What correlation actually measures
Correlation is a number between -1 and +1 that describes how two series tend to move together. A value near +1 means that when one rises, the other usually rises too. Near -1 means they tend to move in opposite directions. Near 0 means there is no consistent linear relationship.
Two details are easy to miss. First, correlation says nothing about size: two assets can have a correlation of 0.8 even if one moves far more than the other. Second, correlation does not say that one thing causes the other. Both may simply be responding to the same news.
Use returns, not price levels
When comparing bitcoin with a stock index, the calculation is normally done on returns, meaning the percentage change from one period to the next, rather than on raw prices. Prices that trend upward over time will often look strongly correlated even if their day-to-day moves have little in common. Using daily or weekly returns removes much of that effect.
Another practical point is timing. Stock markets close overnight and on weekends, while crypto trades continuously. Analysts usually need to line up observations carefully, for example by using the same daily closing time, or by using weekly data, so that the two series refer to comparable moments.
What a rolling correlation is
A single correlation computed over many years gives one number and hides how the relationship changed. A rolling correlation recalculates it repeatedly over a moving window, such as the most recent 30, 90 or 180 days, and plots the results as a line.
The window length is a trade-off.
- Short windows (a month or so) react quickly but are noisy, because a few extreme days can dominate.
- Long windows (many months) are smoother but slow to reveal a change in behaviour.
A tiny illustration shows the idea. Suppose five daily returns for a stock index are +1%, -1%, +2%, -2%, +1%, and bitcoin’s are +2%, -2%, +4%, -4%, +2%. Bitcoin’s moves are exactly double the index’s, so the correlation is +1, even though the sizes differ. If bitcoin’s returns were instead -2%, +2%, -4%, +4%, -2%, the correlation would be -1. Real data is nowhere near this tidy, which is why real correlations usually sit between these extremes and wander over time.
How the relationship has varied
Commentators have often reported that the correlation between bitcoin and large US stock indices, particularly technology-heavy ones, was low or unstable in the early years and then rose noticeably around the market turmoil of early 2020 and again during the 2022 period of rising interest rates, when many risk assets fell together. The exact figures depend on the index, the window and the dates chosen, so it is better to treat any specific number quoted online with some scepticism.
Equally important, the correlation has not stayed high. Rolling charts typically show long stretches where the relationship fades toward zero and occasionally turns negative before rising again. That variability is the main finding, not any particular level.
Why it changes
Several explanations are offered, none of them settled.
- Shared drivers. When markets focus on interest rates or inflation, many assets respond to the same news, pushing correlations up. Some of these links are discussed in the broader overview of how macro conditions affect crypto.
- Changing participants. As larger institutions and investment products entered the market, some people argue crypto became more tied to conventional portfolio decisions. This is plausible, but difficult to demonstrate cleanly.
- Stress periods. In sharp sell-offs, investors often reduce risk across the board, which can lift correlations temporarily. Correlations frequently rise exactly when diversification is wanted most.
- Crypto-specific events. An exchange failure or a large hack can push crypto down while stocks barely react, lowering the measured correlation.
Cautions when reading a correlation chart
- A correlation is backward-looking. A 90-day figure describes the last 90 days, not the next 90.
- Outliers matter. One extreme day can move a short-window figure considerably.
- Different indices give different answers. A broad stock index and a technology index will not behave identically.
- Correlation is not the same as the broader question of how bitcoin relates to other coins, which is a separate topic in the discussion of price correlation.
- A low correlation does not mean an asset is a reliable diversifier, because the relationship can change quickly.
Reading it sensibly
Rolling correlation is a descriptive tool. It can help you notice that bitcoin and equities have recently been moving together, and it can temper stories that crypto is always independent of traditional markets or always follows them. It cannot forecast the future, and it should not be used as a basis for deciding what to buy or sell.
Cryptocurrency prices are volatile, past relationships may not persist, and you can lose all you invest. This article is educational and is not financial, tax or legal advice.