FIFO vs HIFO vs Share Pooling: One Ledger

Introduction

Comparisons of cost basis methods usually stop at a definition each and a sentence about which is most favourable. That leaves the interesting part out, because the interesting part is arithmetic and the arithmetic is not intuitive.

This page takes one fixed ledger — eleven ETH acquired across six purchases, six disposed of across two sales — and works it through four approaches in full, showing the cost relieved at every step. Nothing is adjusted between runs to favour an outcome.

The prices below are illustrative. They were invented for this worked example: they are not historical quotes, not a record of what ETH traded at on those dates, and nothing here should be read as a price claim. What is demonstrated is the mechanism, and the mechanism behaves the same way whatever the prices are.

Which of these approaches you may actually use is a separate question with a firm answer, and it is not the one most summaries give. That is covered near the end, and in more depth in our cost basis methods guide.

The Ledger

Eleven ETH bought across six purchases, six sold across two disposals, all in sterling. The same ledger runs through every approach on this page.

Totals: 11 ETH acquired for £26,075, 6 ETH disposed of for £17,300, and 5 ETH still held. The purchase at trade 7 matters more than it looks: it lands between the two disposals, so it changes what each approach has available the second time.

No purchase in this ledger falls within thirty days of a disposal. That is deliberate — it keeps the first comparison clean, so the four approaches differ only in which units they select. What a repurchase inside thirty days does has its own section below.

What Each Approach Does

All four answer the same question — when you dispose of part of a holding, which units are treated as having left? — and each answers it differently.

First in, first out

The oldest units go first. In a rising market the oldest are usually the cheapest, so this tends to relieve the least cost and report the largest gain.

Last in, first out

The newest go first. Where a holding was topped up near a local high, this relieves that expensive cost early and reports a smaller gain.

Highest basis first

The most expensive units go first, regardless of when they were acquired. This is not really a chronological rule at all — it selects whichever lot carries the most cost, and it minimises the gain reported on any single disposal.

The section 104 pool, and average cost

Pooling refuses the question. Every unit of the asset carries the same cost — total cost divided by total units — so a disposal relieves that average and which particular unit left does not matter.

The UK section 104 pool is an average-cost method, and that is a finding rather than a simplification. On the ledger below it produces numbers identical to plain average cost at every step, not approximately but exactly, because the two are the same arithmetic. A comparison page is under some pressure to manufacture a difference here; there is none to manufacture. Where the UK method genuinely diverges is in what runs before the pool is consulted — the same-day and thirty-day matching described on our UK crypto tax reporting rules — and this page shows only what those rules do to a number.

The First Disposal: 4 ETH for £11,200

At 18 February 2026 the holder has 10 ETH acquired for £23,625. Four go out.

First in, first out

Oldest first: 2.0 at £1,450, then 1.5 at £2,100, then 0.5 at £2,600. Cost relieved £2,900 + £3,150 + £1,300 = £7,350. Gain: £11,200 − £7,350 = £3,850.

Last in, first out

Newest first: 2.5 at £3,150, then 1.0 at £1,900, then 0.5 at £2,600. Cost relieved £7,875 + £1,900 + £1,300 = £11,075. Gain: £125.

Highest basis first

Most expensive first: 2.5 at £3,150, then 1.5 at £2,600. Cost relieved £7,875 + £3,900 = £11,775. Result: £11,200 − £11,775 = a loss of £575.

Pooling and average cost

Pooled cost £23,625 across 10 ETH gives £2,362.50 per ETH. Cost relieved 4.0 × £2,362.50 = £9,450. Gain: £1,750.

The spread on a single disposal

One disposal, one set of proceeds, one set of purchase records — and a spread of £4,425 between the largest gain and the loss. The proceeds never moved. Only the answer to "which units left" changed.

What Each Leaves Behind

This is the part a single-disposal comparison hides. Relieving more cost now means holding less cost later, and the units left over are not interchangeable between approaches.

