Bitcoin ETF for Retirees: Accounts, Drawdowns and Withdrawal Risk

TL;DR. A spot Bitcoin ETF can sit in almost any US retirement wrapper except a standard 401(k) menu, and the wrapper changes the tax outcome far more than the fund choice does. The risk numbers are the part worth knowing before anything else: across 676 trading sessions from 11 January 2024 to 22 September 2026, bitcoin's annualised volatility on our own daily close series was 47.7%, the deepest close-to-close fall was 53.0% (7 October 2025 to 1 July 2026, about nine months), and on 22 September 2026 the price was still 30.7% below that peak. Nothing in the fund pays income. For someone drawing from the portfolio rather than adding to it, the order in which those returns arrive changes the ending balance by tens of thousands of dollars on an identical set of returns. This article explains the mechanics and the arithmetic. It isn't investment advice, and nothing here is a recommendation to hold any particular amount.

What changes when the contributions stop

Most writing about bitcoin allocation is aimed at someone who is still saving. That reader has an option the retiree doesn't: if the position halves, the next paycheque buys more of it and the drawdown becomes an entry point. Take the paycheque away and the same decline is a permanent reduction in the capital that has to produce income for two or three decades. Three things change once withdrawals begin.

• Losses get locked in by the withdrawal itself. Selling shares to fund groceries during a 50% decline converts a paper loss into a realised one, and those shares never participate in the recovery.
• The recovery has to come from a smaller base. An accumulator's portfolio grows during a bear market from contributions alone. A retiree's shrinks from both directions at once.
• The timeline is asymmetric. A 67-year-old may well have a 25-year horizon, but the last fifteen years of it are funded entirely by whatever survives the first ten.

That's the frame for everything below. If you want the general case rather than the withdrawal case, our portfolio allocation article covers it, and what a Bitcoin ETF actually is covers the instrument itself.

Where a retiree can actually hold one

A spot Bitcoin ETF is an ordinary exchange-listed security, so the constraint is never the fund. It's the account. Here's what each wrapper does, with the rules as the IRS states them for 2026.

Two rows deserve a note. The Roth IRA is the only wrapper that combines no lifetime required distributions with no tax on the gain; the phase-outs and conversion arithmetic are in Bitcoin ETF in a Roth IRA. The 401(k) row surprises people, because the answer depends on your employer's plan document rather than on any rule about crypto. Bitcoin ETF in a 401(k) shows how to find out which case you're in, and crypto IRA versus Bitcoin ETF IRA compares the ETF route with holding coins at a specialist custodian.

One tax detail people miss in the taxable row: these funds are grantor trusts, not 1940 Act funds. The trust sells a little bitcoin each month to pay the sponsor fee, and that sale is allocated to you. It arrives on an annual grantor-trust statement as a small amount of proceeds and a matching basis adjustment, so you get paperwork in a year you didn't trade. The US tax guide has the forms.

What bitcoin actually did while the ETFs existed

Every allocation conversation runs on an assumed risk level. Here's the measured one, taken from the daily close series behind our own flow tracker, which starts on 11 January 2024 (the day the US spot ETFs began trading) and runs to 22 September 2026. That's 676 trading sessions.

One caveat about that table. Our series begins the day the ETFs launched, so it contains exactly one full bear market. Bitcoin's earlier declines in 2014, 2018 and 2022 were deeper than 53%, and three years of data gives no basis for treating 53% as a floor. Read it as a lower bound, not a worst case.

Flows say the same thing from the holder's side. Across those 676 sessions the US spot Bitcoin ETFs took in $56.9bn net, but 41% of sessions were net outflows, and the worst 30-session stretch, 15 May to 29 June 2026, pulled out $7.25bn. It ended two days before the price low. Calendar 2026 through 22 September is close to flat at $349m. The series updates daily on our Bitcoin ETF flows page; how to analyse ETF flows explains what a daily number can and can't tell you.

Sequence-of-returns risk, with the arithmetic

This is what separates a retirement portfolio from every other kind, and it's what most bitcoin allocation writing skips. For a buy-and-hold investor the order of annual returns is irrelevant: multiplication commutes, and a 53% fall costs the same first or last. Add withdrawals and that stops being true, because a withdrawal taken in a bad year removes a larger fraction of what's left.

