Crypto · Coins and tokens
Bitcoin in a portfolio: what a small allocation has done
Twelve years of monthly data on adding 1% to 10% bitcoin to a stock and bond mix: return, volatility, drawdown, correlation and what rebalancing changed.
On this page
Whether to hold bitcoin is a judgement about the asset. How a holding of a given size behaves inside a portfolio is arithmetic, and it can be measured. This guide does the second. It takes twelve years of month-end prices, adds 1%, 3%, 5% and 10% bitcoin to a plain stock and bond mix, and reports what happened to return, volatility and losses.
The results are history, and for bitcoin history is short and unusually favourable. The most useful findings are the ones that do not depend on the size of the gain: how much risk a small weight carries, how unstable the correlation is, and how much work rebalancing does. Nothing here is a recommendation to hold any amount.
The data and its limits
All figures use month-end closing prices from Finabase chart data for bitcoin and four exchange-traded funds: SPY (US large companies), BND (US investment-grade bonds), GLD (gold) and QQQ (the Nasdaq-100). The period is the 144 months from November 2014 to October 2026.
Three limits apply to everything below:
- Price only. Fund prices exclude dividends and interest. That understates SPY and especially BND, most of whose return is interest, while bitcoin pays nothing and is not affected. The comparison is tilted in bitcoin's favour.
- No costs or taxes. Trading costs, fund fees beyond what is in the price, exchange fees and tax on rebalancing sales are ignored.
- One period. Twelve years is one stretch of history in which bitcoin went from a niche asset to a $1.29 trillion one. Month-end data also hides falls that happened and reversed within a month.
Bitcoin on its own
| Asset | Total price change | A year | Volatility | Deepest fall (month-end) | Worst month |
|---|---|---|---|---|---|
| Bitcoin | +22,788% | 57.3% | 71.1% | −75.6% | −37.8% |
| QQQ | +581.0% | 17.3% | 18.7% | −33.1% | −13.6% |
| SPY | +274.7% | 11.6% | 14.9% | −24.8% | −13.0% |
| GLD | +215.8% | 10.1% | 15.8% | −26.5% | −11.7% |
| BND | −13.1% | −1.2% | 5.0% | −23.4% | −4.4% |
Volatility here is the standard deviation of monthly returns, annualized, and the deepest fall is the maximum drawdown on month-end prices. Bitcoin's deepest fall ran from December 2017 to January 2019. A second, from October 2021 to December 2022, was 73.0%. Its calendar years show the pattern more plainly than an average does:
| Year | Bitcoin | SPY |
|---|---|---|
| 2015 | +34% | −1% |
| 2016 | +124% | +10% |
| 2017 | +1,369% | +19% |
| 2018 | −74% | −6% |
| 2019 | +92% | +29% |
| 2020 | +303% | +16% |
| 2021 | +60% | +27% |
| 2022 | −64% | −19% |
| 2023 | +155% | +24% |
| 2024 | +121% | +23% |
| 2025 | −6% | +16% |
| 2026 (to October) | −12% | +11% |
Two things to take from the table. Bitcoin's two worst years, 2018 and 2022, were also years when stocks fell. And the growth rate has been falling: the first six years returned 85.5% a year and the second six 33.3%. That is what happens as any asset gets larger. For bitcoin to repeat its twelve-year price change of roughly 229 times from here, its market cap would have to go from $1.29 trillion to about $295 trillion.
Correlation: low on average, higher when it mattered
Correlation measures whether two assets' returns move together, from −1 to +1. A volatile asset can still lower a portfolio's risk if its correlation with the rest is low enough. Here is bitcoin's, on monthly returns:
| Bitcoin with | Full period | 2014–2020 | 2020–2026 |
|---|---|---|---|
| SPY (US stocks) | 0.35 | 0.24 | 0.51 |
| QQQ (Nasdaq-100) | 0.33 | 0.21 | 0.48 |
| BND (US bonds) | 0.12 | 0.11 | 0.13 |
| GLD (gold) | 0.06 | 0.08 | 0.06 |
For reference, SPY and BND had a correlation of 0.38 over the same period, and SPY and GLD 0.09.
The full-period 0.35 with stocks looks like a useful diversifier. The split shows it was not stable: the figure roughly doubled between the two halves.