After the first disposal the holder still has 6 ETH, then buys another at £2,450 on 9 April 2026 — so 7 ETH going into the second disposal. What that 7 ETH carries depends entirely on which units the first disposal consumed.

• First in, first out — £18,725 of cost carried by the remaining 7 ETH.
• Pooling / average cost — £16,625, which is a pool average of £2,375 per ETH.
• Last in, first out — £15,000.
• Highest basis first — £14,300.

Highest-basis selection reported a loss on the first disposal and is now holding £4,425 less cost than first-in-first-out against the same seven units. That is the same number as the spread in reported gain, and it has not gone anywhere. It has moved to the other side of the ledger.

The Second Disposal: 2 ETH for £6,100

First in, first out

The oldest surviving lot is the January 2025 purchase: two units at £2,600. Cost relieved £5,200, gain £900.

Last in, first out

Newest first: the whole April 2026 unit at £2,450, then one at £2,600. Cost relieved £5,050, gain £1,050.

Highest basis first

Most expensive surviving lot first: 1.5 at £2,600, then 0.5 at £2,450. Cost relieved £5,125, gain £975.

Pooling and average cost

Pool average £2,375 per ETH. Cost relieved 2.0 × £2,375 = £4,750, gain £1,350.

The order reverses

On the first disposal first-in-first-out reported the largest gain and highest-basis selection a loss. On the second, first-in-first-out reports the smallest gain of the four. Nothing about the approaches changed. It had already used up the cheap units, so the cost available to it the second time was higher.

Over the Life of the Holding

Highest-basis selection has realised £4,350 less gain than first-in-first-out. It is also holding £4,350 less cost against the units it still owns — the same number, on the other side.

The arithmetic that settles what these approaches are for

Suppose the remaining 5 ETH are sold one day for some amount, X. Total lifetime gain is total proceeds minus total cost:

Total proceeds of £17,300 plus X, less total cost of £26,075, gives a lifetime gain of X − £8,775 in every case.

That expression contains no reference to the approach used. Whichever it was, the cost relieved cancels: it was subtracted once from the realised gain and added once to the remaining basis. Sell the last five units for £15,000 and every approach lands on £6,225 of lifetime gain — first-in-first-out as £4,750 already realised plus £1,475 to come, highest-basis as £400 realised plus £5,825 to come.

An approach moves gains between years. It does not remove them. Anything one appears to save on a single disposal is carried forward in the cost of what is still held, so the choice matters for the timing of a liability rather than for its eventual size.

Why the Spread Is as Large as It Is

A spread of £4,425 on a single disposal is not a property of the approaches themselves. It is a property of the holding they were applied to, and understanding what drives it tells you when a method decision is worth attention and when it barely matters.

The driver is the range of acquisition prices

The ledger on this page contains units acquired at £1,450 and units acquired at £3,150 — a range of more than two to one. Selection approaches differ precisely by choosing between those extremes, so the wider the range in a holding, the further apart the answers sit.

A holder who accumulated steadily at similar prices would see the four approaches converge almost completely. The arithmetic is identical; there is simply less to choose between when every lot carries a similar cost. That is why the same comparison run over a regular monthly purchase plan produces a far duller table than the one above, and why an article demonstrating dramatic differences has usually chosen a volatile ledger to demonstrate them on.

The second driver is how much of the holding leaves

The first disposal here took four of ten units — a substantial fraction, so which four mattered a great deal. A disposal of a tenth of a holding under the same four approaches would produce a proportionally narrower spread, because each approach would be reaching into a smaller part of the range.

Put the two together and a useful rule of thumb appears. A method decision is worth thinking about when a wide-ranging holding is being reduced substantially. It is worth much less attention when the holding is uniform or the disposal is small, and treating it as a major lever in those cases is spending effort where the arithmetic does not repay it.

What Happens When Prices Fall Instead

Every worked comparison of these approaches, including this one, tends to be built on a holding that has appreciated. That makes the conclusions feel more general than they are, so it is worth stating what changes in the other direction.

