JPMorgan’s long-term assumptions put a 10% return on bitcoin, and keep it out of the portfolios

J.P. Morgan Asset Management has put a return figure on bitcoin in its Long-Term Capital Market Assumptions for the first time, and kept it out of every portfolio the document builds. The 2027 edition, the 31st, came out on 19 October 2026. Its press release says a range on bitcoin return “is introduced, but bitcoin is not included as an allocation”. The release itself prints no bitcoin number.

The numbers circulating since come from coverage, not from that release. Bitcoin.com News, Forkast, TFTC and Coindoo all report that JPMorgan assumes a compound annual return of 10.0% for bitcoin over 10 to 15 years, an arithmetic return of 20.9% and annualised volatility of 55%. The four outlets give the same three numbers. JPMorgan’s full report and its return matrices sit behind a registration wall on its site, and Coindaro could not open them to check the bitcoin line itself.

These are long-term model assumptions used for asset allocation. The release says its figures are in US dollars unless stated, and that they leave out fees, taxes and trading costs. They are not a bitcoin price forecast or a target, JPMorgan does not turn them into a holding in its portfolios, and nothing here is investment advice.

Bitcoin traded near $81,450 at 10:22 UTC on 20 October, down 0.85% over 24 hours and 2.5% over seven days, according to Coindaro’s market data. The price shows no clear reaction to the document: bitcoin had been holding near $82,000 through the Asian morning, driven by other news.

What JPMorgan’s long-term capital market assumptions say about bitcoin

The LTCMA is one of the reference documents of the asset management industry. Pension funds, endowments and private banks use its numbers as inputs when they decide how much to put in stocks, bonds and alternatives over a decade or more. The 2027 edition uses data as of 30 September 2026 and covers more than 20 base currencies.

On bitcoin, JPMorgan’s own words are few. The LTCMA landing page advertises a “first look at bitcoin”. It says the team puts “a range on bitcoin returns” and that “we do not yet include bitcoin in our main covariance matrices”. Those matrices are the tables of volatilities and correlations that feed the allocation models. Bitcoin gets a return range, and stays out of the tables that build the portfolios.

The landing page does not say which base currency the bitcoin figures use, nor whether the “range” is a band or one central number. The coverage reports a single 10.0%.

The report was led by John Bilton, Global Head of Strategic Research for Multi-Asset Solutions, with David Kelly, Karen Ward, Tai Hui, Gabriela Santos, Pulkit Sharma and David Lebovitz among the listed authors. None of the quotes in the press release is about bitcoin.

Bitcoin at 10.0% a year against 7.5% for US stocks and 5.5% for gold

Placed next to the rest of the document, the reported bitcoin figure sits at the top of the return column. The table below puts the press release figures next to the reported bitcoin line and the cash figure given by Bitcoin.com News and Coindoo. All are compound annual returns in US dollars over 10 to 15 years.

Sources: J.P. Morgan Asset Management press release of 19 October 2026; Bitcoin.com News, Forkast, TFTC and Coindoo, 20 October 2026. Bitcoin.com lists US high yield at 6.6%; the press release says 6.7%, and the table uses the press release.

Several headlines have turned this into bitcoin beating stocks, bonds and gold. That holds only inside the listed table, and only on the compound figure. The narrower reading: in JPMorgan’s model, bitcoin’s assumed compound return is 2.5 points above US large caps and 4.5 points above gold, with about three times their assumed volatility.

A near tie with private equity, 0.2 point apart

The gap between bitcoin at 10.0% and private equity at 9.8% is 0.2 percentage point. Over 10 to 15 years, for an asset whose assumed volatility is 55%, that difference sits well inside any reasonable error band. It is a tie, not a ranking.

The comparison with US large caps is more telling. Stocks gained 0.8 point in this edition, from 6.7% to 7.5%, so the 2.5-point edge bitcoin shows over them is smaller than it would have looked against last year’s equity numbers. Gold, at 5.5%, is assumed to deliver a little more than half of bitcoin’s compound return for under a third of its volatility. Bitcoin.com News and TFTC report gold’s volatility assumption at 17.2%; Forkast and Coindoo put US large caps near 18%.

Why 10.0% compound and 20.9% arithmetic describe the same assumption

The two return figures look contradictory and are not. The arithmetic return is the average of the yearly returns the model expects. The compound, or geometric, return is the rate at which money actually grows when those years are chained together. The difference between them is called volatility drag.

A simple case shows it. An asset that gains 50% one year and loses 50% the next has an arithmetic average of zero. Start with $100 and you end with $75. The bigger the swings, the larger the gap between the average year and what the holder keeps.

On the reported figures, with 55% annual volatility, the model’s average year for bitcoin is 20.9%, yet the compounded result over the horizon is 10.0%. Almost 11 points of return vanish in the arithmetic of large swings. For US stocks or gold, with volatility near 17% to 18%, the gap between the two measures is far smaller. Newcomers who want the basics of what drives bitcoin’s price can start with our guide to what bitcoin is and how it works.

55% volatility keeps bitcoin out of the main covariance matrices

JPMorgan’s landing page gives no reason for leaving bitcoin out of the main matrices. The coverage points to two: the 55% volatility, and how little the model can say yet about how bitcoin moves with other assets. Bitcoin.com News and TFTC both write that, outside the main correlation data, JPMorgan cannot yet place bitcoin against the rest of a portfolio. The landing page says bitcoin is not “yet” in those matrices, which leaves a later edition free to change that.