Over rolling three-year windows, bitcoin's correlation with SPY ranged from −0.01 in May 2019 to 0.61 in January 2023, and was 0.48 at the end. With gold it stayed between −0.13 and 0.36 and ended at 0.05. Bitcoin is sometimes called digital gold; on this measure the two have had almost nothing to do with each other.
A correlation is an average across all months. What a holder cares about is the bad ones. SPY fell 5% or more in 14 of the 144 months. Bitcoin fell in 11 of them, and its average change across the 14 was −5.6%.
| Month | SPY | Bitcoin |
|---|---|---|
| March 2020 | −13.0% | −25.1% |
| June 2022 | −8.6% | −37.8% |
| May 2019 | −6.4% | +60.2% |
So bitcoin has not behaved as protection against stock market falls. It has more often added to the loss. May 2019 shows that it does sometimes move the other way, which is why the correlation is 0.35 and not 0.9.
Size is the risk control
The most reliable arithmetic in this guide needs no forecast:
| Bitcoin weight | Bitcoin falls 50% | Bitcoin falls 75.6% (its deepest) | Bitcoin goes to zero |
|---|---|---|---|
| 1% | −0.5% | −0.8% | −1% |
| 3% | −1.5% | −2.3% | −3% |
| 5% | −2.5% | −3.8% | −5% |
| 10% | −5.0% | −7.6% | −10% |
| 25% | −12.5% | −18.9% | −25% |
The same works for volatility. A 5% weight in an asset with 71.1% volatility carries about 3.6 percentage points of volatility on its own (5% × 71.1%), against 9.9% for the whole 60/40 mix in this data. A small weight is not a small share of the risk. See Concentration risk: when one holding dominates a portfolio.
A worked mix: 60/40 plus bitcoin
The base portfolio is 60% SPY and 40% BND. Each variant puts a share in bitcoin and splits the rest 60/40, then resets to those weights every December. A 5% variant is therefore 5% bitcoin, 57% SPY and 38% BND.
| Bitcoin share | Return a year | Volatility | Deepest fall | Worst month | Return ÷ volatility |
|---|---|---|---|---|---|
| 0% | 6.6% | 9.9% | −21.2% | −8.1% | 0.67 |
| 1% | 8.2% | 10.2% | −21.6% | −8.3% | 0.80 |
| 3% | 10.9% | 11.5% | −22.3% | −8.8% | 0.95 |
| 5% | 13.4% | 13.1% | −23.4% | −9.2% | 1.02 |
| 10% | 18.8% | 17.4% | −26.8% | −10.2% | 1.08 |
Over the full period every step up in bitcoin raised the return, and raised volatility, the deepest fall and the worst month too. Return rose faster than risk, so the last column improved. That is the result that gets quoted.
Now split the same twelve years in two.
| Bitcoin share | 2014–2020: return, volatility, deepest fall | 2020–2026: return, volatility, deepest fall |
|---|---|---|
| 0% | 5.5%, 8.8%, −11.9% | 7.8%, 10.9%, −21.2% |
| 1% | 7.9%, 9.3%, −12.2% | 8.2%, 11.1%, −21.6% |
| 5% | 16.2%, 13.5%, −13.1% | 10.1%, 12.4%, −23.4% |
| 10% | 24.6%, 19.3%, −14.3% | 12.2%, 14.6%, −26.8% |
In the first half, 5% bitcoin added 10.7 percentage points a year to the return. In the second half it added 2.3. The risk it added did not shrink in the same way: the deepest fall with 10% bitcoin was 26.8% in the second half against 21.2% without. Return per unit of volatility in the second half went from 0.72 with no bitcoin to 0.81 with 5%, a far smaller improvement than the full-period table suggests.
The honest reading is that most of the twelve-year result came from the first six years, when bitcoin was a much smaller asset. A backtest that starts in 2014 measures that period as much as it measures bitcoin.