The ranking inverts

In a holding where the oldest units were the most expensive — accumulated near a peak and added to on the way down — first-in-first-out relieves the largest cost rather than the smallest. It becomes the approach that reports the smallest gain, or the largest loss, and highest-basis selection and it converge, because in that shape the highest basis often is the oldest.

Nothing about the approaches changed. What changed is which end of the price range the oldest units sit at, which is the only thing chronological rules are actually tracking. Descriptions that present first-in-first-out as inherently unfavourable, or highest-basis selection as inherently advantageous, have silently assumed an appreciating holding and stated the consequence as though it were a property of the rule.

And losses behave the same way

The identity at the heart of this page holds for losses exactly as it does for gains. An approach that realises a larger loss now leaves less cost in the holding, so the same total is recovered across the life of the position rather than created by the selection.

That is worth saying because loss harvesting is often described as though the approach chosen produced the loss. It determines when a loss is realised and against which year's gains it can be set, which can be genuinely valuable — but the loss itself came from the price movement, not from the rule applied to it.

What a Repurchase Inside Thirty Days Does

The ledger above has no repurchase close to a disposal, which is why the pooling column behaved as plain average cost throughout. Move trade 7 from 9 April 2026 to 10 March 2026 — twenty days after the February disposal — and the United Kingdom answer changes shape.

That later purchase is then matched against the earlier disposal before the pool is consulted at all, so the cost relieved is the cost of the March unit rather than the pool average, and the pool is left untouched by that much.

The point for this page is arithmetic only: the same trades, re-dated and not re-priced, produce a different gain under pooling while the three selection approaches are unaffected by the calendar gap. The rule that produces the effect — same-day matching first, then the thirty-day window, then the pool — is explained in full on our UK crypto tax reporting rules, which works a holding through all three stages.

Which of These You May Actually Use

Having shown what the four produce, the question that decides which one matters to you is not preference. It is what your jurisdiction's rules permit — and the answer is narrower than the four-item menu this page's own title implies.

The United Kingdom

No choice is offered. The UK computation uses pooling; specific-identification approaches do not enter it. A disposal is matched against same-day acquisitions, then against acquisitions in the following thirty days, then against the section 104 pool, in that order.

HMRC's Cryptoassets Manual does not name first-in-first-out, last-in-first-out or highest-basis selection as options for individuals. It has not said it prohibits them either — it says nothing about those names at all, and the accurate statement is structural rather than permissive.

The United States

Selection is available, through a single mechanism: specific identification, which must be recorded no later than the date and time of the disposal. Where no identification is made, a default ordering applies — earliest acquired first, within the account or wallet concerned rather than across a whole portfolio.

And highest-basis selection is not a method either of them names

This is the correction that matters most for a page with this title. The United States regulation refers to a method of specifically identifying units "for example, by the earliest acquired, the latest acquired, or the highest basis" — so highest-basis selection is one way of exercising specific identification, not a separate approved method. The acronym does not appear in the provision at all.

Which means the familiar list of four alternatives misdescribes the structure it summarises. There is specific identification, exercisable by any consistent rule including highest basis; and there is a default that applies when no identification was made in time. The full treatment, including why "the IRS permits HIFO" is not a statement the text supports, is in our cost basis methods guide.

So what is this comparison for?

Two things. For a United States holder, it shows what a selection rule is worth before you commit to one — which is exactly the decision the timing requirement forces you to make in advance rather than in hindsight. Knowing that the spread on a substantial disposal can reach several thousand pounds, and that the ranking reverses over successive disposals, is the difference between choosing a standing order deliberately and defaulting into whichever rule a piece of software happened to preselect.

For everyone, it demonstrates the identity above: that these approaches redistribute gains across years without changing their total. That is worth knowing whether or not you have a choice, because it is the strongest argument against treating a method decision as a saving.

The Two Disposals Fall in Different Tax Years

Something in the ledger has gone unremarked so far, and it changes what the cumulative table means. The February disposal and the July disposal are not in the same United Kingdom tax year.

