Key Takeaways

  • JPMorgan’s inputs: 0.33 correlation, 55% versus 18% volatility
  • 20% stock drop implies roughly 20% Bitcoin drop
  • IMF found correlation rose from 0.01 to 0.36
  • Stress case: 5% Bitcoin adds 1.4-point loss

In a November 4 note, Grayscale Head of Research Zach Pandl called Bitcoin a potentially attractive portfolio diversifier, pointing out that its estimated correlation with U.S. large caps “is just 33%.” Grayscale sells Bitcoin investment products, and the note builds on figures from J.P. Morgan Asset Management’s 2027 Long-Term Capital Market Assumptions, released October 19.

What “just 33%” actually measures

Correlation tells you how consistently two assets move in the same direction. It says nothing about how far each one moves. For that, the 0.33 has to be read next to the other two numbers in JPMorgan’s standalone Bitcoin exhibit: 55% annualized volatility for Bitcoin and 18% for U.S. large-cap stocks.

Multiply the correlation by the ratio of the two volatilities and you get the average Bitcoin move that accompanies a 1% move in stocks. On JPMorgan’s inputs, 0.33 × (55 ÷ 18) comes to about 1.0. A loose link to an asset three times as volatile produces a one-for-one response, so a 20% equity decline carries an expected Bitcoin decline of about 20%.

That average comes with a very wide spread. The part of Bitcoin’s volatility not tied to stocks still works out to about 52% a year on the same inputs, and that independent movement is what Grayscale’s case rests on. Over a full decade it can lift a portfolio’s return per unit of risk; Coindoo’s breakdown of JPMorgan’s Bitcoin assumptions shows a textbook optimizer reaching for roughly a quarter of a two-asset portfolio on these numbers. Within any single sell-off, though, that independent swing can land on either side of the average, so it carries no promise of cushioning.

Why JPMorgan won’t build on the 0.33

JPMorgan treats its own correlation figure as too unreliable to drive allocations. Its executive summary names Bitcoin’s volatility and unstable correlations as obstacles to wider portfolio use, and keeps the asset out of the main covariance matrices that feed the firm’s strategic allocations.

The record behind that caution is short and steep. IMF research measured Bitcoin’s correlation with the S&P 500 at 0.01 in 2017–19 and 0.36 in 2020–21, with spillovers between the two markets rising during turbulent stretches such as March 2020. JPMorgan’s figure sits close to the later reading, and the jump between the two shows how far the number can travel in a few years. A higher reading during the next sell-off is the scenario worth pricing in.

One 20% stock drop, three correlation settings

The test uses the 60/40 stock-bond mix JPMorgan treats as its balanced benchmark, scaled to $100. Stocks fall 20% and bonds hold flat, which leaves the plain portfolio down 12.0%. Bitcoin replaces part of the stock slice, and its move is set at the average JPMorgan’s volatility figures imply for each correlation level.

Portfolio loss by Bitcoin share and correlation

Correlation 0.33

JPMorgan’s published figure

Correlation 0.6

Hypothetical stress setting

Correlation 0.8

Hypothetical stress setting

Coindoo calculations from JPMorgan’s 2027 LTCMA inputs: 55% Bitcoin volatility, 18% U.S. large-cap volatility, 0.33 correlation. The 0.6 and 0.8 settings are hypothetical, not measured values. Implied moves are statistical averages, and actual moves in any single sell-off can differ widely. No leverage, fees, taxes or rebalancing. Allocation sizes are arithmetic inputs, not recommendations.

A cushion at none of the settings, a cost that scales with size

At JPMorgan’s published correlation, the swap changes nothing. Bitcoin’s implied average fall matches the stocks it replaced, so the portfolio loses exactly what the plain 60/40 loses. On these inputs, Bitcoin behaves like extra equity exposure during a sell-off rather than protection against one.

Once the correlation climbs, position size decides how much it hurts. The 5% position adds about two and a half times the extra loss of the 2% position, in line with its larger weight. A position bought at 2% also rarely stays at 2% through a rally, a problem Coindoo’s look at how small Bitcoin stakes drift works through.

The funding choice matters as well, though for a reason that has nothing to do with Bitcoin. Taking the slice from bonds instead of stocks keeps the full 60% equity weight in place, which adds the stock decline on that slice to every figure above: 0.4 points for a 2% position and 1.0 point for 5%. The flat-bond assumption is also generous. JPMorgan’s own portfolio chapter expects that “stock-bond correlation will be more positive and unstable,” which would leave the bond side a weaker shock absorber too.

The number to watch next October

Each 0.1 rise in the correlation lifts the implied Bitcoin move by about 0.3% for every 1% fall in stocks, as long as the two volatility figures stay where JPMorgan set them. That makes the correlation line, more than the 10% return assumption, the figure that decides whether Grayscale’s diversification case survives a sell-off.

The 2028 edition of JPMorgan’s assumptions, due around October 2027 on the firm’s annual cycle, is the checkpoint. Bitcoin moving into the main matrices with a correlation at or below 0.33 would support Grayscale’s reading. A higher figure at unchanged volatility would push Bitcoin’s implied sell-off behavior closer to a leveraged stock position, and the 5% column above would become the one that matters.

Until then, holders can run the check without a model. Multiply the published correlation by about three, Bitcoin’s volatility relative to stocks, and the result is the average Bitcoin move JPMorgan’s inputs attach to each 1% move in the stock market.

This article is provided for informational purposes only and does not constitute financial or investment advice. Model assumptions and illustrative calculations do not guarantee future performance.