What rebalancing did
The tables above assume the mix is reset each December. Without that, the result is a different portfolio:
| Starting bitcoin share | Never rebalanced: return a year | Volatility | Deepest fall | Bitcoin share at the end |
|---|---|---|---|---|
| 1% | 14.1% | 23.3% | −45.2% | 47.1% |
| 3% | 20.5% | 35.5% | −59.4% | 73.2% |
| 5% | 24.5% | 41.6% | −64.1% | 82.3% |
| 10% | 30.9% | 49.6% | −68.4% | 90.7% |
A 1% position left alone became 47% of the portfolio, and the portfolio's deepest fall was 45% where the rebalanced version's was 22%. The investor who "only put in 1%" ended up holding something close to a half-bitcoin portfolio with a risk to match. Starting six years later the drift was milder, with a 5% position ending at 14.7%, but still nearly three times the target.
Rebalancing is what kept the weight, and so the loss table above, true. It has two costs. In this data it gave up return, because it kept selling an asset that went on rising: 13.4% a year against 24.5% for the 5% mix. And even yearly resets let the weight wander between dates: the 5% target reached 40.8% of the portfolio in December 2017 just before being reset, at the month-end from which bitcoin's deepest fall began. A band rule, which resets whenever the weight crosses a limit, would have acted sooner. In a taxable account each reset that sells bitcoin at a gain is a taxable sale under IRS rules, which treat crypto as property.
Pitfalls
- Extrapolating the return. The 57% a year belongs to an asset growing from very small to very large. The second half's 33% is already much lower, and neither is a forecast.
- Trusting one correlation number. It doubled between the two halves and reached 0.61.
- Calling it a hedge. In most of the months when stocks fell hard, so did bitcoin, and by more.
- Setting a weight and not maintaining it. An unrebalanced small position can become the whole portfolio's risk.
- Sizing by the gain you want. Size by the loss in the table that you could carry without selling at the bottom. A fall of about three quarters has happened twice in twelve years.
- Counting it twice. Shares of bitcoin miners, exchanges and companies that hold bitcoin in treasury move with it. So do other crypto assets; see Types of cryptocurrency: a map of the largest coins.
- Ignoring how it is held. A spot ETF, an exchange account and self-custody have different costs and different ways to fail. See Spot bitcoin and ether ETFs: how they work and what they cost and Crypto wallets and custody: keys, seed phrases, hot and cold storage.
Ways to hold the position compared
| Route | What you own | Main extra risk | Rebalancing |
|---|---|---|---|
| Spot bitcoin ETF (approved in the US from January 2024) | Fund shares in a brokerage account | Fund fee; trades only in market hours | Same as any fund, alongside stocks and bonds |
| Exchange account | A claim on the exchange for bitcoin | The exchange failing or freezing withdrawals | Needs transfers between the exchange and a broker |
| Self-custody wallet | Bitcoin itself, controlled by your keys | Losing the keys or being tricked into sending | Slowest; network fee on each move |
How Finabase does it
Finabase does not suggest an allocation, calculate a correlation between two symbols or rebalance a portfolio.
- Comparison chart. Open bitcoin against SPY, BND and GLD to see the four as percent changes from a common start. Changing the chart range shows how much the picture depends on the starting date.
- Portfolio weights. A portfolio page lists each position's market value with the total, and the allocation bar shows the largest holdings as shares of the whole. A position's value divided by the total is its current weight. Bitcoin held as BTC.X and a spot bitcoin ETF such as IBIT are separate rows, so add them for the true exposure.
- Alerts as a band. A price alert on BTC.X can mark the price at which your holding would cross a weight you have chosen.
The figures in this lesson were calculated from the same month-end prices the charts use.
Frequently asked questions
What percentage of a portfolio should be in bitcoin?
There is no figure that can be derived from data, and this guide does not give one. What the data supports is a method: decide the largest loss from this one position you could sit through, divide it by a fall of 75% or more, and that bounds the weight. Zero is a valid answer.
Is bitcoin a hedge against stock market falls?
Not on this evidence. In the 14 months since late 2014 when SPY fell at least 5%, bitcoin fell in 11 and averaged −5.6%. Its three-year correlation with stocks peaked at 0.61 in January 2023 and was 0.48 at the end of the period.
Is bitcoin like gold in a portfolio?
Their returns have been almost unrelated: a correlation of 0.06. Gold's volatility over the period was 15.8% and bitcoin's 71.1%, and gold's correlation with stocks was 0.09 against bitcoin's 0.35. They may share a story about scarcity, but they have not shared behaviour.