Which year each lands in

A United Kingdom tax year runs from 6 April to the following 5 April. The disposal on 18 February 2026 therefore falls in the 2025 to 2026 tax year, and the disposal on 15 July 2026 falls in the 2026 to 2027 tax year.

Each year brings its own annual exempt amount, and gains and losses are set against each other within a year rather than across the pair. So the cumulative figures in the table above — £4,750 under first-in-first-out, £400 under highest-basis selection — are sums of two separate annual computations, not a single one.

Why that matters more than the total

Because an annual exempt amount that is not used is not carried forward. A holder whose February gain sat below the allowance, and whose July gain sat above it, has a very different outcome from one whose gains were evenly split — even where the total is identical.

This is where a timing difference stops being purely a deferral and starts being worth something in itself. Moving a gain from a year where the allowance is already spent into one where it is not is a real effect, and it is the legitimate version of what method comparisons are usually gesturing at when they talk about savings.

And it is not a lever a UK holder can pull

The obvious follow-up — arrange the selection to put more gain in the lighter year — is not available under a pooling system, because there is no selection to arrange. What a United Kingdom holder can influence is the timing of the disposals themselves, which is a different decision with its own risks and is not a matter of cost basis at all.

The rates and allowances that make this arithmetic concrete, each dated to its tax year, are on our UK crypto tax reporting rules.

How Your Trading Pattern Changes the Picture

The ledger on this page has eight transactions. Real holdings range from a handful of purchases held for years to an active trading record running to thousands of executions, and the practical weight of a method decision differs enormously between them.

A long-term investment portfolio

A portfolio built by periodic purchases and reduced occasionally has few disposals and a long horizon. Each disposal is individually significant, the acquisition prices in the holding may span a wide range, and the spread between approaches on any one of them can be large — much as it is above.

But the horizon cuts the other way. Over a holding period measured in years the identity at the centre of this page does most of the work: whatever an approach defers, a later disposal collects. Unless the deferral crosses a boundary that matters — a year with unusual income, an allowance that would otherwise go unused — its value is the value of holding money for a while rather than a permanent reduction.

An active trading record

At high transaction volume the calculus changes in two ways. Individual disposals matter less because there are so many of them, and the range of acquisition prices in the holding is usually narrower, because trades cluster in time and therefore in price. Both push the four approaches closer together.

What grows instead is the operational burden. Under a selection system each disposal needs its identification recorded in time, and at high volume that is the constraint that actually bites — which is why standing orders exist and why an approach that cannot be maintained in practice is worth less than a slightly worse one that can.

A holding that is both

Many real positions are a long-term core with an active layer around it, and the two parts want different treatment. A selection system will not let you apply one rule to the core and another to the trading layer unless the units genuinely sit in different accounts or wallets, because the identification rules operate at that level rather than by intention.

Where the units do sit separately, the default orderings apply separately too — which is a structural feature worth knowing before a position is arranged rather than after.

This Ledger, for a United Kingdom Reader

Most readers of this page are working out a United Kingdom position, and for them three of the four columns above are informational rather than available. It is worth stating plainly what the ledger produces under the rules that actually apply.

The answer is the pooling column

A United Kingdom holder with this ledger reports a gain of £1,750 on the February disposal and £1,350 on the July one, for a cumulative realised gain of £3,100, and carries £11,875 of pooled allowable cost against the five units still held.

There is no decision to make about that. The pooling column is not the option a UK taxpayer selected as most suitable; it is the computation the rules prescribe, arrived at without any choice having been available at any point.

That is worth dwelling on for a moment, because it changes how the rest of this page should be read by most of the people reading it. A comparison of four approaches naturally invites the question of which to adopt, and for a United Kingdom holder that question has no content — the same way a comparison of driving conventions does not invite a choice about which side of the road to use. What remains useful is understanding the shape of what the rules prevent, and recognising when a tool or an article is quietly describing a different jurisdiction's answer as though it were yours.

Why the other columns are still worth reading

Two reasons, and neither is about choosing between them. The first is diagnostic: if your software reports a figure close to the first-in-first-out column rather than the pooling column, it is computing a different country's rules, and you have found that out by comparison rather than by audit.