Here's a worked illustration. It's arithmetic on stated assumptions, not a forecast, and the assumptions are simple enough to check.

• Starting portfolio: $1,000,000, at the point withdrawals begin.
• Withdrawal: $45,000 in year one, rising 3% a year, taken at the end of each year, for 20 years.
• Everything that isn't bitcoin returns a flat 4% a year.
• The bitcoin sleeve returns 12% in every year except one, when it falls 53% (the measured drawdown above).
• Withdrawals come out of both sleeves in proportion to their size, which is what most automatic withdrawal plans do.

The only thing that changes between columns is which year the 53% fall lands in.

Same asset, same returns, same portfolio, same withdrawals. Move one bad year from tenth place to first and the 20-year outcome falls by 2.1%, 7.5% or 16.0% of starting capital, depending on the sleeve. That gap is what sequence-of-returns risk means in dollars. Notice that the fourth column grows much faster than the others too: the larger sleeve wins by far the most when nothing goes wrong, which is exactly why this is a trade-off rather than a right answer.

Why the withdrawal order rule exists, and when it fails

The conventional response is to fund spending from cash and bonds first and leave the volatile sleeve alone. Run the same model drawing only from the non-bitcoin side and the 10% sleeve ends at $644,257 whether the crash lands in year one or year ten, because the bitcoin shares are never sold into the decline. Against $489,830 for pro-rata draws into a year-one crash, that's $154,000 from a spending rule rather than a market call.

The catch is that the rule only helps if the sheltered asset outperforms. The model has bitcoin compounding at 12% against 4% elsewhere. Reverse that and never selling it is the worst possible policy: you'd have spent the good asset and kept the bad one. A cash buffer buys time, and time is only worth something if the asset recovers. Nobody knows in advance which case they're in.

Required minimum distributions turn you into a forced seller

The IRS requires withdrawals from a traditional IRA, SEP IRA, SIMPLE IRA and pre-tax 401(k) balances from age 73. Roth IRAs have no required distributions during the owner's lifetime, and neither do designated Roth accounts inside a 401(k). That asymmetry matters more for a volatile holding than for a bond fund. The amount is the prior 31 December balance divided by a factor from the IRS Uniform Lifetime Table, and the factor shrinks with age.

Three consequences follow, and none of them is about bitcoin specifically.

• The divisor uses last year's balance. If bitcoin ran up through December and collapsed in February, the required amount is still calculated on the December figure. That's the forced-seller scenario in its purest form.
• The distribution doesn't have to be cash. An in-kind transfer moves the ETF shares to a taxable account and satisfies the RMD at the transfer-day value. You owe the income tax, but you keep the position and the price exposure.
• The withdrawal raises MAGI. IRA distributions aren't net investment income themselves, so the 3.8% NIIT doesn't hit them directly, but they lift MAGI and can pull other investment income over the $200,000 single or $250,000 joint threshold. The same MAGI sets the Medicare surcharge two years later, on top of the 2026 standard Part B premium of $202.90 a month. Roth distributions don't count toward it.

Fees compound too, in the wrong direction

Over a five-year horizon a difference of ten basis points is noise. Over the 20 or 30 years a 65-year-old should be planning for, it isn't. Sponsor fees on US spot Bitcoin ETFs currently span roughly 0.15% to 1.50%, and the fee is paid by selling bitcoin out of the trust, so the coin backing each share declines by the fee every year regardless of price.

The 8% growth rate is an illustration, not a projection; the point is the ratio between rows, which holds at any rate. The gap between 0.15% and 0.25% is modest. The gap between either and 1.50% is a quarter of the position over two decades. Inside an IRA you can switch funds with no tax consequence, which removes the usual reason for sitting in an expensive legacy product; in a taxable account the switch realises a gain, and the tax-loss harvesting article has that arithmetic. Current fees and assets for every fund are on our funds page, and the selection checklist covers what else to compare.

What happens to the position when you die

Estate outcomes differ more across wrappers than growth outcomes do, and for an asset that might be held untouched for twenty years, that's not a footnote.