What 55% means once bitcoin sits in a portfolio

Other managers have tried to size the problem. In its paper on sizing bitcoin in portfolios, published in December 2024, the BlackRock Investment Institute found that a 1% to 2% bitcoin position in a 60/40 portfolio adds about as much risk as one of the large “Magnificent 7” tech stocks. At 4%, bitcoin would account for about 14% of the portfolio’s total risk. BlackRock has since gone further on the product side, filing for a bitcoin ETF that caps the upside to limit losses, a structure built around that same volatility.

JPMorgan stops one step earlier than BlackRock’s sizing exercise. It puts a return assumption on bitcoin, which clients can feed into their own models, and leaves it out of the portfolios it builds. Forkast’s Nolan Pratt called that “permission to look, not permission to buy”.

Bitwise’s Matt Hougan says the volatility input is too high

The sharpest reaction came from Matt Hougan, chief investment officer of Bitwise, a firm that sells crypto funds. In a post on X on 20 October, reported by TFTC, Bitcoin.com News and Coindoo, he welcomed bitcoin’s place in the report but argued that the 55% volatility assumption is too high and the 10% return too low. TFTC and Bitcoin.com News both quote his opening line, “First they ignore you, then they laugh at you”. Coindaro did not see the post itself.

His argument goes straight to the volatility drag above. Lower the volatility assumption and, for the same arithmetic return, the compound figure rises. Whether 55% is too high is the open question. JPMorgan has not answered Hougan publicly, and its published pages do not explain how the volatility figure was set.

In the same Forkast piece, Pratt read the split between a return range and no allocation as a sign that the report is “a map of an evolving financial world, not a mandate”.

Crypto asset managers assume far more than 10%

JPMorgan’s 10% is modest next to what crypto-native houses publish. VanEck published its own bitcoin capital market assumptions on 8 January 2026, signed by Matthew Sigel and Patrick Bush. Its base case is a 15% compound annual return over 25 years, from 2026 to 2050, and it names a strategic allocation of 1% to 3% for diversified portfolios. VanEck sells bitcoin products and says it may hold the assets it writes about. JPMorgan’s document, by contrast, puts no bitcoin in any of its model portfolios.

JPMorgan’s assumption should also not be confused with notes from JPMorgan’s investment bank strategists, a separate team that publishes its own bitcoin research, at times with price levels drawn from comparisons with gold. The LTCMA is a different kind of document: a return and risk input for asset allocation, with no price attached.

Some outlets have also turned the 10% into a bitcoin price path over 10 or 15 years. That arithmetic is theirs. JPMorgan gives no starting price and no target, and the 10-to-15-year window has no fixed start date beyond the data cut-off of 30 September 2026.

From the 2026 edition to the 2027 edition: higher returns across the board

The 2026 LTCMA, released on 23 October 2025, said nothing about bitcoin or crypto in its press release. This year’s numbers are higher almost everywhere. The USD 60/40 portfolio moves from 6.4% to 7.0% a year. The 60/40 with 30% in diversified alternatives moves from 6.9% to 7.5%, with a Sharpe ratio about 21% higher than the plain 60/40, against about 25% higher last year.

Bonds are the other headline. JPMorgan puts global aggregate bonds at 5.3%, which it calls the highest in two decades, against 2.1% in 2021. US high yield rises to 6.7% from 6.1%, and US intermediate Treasuries sit at 4.6%. Those numbers matter for bitcoin too. When cash and bonds are assumed to return 3% to 5%, an asset with 55% volatility has to clear a higher bar: on the reported figures, bitcoin’s 10.0% is 5.4 points above intermediate Treasuries at 4.6% and 6.8 points above cash at 3.2%.

What it changes for investors in the US, France and Italy

For US investors, the change is one of status, not of access. Bitcoin already trades through spot exchange-traded funds, and advisers can buy it for clients. What JPMorgan adds is a reference figure that an adviser can put in a planning model, with a clear warning attached that the firm itself does not allocate to it.

For readers in the euro area, two points apply. JPMorgan’s headline figures are in US dollars, so a euro-based holder carries the euro-dollar exchange rate on top of bitcoin’s own swings, and the 55% volatility figure is a dollar figure. The LTCMA does come in more than 20 base currencies, but Coindaro could not confirm whether a euro version of the bitcoin range is published. Neither the AMF in France nor Consob in Italy has said anything about this document, and local tax on crypto gains applies whatever any model assumes.

What is still unknown

Several points remain open until the full matrices are checked:

• Whether the bitcoin “range” is a band around 10.0% or a single central figure, as the coverage reports.
• Which base currency and starting date the bitcoin line uses, and whether a euro version exists.
• The volatility and correlation inputs behind the bitcoin line, which JPMorgan’s public pages do not show.
• Whether a later edition moves bitcoin into the main covariance matrices, the step that would let JPMorgan’s models allocate to it.

The next LTCMA, the 2028 edition, would normally follow in October 2027. Until then, bitcoin keeps a line in JPMorgan’s long-term table and no weight in its portfolios. The bitcoin page carries the live price and market data.