Does adding bitcoin always improve risk-adjusted return?
It did over this period, and by much less in the second half than the first. The result depends entirely on bitcoin's return staying high enough to pay for its volatility. With a lower return and the same volatility, the same arithmetic gives a worse portfolio.
How often should a bitcoin position be rebalanced?
More volatile assets drift faster, so a calendar rule leaves longer gaps than it would for stocks. Yearly resets in this data still let a 5% target reach 40.8%. A band checked monthly reacts to the size of the drift instead of the date. See Rebalancing methods: calendar, threshold and what each one costs.
Summary
From late 2014 to October 2026, a small bitcoin allocation raised the return of a stock and bond mix, and raised its volatility and drawdowns proportionally less. Most of that benefit came in the first six years. Bitcoin's volatility was 71%, it fell by about three quarters twice, and its correlation with stocks rose to around 0.5, with losses arriving in the same months as stock losses. The dependable tools are position size, which fixes the worst case, and rebalancing, which keeps the size from changing on its own. Past results describe one period and are not a forecast.
Check your understanding
5 questions on this lesson. Each answer comes with the reason, and nothing is scored.
A portfolio holds 5% in an asset that then falls 75.6%. Everything else is unchanged. How much does the portfolio lose?
- About 3.8%
- About 7.6%
- 75.6%
Show the answer
A. About 3.8% Portfolio loss = weight × the position’s fall: 5% × 75.6% = 3.8%. A 10% weight would lose 7.6%. Position size fixes the worst case without any forecast.
From November 2014 to October 2026, what happened to the correlation between bitcoin’s and US stocks’ monthly returns?
- It stayed near zero throughout
- It rose, from 0.24 in the first six years to 0.51 in the second
- It was negative, so bitcoin offset stock losses
Show the answer
B. It rose, from 0.24 in the first six years to 0.51 in the second The correlation with SPY roughly doubled between the two halves, and over rolling three-year windows it reached 0.61 in January 2023. A single full-period figure of 0.35 hides that.
A 60/40 mix started with 1% in bitcoin in November 2014 and was never rebalanced. Roughly what share was bitcoin by October 2026?
- Still about 1%
- About 5%
- About 47%
Show the answer
C. About 47% Left alone, the position grew to 47.1% of the portfolio, and the portfolio’s deepest fall was 45.2% against 21.6% for the version reset to 1% each December.
Why does the price-only data in this lesson flatter bitcoin against the stock and bond funds?
- Bitcoin prices are quoted in dollars
- Fund prices leave out dividends and interest, which are part of stock and bond returns, while bitcoin pays neither
- Bitcoin trades at weekends
Show the answer
B. Fund prices leave out dividends and interest, which are part of stock and bond returns, while bitcoin pays neither SPY’s and especially BND’s total returns include income that a price series omits. Bitcoin has no income, so its price change is its whole return.
With 5% bitcoin, the rebalanced mix returned 16.2% a year in 2014–2020 and 10.1% in 2020–2026, against 5.5% and 7.8% without. What is the fairest conclusion?
- Bitcoin adds about 7 points a year to any portfolio
- Most of the twelve-year benefit came from the first six years, so the full-period figure depends heavily on the start date
- The benefit is growing over time
Show the answer
B. Most of the twelve-year benefit came from the first six years, so the full-period figure depends heavily on the start date The lift was 10.7 points a year in the first half and 2.3 in the second. A backtest starting in 2014 measures that early period as much as it measures bitcoin.
Sources and further reading
- Finabase chart data — month-end closing prices of bitcoin (BTC.X), SPY, BND, GLD and QQQ, October 2014 to October 2026, frozen November 18, 2026; all returns, volatilities, correlations, drawdowns and portfolio mixes calculated by Finabase
- CoinMarketCap listings as shown on Finabase (finabase.io/crypto), read November 18, 2026 — bitcoin and total crypto market cap
- SEC — Order approving the listing of spot bitcoin exchange-traded products (January 10, 2024)
- Investor.gov (SEC) — Beginners’ Guide to Asset Allocation, Diversification, and Rebalancing
- IRS — Digital assets (sales and exchanges of digital assets are taxable)
