said to move on its own, regardless of stocks. Economists at the International Monetary Fund checked the claim on data and wrote up the result in a short, widely cited post. The numbers are specific, so they are worth reading carefully.

The numbers

Adrian, Iyer and Qureshi report that “before the pandemic, Bitcoin and Ether showed little correlation with major stock indices”. The correlation coefficient of Bitcoin and S&P 500 returns was 0.01 in 2017–19. “That correlation jumped to 0.36 for 2020–21”, so the two increasingly moved together.¹

A correlation coefficient runs from −1 (opposite moves) through 0 (no linear relationship) to +1 (identical). 0.01 means no relationship; 0.36 is a moderate positive one: on days when stocks rose, Bitcoin was more likely, but far from always, to rise too.

The post adds more figures:¹

• for emerging-market stocks (the MSCI emerging markets index) the correlation with Bitcoin was 0.34 in 2020–21, “a 17-fold increase” from the years before;
• Bitcoin’s correlation with stocks was higher than the correlation between stocks and gold, investment-grade bonds or major currencies, “pointing to limited diversification benefits”;
• spillovers, meaning how much of one market’s moves are explained by the other’s, rose sharply: Bitcoin volatility explains about one-sixth of S&P 500 volatility during the pandemic, and about one-tenth of the variation in S&P 500 returns;
• spillovers “tend to increase in episodes of financial market volatility”, for example March 2020 and the swings of early 2021.

The IMF also noted that crypto assets rose from $620 billion in 2017 to nearly $3 trillion in November 2021, and drew the conclusion that their growing adoption “has coincided with significantly higher correlation” with traditional holdings.¹

What it means for a trader

1. “Uncorrelated” was a fact about 2017–19, not a law. The same pair of assets gave 0.01 in one window and 0.36 in the next. Any claim of the form “Bitcoin is a hedge” or “Bitcoin follows the Nasdaq” is a statement about a window.
2. Correlation rises when it hurts. The post links larger spillovers to stress episodes. A hedge that stops hedging in a crisis is not a hedge.
3. Correlation is not causation. The post measures co-movement and spillovers in both directions, “from Bitcoin returns and volatility to stock markets, and vice versa”. It does not show that stocks move Bitcoin, or the reverse, and it does not claim to.
4. Check your own window. Correlation changes with the period, the data frequency and the market. If you use it in a model, compute it on your data, with the dates written down, rather than copy a number from an article, this one included. For other known time-varying relations see crypto momentum and attention.
5. It is about Bitcoin, Ether and Tether. The same work estimates that spillovers from Tether to U.S. equity markets were smaller than those from Bitcoin, explaining “about 4 percent to 7 percent” of the variation in U.S. equity returns and volatility. For how Tether itself was studied, see did Tether inflate Bitcoin in 2017.

The sentiment around the market can be tracked on our Fear & Greed page. It is a sentiment index, not a correlation measure.

This article is educational material, not investment advice. The figures are those published by the IMF in January 2022; they cover 2017–2021 and may differ in later data.