The second is that the comparison demonstrates what pooling is protecting against. A system where a holder could select units would let this ledger produce anything between a £575 loss and a £3,850 gain on the same disposal — which is precisely the discretion the United Kingdom structure removes, and seeing the size of it makes the design choice legible.

The stage this ledger does not exercise

One caveat specific to the United Kingdom column. This ledger was built with no purchase within thirty days of a disposal, so the pooling figures above reflect the section 104 pool alone, with neither of the matching stages engaged.

A real holding frequently does have a repurchase in that window, and when it does the United Kingdom answer moves while the selection columns do not — which is the effect the earlier section describes and a worked holding on our UK crypto tax reporting rules runs in full.

What This Comparison Cannot Tell You

A worked comparison is persuasive in a way that makes it worth stating its limits explicitly.

• It says nothing about whether to dispose at all. Every figure here assumes the trades in the ledger happened. The tax consequence of a disposal is a reason to plan its timing, not a reason to make or avoid an investment decision on tax grounds alone.
• It says nothing about market conditions. The prices are illustrative and the pattern they describe — a broadly rising market with one lower purchase — is one shape among many. The section on falling prices above covers the inversion, but no single ledger covers the range.
• It does not model a tax liability. Gains are shown, not tax due, because a liability depends on allowances, other gains and losses, and income — none of which belongs on a page comparing selection rules.
• It assumes complete records. Every approach here requires knowing what each unit cost. A holding whose acquisition history is partial cannot be computed accurately under any of them, and that problem is prior to the choice rather than solved by it.

What it does establish

Three things, all of them arithmetic rather than opinion. That a selection rule can move a single disposal's result by a very large margin. That the ranking between rules reverses over successive disposals, so a rule that looks favourable once is not reliably favourable. And that the lifetime total is invariant, which is the strongest reason to treat this as a timing question rather than a saving.

Those three survive any change to the prices, the quantities or the dates. They are properties of how selection and pooling work, not of the particular ledger used to display them.

There is a fourth thing worth adding, because it is the one that most often gets lost between the tables. Every figure on this page assumes the acquisition history is complete and accurate — that each of the six purchases is recorded with the right date, quantity and consideration, and that no acquisition is missing. None of the four approaches degrades gracefully when that assumption fails. A missing purchase does not produce a slightly wrong answer under one of them and a correct answer under another; it produces a holding whose cost is understated under every single one, with the error propagating into each subsequent disposal. The comparison above is therefore a study in what to do once the record exists, and the record is the part that has to come first.

Conclusion

One ledger, four approaches, and three results worth carrying away.

On a single disposal the spread is large — £4,425 here, between a £3,850 gain and a £575 loss on identical proceeds. On the next disposal the ranking reverses, because the approach that relieved the most cost first had the least left to relieve later. And across the life of the holding the total is identical: lifetime gain works out to X − £8,775 whichever was used, because the cost relieved cancels on both sides.

The other finding is one this page declined to dress up. The UK section 104 pool and plain average cost produce identical numbers at every step of this ledger, because arithmetically they are the same method. What separates them is the matching that runs before the pool is reached, not the averaging itself.

And the answer to the question the title asks: a United Kingdom holder does not choose between these at all, and a United States holder is choosing how to exercise one mechanism rather than picking from a menu of four.

One thing the cumulative table does not show on its face. The two disposals here fall in different United Kingdom tax years, so a gain an approach defers is deferred into a year with its own allowance and its own rate. That is what the choice is actually about. Not how much gain exists, because that is fixed, but which year has to account for it — and each year prices it separately.

Sources

Every figure on this page was produced by running the ledger above and cross-checked as proceeds less cost relieved; all four approaches reconcile. The rules referenced are cited in full on the pages that own them.