• Taxable account. Heirs receive a cost basis stepped up to the date-of-death value. A position bought at $20,000 and worth $200,000 passes with $180,000 of embedded gain erased.
• Traditional IRA. No step-up. Most non-spouse beneficiaries fall under the 10-year rule and must empty the account by 31 December of the tenth year after death, paying ordinary income tax on every dollar at their own marginal rate.
• Roth IRA. The 10-year rule applies here too, but the distributions are tax-free for a qualified account. The heir gets a decade of continued tax-free compounding and then a clean exit.
• HSA. The harshest case. A spouse can treat it as their own; anyone else takes the entire balance into income in the year of death, with no 10-year spread.
• Spousal inheritance is the exception throughout. A surviving spouse can generally roll an IRA into their own and restart the clock, which means none of the 10-year arithmetic applies.

If leaving assets to heirs is part of the plan, note the tension: the asset with the largest expected gain is both the one where a step-up is worth most and the one where a Roth shelter is worth most. Which wins depends on your bracket, your heirs' brackets and your spending needs, and that's a question for someone who can see the whole return.

Who this clearly doesn't fit

There are situations where the arithmetic above simply doesn't work, and recognising one of them is worth more than any allocation number.

• A portfolio that barely covers essential spending. If the safe withdrawal rate is already tight against fixed costs, there's no capacity to absorb a 53% fall in any sleeve. Volatility isn't a risk you're taking, it's a risk you can't pay for.
• A horizon under about five years. Our data shows a nine-month decline followed by more than a year without full recovery, in the mildest bear market bitcoin has had.
• Anyone who would sell at the bottom. The most common failure mode, and a behavioural one. If a 50% fall in a line item would make you liquidate, the size is wrong whatever a spreadsheet says, and nobody learns their own answer until it happens.
• Money already spoken for. A house purchase, a care deposit or a tax bill has a deadline, and a volatile asset doesn't respect deadlines.
• Anyone who needs the income now. A Bitcoin ETF pays nothing. A portfolio built to generate cash flow gets no help from it, and every dollar of spending it funds comes from a sale.

The flip side is worth stating too. None of this rules out a position for someone with surplus capital, a long horizon and a demonstrated tolerance for watching a line item halve. The arithmetic here is neutral on that; it only makes the trade-off visible.

A practical checklist before the first purchase

1. Establish which accounts can hold it before deciding anything about size. For a 401(k) that means reading your plan's summary description, not a general article.
2. Write down the drawdown you're accepting in dollars. Not a percentage of the sleeve, a dollar figure on the total portfolio. A 3% sleeve of $1,000,000 falling 53% is a $15,900 loss; at 15% it's $79,500.
3. Write down the withdrawal order and the rebalancing rule before you need them. Pro-rata versus cash-first was worth $154,000 in the model above, and both are far harder to decide calmly mid-decline.
4. Check the fee and the account type together. Switching funds inside an IRA is free; in a taxable account it isn't, so the fund you pick there is one you'll live with.
5. Talk to a fiduciary adviser and a tax professional. The RMD, Medicare and estate interactions above depend on your full tax picture, and this article can't see it.

To watch what other holders are doing rather than just the price, our daily Bitcoin ETF flow tracker publishes net creations and redemptions for every US fund each business day, and the Crypto ETF Flow Index compresses the last month into one reading. It sat at 87 for bitcoin on 22 September 2026, near the top of its 0 to 100 range, after the price ran from $76,417 to $86,416 in four sessions.

FAQ

Sources and further reading

• IRS, Retirement plan and IRA required minimum distributions FAQs (RMD start age 73; Roth IRAs and designated Roth accounts exempt during the owner's lifetime).
• IRS Publication 590-B, Distributions from Individual Retirement Arrangements (Uniform Lifetime Table factors; the 10-year rule for designated beneficiaries).
• IRS, Net Investment Income Tax (3.8% rate; $200,000 single and $250,000 joint MAGI thresholds).
• Medicare.gov, Medicare costs (2026 standard Part B premium of $202.90 a month and the income-related adjustment).
• Price, volatility, drawdown and flow figures: our own daily series behind the Bitcoin ETF flow tracker, 676 sessions from 11 January 2024 to 22 September 2026.

This article explains mechanics and arithmetic. It is not investment, tax or legal advice, and it does not recommend any allocation, fund or account. Retirement decisions depend on your full financial picture; consult a fiduciary adviser and a tax professional before acting.