Key takeaways
- Over the 10.3 years to 2 September 2026, bitcoin compounded at 64.7% a year and ether at 65.8%, a gap of 1.1 percentage points that is inside the noise of a decade.
- Ether's annualised volatility over that sample was 93.3% against bitcoin's 67.6%, so ether carried 1.38 times the risk for the same compounded return.
- Bitcoin's deepest fall in the sample was 83.8% and ether's was 94.0%. Recovering the first needs a 517% gain; recovering the second needs 1,569%.
- Their daily returns correlated at 0.70 across 3,757 observations, so holding both is closer to one position sized twice than to two positions.
- Across seven start dates the return ranking flipped repeatedly, but ether's volatility stayed between 1.31 and 1.38 times bitcoin's in every window.
Ethereum vs bitcoin: the returns nearly matched, the risk did not
If you are choosing between ether and bitcoin for a portfolio, the question underneath is whether ether's extra volatility bought you anything. Over the longest common daily history the two assets share, it didn't.
We pulled daily closes for BTC-USD and ETH-USD straight from the Coinbase Exchange API, covering 18 May 2016 to 2 September 2026, and ran them through the same risk engine LedgerTouch uses on a live portfolio. Over those 10.3 years bitcoin compounded at 64.7% a year. Ether compounded at 65.8%. That's a gap of 1.1 percentage points a year across a decade, which is noise.
The risk was not close. Across the same 3,757 daily observations, bitcoin's annualised volatility was 67.6%. Ether's was 93.3%. Divide one by the other and ether carried 1.38 times bitcoin's volatility for the same compounded return. Per unit of volatility, bitcoin returned 0.96 and ether 0.70.
Deep losses say it more plainly. Bitcoin's worst peak-to-trough fall in the sample was 83.8%, from 16 December 2017 to 15 December 2018. Ether's was 94.0%, from 13 January 2018 to 14 December 2018. Those two numbers look adjacent. They are not. Climbing back from 83.8% takes a 517% gain. Climbing back from 94.0% takes 1,569%.
That's the whole ethereum vs bitcoin comparison in one line: the returns were a coin toss, the losses were not.
Ether and bitcoin are one exposure sized twice, not two
Daily returns on the two correlated at 0.70 over the full sample. Correlation of 1.0 means two things move in lockstep, and 0 means one day's move in the first tells you nothing about the second. At 0.70, a portfolio holding both is mostly holding one thing at a larger weight.
Against equities the picture changes and the two assets stop differing. Matching both series to the S&P 500's daily closes from the St. Louis Fed's FRED database gives 2,512 common trading days from 6 September 2016 to 2 September 2026. Bitcoin's correlation with the index over that stretch was 0.26. Ether's was 0.27.
So the ether and bitcoin correlation is high enough to make the pair redundant, while their correlations with equities are close enough to be indistinguishable. Neither of those correlations is stable, which is the point of our work on bitcoin equity correlation and on how crypto diversification behaves in the weeks equities actually fall.
One asset caps its supply in code; the other deliberately does not
The two designs promise different things, and the filings are unusually clear about it. Grayscale's annual report on Form 10-K for the year ended 31 December 2024 states it flatly: "Bitcoin has a maximum supply of 21 million." The same filing says the reward "for solving a new block is 3.125 Bitcoin per block" and halves again "after the next 210,000 blocks", expected in mid-2028. Roughly 19.8 million bitcoin were outstanding at that year end.
Ether has no equivalent. Fidelity's prospectus for its ether fund, filed in July 2024, is explicit: "There is no hard cap which would limit the number of outstanding ether at any one time to a predetermined maximum." The same document says that after the Merge "approximately 1,700 ether are issued per day, though the issuance rate varies based on factors such as recent use of the network."
The Merge is the second structural difference. Ethereum switched to proof of stake on 15 September 2022, and the Ethereum Foundation's own documentation puts the fall in the network's energy use at "an estimated 99.95%". Under ethereum proof of stake, "a validator must stake 32 ether", which creates an income stream in ether that bitcoin has no analogue for.