• The United Kingdom matching order and the section 104 pool — HMRC Cryptoassets Manual CRYPTO22200, cited and quoted on our UK crypto tax reporting rules.
• The United States specific-identification rules, the default orderings and the highest-basis wording — United States Treasury regulations at section 1.1012-1(j), cited on our cost basis methods guide.
• Which transactions reach this arithmetic in the first place — which movements are disposals and which are not — is the subject of our taxable events guide.
• The records a ledger like the one above has to rest on, and how to rebuild one that is incomplete — our records and reconstruction guide.
• Spending crypto is itself a disposal, so the same arithmetic runs behind a card purchase — worked through in our guide to what a crypto card costs.
• Primary text for the United Kingdom half: HMRC's Cryptoassets Manual at CRYPTO22200 for the matching order, and CRYPTO22250 for the section 104 pool. Read 21 September 2026.
• Primary text for the United States half: the broker reporting and basis determination regulation as published in the Federal Register, which carries the specific-identification wording this page's selection columns rest on. Read 21 September 2026.
• The broker-reporting context for United States identifications — our Form 1099-DA guide.
• Where method choice sits among everything else a holder has to get right is mapped in our complete crypto tax guide.

Frequently Asked Questions

Which cost basis method produces the lowest tax?

On a single disposal, whichever relieves the most cost — on the ledger on this page that was highest-basis selection, which turned a £3,850 gain under first-in-first-out into a £575 loss on the same £11,200 of proceeds. Over the life of the holding the answer is none of them: total lifetime gain came to X − £8,775 whichever approach was used, because the cost relieved is subtracted once from realised gain and added once to the cost still carried. The choice moves a liability between tax years rather than reducing it.

Is the UK section 104 pool the same as average cost?

Arithmetically yes — on the ledger on this page they produce identical figures at every step, because they are the same computation. The difference is what runs before the averaging: the United Kingdom applies same-day matching and then a thirty-day rule, both of which remove units from the computation before the pool is reached. Plain average cost has no equivalent stages, so the two diverge whenever a purchase lands close to a disposal.

Can a UK taxpayer choose FIFO or HIFO for crypto?

No. The United Kingdom computation uses pooling and specific-identification approaches do not enter it: a disposal is matched against same-day acquisitions, then acquisitions in the next thirty days, then the section 104 pool. HMRC's Cryptoassets Manual does not name first-in-first-out, last-in-first-out or highest-basis selection as options for individuals — and it has not stated a prohibition on them either, which is why the accurate description is structural rather than one of permission.

Is HIFO an approved method in the United States?

Not as a named method. The United States regulation describes a method of specifically identifying units "for example, by the earliest acquired, the latest acquired, or the highest basis", so highest-basis selection is one way of exercising specific identification rather than a separate approved approach, and the acronym appears nowhere in the provision. The binding requirement is that the identification exists no later than the date and time of the disposal.

Why did FIFO report the smallest gain on the second disposal?

Because it had already consumed the cheapest units. On the first disposal first-in-first-out relieved £7,350 of cost — the lowest of the four — leaving the most expensive units in the holding. By the second disposal its available cost was higher than the others', so it reported £900 against £1,350 for pooling and £1,050 for last-in-first-out. Nothing about the approach changed between the two disposals; what changed was what it had left.

Do these figures depend on the prices chosen?

The specific amounts do; the structure does not. The prices in the ledger on this page are illustrative and were invented for the example rather than taken from any market record. Change them and every figure changes, but the pattern holds: a wide spread on a single disposal, a reversal of ranking on a later one, and an identical lifetime total across all four approaches.

Does a repurchase soon after a sale change any of this?

It changes the United Kingdom answer and leaves the three selection approaches untouched. If a purchase falls within thirty days after a disposal, the UK rules match that purchase against the earlier disposal before the section 104 pool is consulted, so the cost relieved is the repurchase price rather than the pool average. The selection approaches have no equivalent rule, so the same trades re-dated produce the same result under them.

Financial Disclaimer

This content is not financial advice. All information provided is for educational purposes only. Cryptocurrency investments carry significant investment risk, and past performance does not guarantee future results. Always do your own research and consult a qualified financial advisor before making investment decisions.