Why bitcoin follows the Nasdaq — and when the correlation breaks

The observation is easy to make and easy to over-generalise: bitcoin often moves with the Nasdaq, sometimes almost tick for tick through a US session, and then spends weeks ignoring it entirely. Both behaviours are normal, and the reason for each is the same — correlation is not a property of the assets, it is a property of whatever is currently driving the news flow.

The transmission channel

A high-multiple technology company is valued mostly on cash flows expected years out. Discount those at a higher real rate and the present value falls sharply. That is what makes the Nasdaq a long-duration asset: it is unusually sensitive to the rate at which the future is discounted.

Bitcoin has no cash flows to discount, but it occupies the same position in a portfolio. It is held for a payoff that is expected later, it pays nothing in the meantime, and it competes directly with the risk-free rate for capital. When short-term real rates rise, the opportunity cost of holding a non-yielding asset rises with them. The mechanism differs; the direction of the sensitivity does not.

• A hot inflation print raises expected policy rates, which raises real yields, which pressures both.
• A dovish central-bank surprise does the reverse and typically lifts both within the same hour.
• A liquidity shock — funding stress, a credit event — hits the highest-beta expressions of risk first, and crypto is usually one of them.
• A growth scare can split them: equities fall on earnings risk while crypto responds mainly to the rate implication.

This is why the correlation is at its tightest around scheduled macro events. On CPI day and FOMC day the entire market is trading one variable, and every asset that is sensitive to that variable moves together for as long as it dominates.

Correlation is a regime, not a constant

Rolling 30-day correlation between BTC and the Nasdaq has spent long periods above 0.6, long periods near 0, and occasional stretches negative. Any statement of the form "bitcoin is correlated to equities" is incomplete without a window and a date. The useful version of the question is: what is the correlation now, over what window, and what is driving it?

The last row deserves emphasis. A correlation computed over a period when neither asset moved much is statistically noisy and behaviourally meaningless. Correlation matters most exactly when it is least stable — during the moves you would want it to protect you from.

Measuring it without fooling yourself

Three mistakes account for most published correlation figures that turn out to be meaningless, and all three are avoidable in the first minute of the calculation.

1. Correlate returns, not prices. Two series that both trend upward produce a high correlation of levels regardless of whether they move together day to day. Percentage changes are the only inputs that answer the question being asked.
2. Align the clocks. Crypto trades continuously and equities do not. A daily correlation that compares a 24-hour crypto candle with a 6.5-hour equity session is measuring two different windows; either restrict crypto to the equity session or accept that the overnight move is noise in the estimate.
3. State the window. A 30-day rolling correlation and a 200-day correlation routinely disagree, and neither is wrong. Publishing one without the other is how a regime observation gets mistaken for a property of the asset.

A fourth, subtler point: correlation says nothing about magnitude. Two assets can be correlated at 0.8 while one moves three times as far, which is the usual relationship between crypto and equities. Beta answers the size question and correlation answers the direction question, and conflating them is how a position that looked like a hedge turns out to be a leveraged copy of the thing it was hedging.

What actually breaks the link

Crypto decouples when the marginal news is crypto-specific and large enough to dominate. A venue failure, a major protocol event, a regulatory decision, a large forced seller, or a structural flow like a fund launch will all pull the asset out of the macro regime for as long as the story is the primary driver. The tell is straightforward: the move happens without a matching move in rates, the dollar or equity index futures.

That cross-market check is the fastest way to classify a move in real time. If BTC drops 3% while Nasdaq futures, the dollar and front-end yields are unchanged, the cause is inside crypto — a liquidation cascade, a large seller, or a story that has not reached general coverage yet. If all four moved together, the cause is macro and the crypto-specific search will waste your time.

Using this in practice

1. Before an event, check whether the current regime is macro-driven. If it is, treat crypto exposure as an expression of the same trade as your equity exposure.
2. During a move, ask which market led. A move that starts in rates and arrives in crypto is a different trade from one that starts on a crypto venue.
3. After the event, record what correlated and what did not. Regimes persist for weeks, and the record tells you which regime you are in faster than a rolling statistic will.
4. Cap combined risk across assets sharing a driver, as covered in position sizing.

For a repeatable way to measure the second step rather than eyeball it, see measuring news impact on price, which covers how to define a window and a baseline so cross-asset comparisons mean something.