Here is the part that matters for a portfolio: none of this shows up as lower measured risk. A hard supply cap did not make bitcoin calm, and a scarcer issuance schedule after the Merge did not make ether calmer than bitcoin in any window we tested. Monetary design is an argument about the long run. Volatility is a measurement of what already happened.
A US listing arrived in 2024, and the funds it created cannot stake
Access changed on 23 May 2024, when the SEC issued Release No. 34-100224, an "Order Granting Accelerated Approval of Proposed Rule Changes, as Modified by Amendments Thereto, to List and Trade Shares of Ether-Based Exchange-Traded Products". Ether went from an exchange balance to a line on a brokerage statement.
What did not come with it was the staking return. Fidelity's prospectus says the trust "will not participate in the proof-of-stake validation mechanism of the Ethereum network". The iShares ether trust's prospectus, dated 31 July 2025, goes further and bars anyone associated with the trust from employing its ether in "Staking Activities". So an ether ETF holder gets ether's volatility without the one economic feature ether has that bitcoin lacks.
For a UK reader the framing is narrower still. The FCA's InvestSmart guidance, "Crypto: The basics", as it stood in September 2026, says that "while not all cryptoassets are the same, they are all high risk and speculative as an investment", that "crypto is largely unregulated in the UK, so it is highly unlikely you will be covered by the Financial Services Compensation Scheme", and that anyone investing "should be prepared to lose all your money". The 94.0% fall above is what that sentence looks like as data.
What a 1%, 3%, 5% and 10% sleeve did to an equity portfolio
Sizing is where the difference becomes concrete. We took the S&P 500 price index as the base, added a crypto sleeve at four weights, and rebalanced back to the target weight at each calendar year end, from 6 September 2016 to 2 September 2026.
| Portfolio | Annualised return | Annualised volatility | Deepest fall |
|---|---|---|---|
| Equities only | 13.4% | 18.1% | 33.9% |
| +1% bitcoin | 15.1% | 18.1% | 33.9% |
| +1% ether | 20.6% | 20.1% | 34.2% |
| +3% bitcoin | 18.2% | 18.6% | 34.0% |
| +3% ether | 29.7% | 25.5% | 35.0% |
| +5% bitcoin | 21.0% | 19.4% | 34.1% |
| +5% ether | 36.1% | 29.5% | 35.7% |
| +10% bitcoin | 27.0% | 21.6% | 34.3% |
| +10% ether | 47.1% | 36.2% | 42.2% |
Two things stand out, and only one of them is about returns. The first is that annual rebalancing kept the deepest fall almost fixed for every bitcoin sleeve: 33.9% with no crypto, 34.3% with a tenth of the portfolio in it. Capping the weight once a year caps how much of a crypto bear market reaches the portfolio.
The second is the chart above, which plots the volatility column. A 1% bitcoin sleeve left volatility at 18.1%, unchanged. A 1% ether sleeve took it to 20.1%. Reaching that same 20.1% with bitcoin would have taken something between a 5% and a 10% weight. That is the ether volatility premium expressed as position size: roughly one unit of ether does the risk work of five to ten units of bitcoin at the small end. How you hold that sleeve between rebalances matters too, which is the subject of our piece on crypto rebalancing and tolerance bands.
The strongest objection: the ether column is a start-date artefact
Look again at that table and the honest reaction is suspicion. The ether column is spectacular, and it is spectacular because the sample begins when ether closed at $13.18. Start a backtest at a price like that and almost any conclusion is available to you.
So we ran the same two statistics from seven different start dates, each ending on 2 September 2026.
| Start | Bitcoin return | Ether return | Bitcoin volatility | Ether volatility | Ether ÷ bitcoin volatility |
|---|---|---|---|---|---|
| 18 May 2016 | 64.7% | 65.8% | 67.6% | 93.3% | 1.38 |
| 1 January 2018 | 22.3% | 14.2% | 64.8% | 85.2% | 1.31 |
| 1 January 2019 | 48.0% | 44.8% | 62.0% | 81.4% | 1.31 |
| 1 January 2020 | 42.8% | 54.7% | 60.6% | 81.3% | 1.34 |
| 1 January 2021 | 18.6% | 23.3% | 57.5% | 77.7% | 1.35 |
| 15 September 2022 | 41.2% | 13.0% | 47.0% | 64.7% | 1.38 |
| 1 January 2023 | 52.1% | 20.7% | 46.8% | 63.6% | 1.36 |
The return columns have no stable answer in them. Ether wins from a 2020 or a 2021 start and loses badly from 2018, from the Merge date, and from 2023. Bitcoin's lead measured from the Merge, 41.2% a year against 13.0%, is larger than ether's lead from any start date, in the opposite direction. An estimate that swings that far with the choice of first day is not an estimate you can size a position with.
The volatility column does have a stable answer. Ether's annualised volatility exceeded bitcoin's in all seven windows, and the multiple sat between 1.31 and 1.38 every time. It never dipped below 1.3 and never reached 1.4. Whatever the ranking of returns, the ranking of risk didn't move.
That reframes the sleeve table rather than rescuing it. Its ether return column is measuring 2016 and 2017. Its volatility column is measuring something that repeated in every sub-period we could construct.
Where our numbers meet the numbers in the filings
A backtest is worth more when someone else's arithmetic lands near it. The iShares Bitcoin Trust ETF prospectus dated 31 July 2025 states that "the average annualized one-year trailing volatility of bitcoin over the past ten years to date remains elevated at 65%". Averaging every rolling one-year window in our own series gives 66.9%, a gap of 1.9 percentage points on two independent constructions. Ether's equivalent average was 89.9%.
The same prospectus and our data disagree on drawdown, and the disagreement is instructive. iShares reports that "in the 2021-2022 cycle, the price of bitcoin peaked at $67,734 and bottomed at $15,632, marking a steep 77% drawdown". Our Coinbase series finds a deeper fall, 83.8%, because it reaches back to the 2017 peak that the 2021-2022 window excludes. Neither figure is wrong. A drawdown statistic is a statement about the window it was measured over, and a filing that starts its window later reports a gentler asset.
As at 2 September 2026 both are well below their highs. Bitcoin closed at $77,307.36, which is 38.0% under the highest close in this series, $124,720.09 on the Coinbase daily bar dated 6 October 2025. Ether closed at $2,390.83, which is 50.5% under its own high of $4,831.24, dated 22 August 2025 on the same convention. Trailing one-year volatility has fallen for both, to 44.0% for bitcoin and 64.1% for ether, but the ratio between them widened to 1.46, above the 1.31 to 1.38 band of the longer windows.
What this data cannot tell you
Several things, and they matter more than the decimal places.
The price series comes from one venue. Coinbase closes stamp at a fixed hour and another exchange would give slightly different daily returns, though not different enough to move a volatility ratio of 1.38.
The equity base is a price index. FRED's SP500 series excludes dividends, so the equities-only return of 13.4% understates what an index fund delivered, and every sleeve's apparent contribution is correspondingly overstated. The volatility and drawdown columns are far less affected by that, which is another reason the risk side of this study is the sturdier side.
Ten years is one sample, not a law. It contains two full crypto bear markets and one equity crash, and volatility ratios estimated from it are estimates. Backtests describe what a rule would have produced, not what it produces.
Nothing here prices custody risk, exchange failure, spreads, UK tax, or the ether staking income that the price series and the US funds both exclude. If you held ether directly and staked it, your outcome differed from this series in a direction it cannot measure. And a sleeve is only one way to hold crypto; sizing evidence for the bitcoin case specifically sits in our piece on a small bitcoin allocation.
What would change this conclusion
Three measurable things, and none of them is a price forecast.
The first is the volatility multiple. It has sat between 1.31 and 1.38 in every window here, and at 1.46 over the past year. If trailing ether volatility settled at or below bitcoin's across a full cycle, the claim that ether is bitcoin with more risk attached would stop describing the data, and the sleeve arithmetic above would need rebuilding from scratch.
The second is the correlation. At 0.70 the pair is close to redundant. If that halved and stayed halved through a drawdown rather than only in calm markets, holding both would start doing something a larger position in either could not.
The third is staking. Ethereum pays validators who lock up 32 ether, and every US spot ether product quoted above is contractually barred from earning it. If that restriction lifts, ether's return arithmetic gains a term this study has no data for, and the ethereum vs bitcoin comparison has to be run again with it included.