The s.104 pool: how UK companies are taxed on listed-share disposals

This post sets out the chargeable-gains regime end-to-end for a UK company holding listed equities — the matching order that applies to every disposal, the running average-cost pool, the December 2017 indexation freeze and the company-specific rules that diverge from those written for individuals. It is a reference piece; the worked mechanics are in Applied A1.

When a UK company sells listed shares, the gain is not worked out share-by-share against whatever it paid for a particular lot. Disposals are matched in a fixed statutory order — same-day acquisitions first, then acquisitions in the previous nine days, then the s.104 pool: the running average-cost record of every share of that class the company holds. The chargeable gain is proceeds less the proportionate pooled cost. Two things trip people up: the 30-day rule they may know from personal capital-gains tax does not apply to companies, and the gain is computed on the pooled cost, never on the moving accounts value.

In brief:

• A company’s listed-share disposals are matched in a fixed order: same-day acquisitions (TCGA 1992 s.105), then acquisitions in the previous nine days (s.107(3)), then the s.104 pool.
• The s.104 pool is the running average-cost record of every share of one class a company holds — quantity and sterling cost, growing on acquisitions and shrinking proportionately on disposals.
• The 30-day “bed-and-breakfast” rule that applies to individuals (s.106A) does not apply to companies — the company analogue runs against the previous nine days, not the following 30.
• Indexation allowance is frozen at December 2017; acquisitions on or after 1 January 2018 attract none.
• The chargeable gain is computed on the pooled sterling cost, never on the FRS 102 carrying amount — the two are tracked in parallel and only the pooled-cost track drives the tax line.

---

The legal framework: how companies are charged on gains

Companies do not pay capital gains tax. They pay corporation tax on chargeable gains — computed under TCGA 1992 rules but folded into the CT computation (TCGA 1992 s.1(2); TCGA 1992 s.2A, substituted by FA 2019 Schedule 1 paragraph 3). The current authority for charging chargeable gains on companies is s.2A. The pre-FA-2019 framing under the original s.2 is no longer operative.

The gain itself is governed by TCGA 1992; the rate at which it is taxed is governed by CTA 2010 Parts 3 and 3A (main rate, small profits rate, marginal relief and the close-investment-holding-company overlay). Profits including chargeable gains are charged under CTA 2009 s.2, with “profits” defined in s.4 to include both income and chargeable gains.

One structural point matters before anything else: the chargeable gain is computed on the TCGA pooled cost, never on the FRS 102 carrying amount. Under FRS 102 a listed equity is normally held at fair value through profit or loss, so its accounting value moves every period. That movement is an accounting result, not a tax event. The engine that matters for chargeable gains is the s.104 pool — a parallel, sterling-denominated, event-driven record that has no connection to fair-value movements. Portive maintains both tracks; only the pooled-cost track feeds the corporation-tax computation. (The fuller account of how the two tracks interact is in the FRS 102 vs corporation-tax reconciliation — the accounting/tax mismatch is the thesis of that piece.)

What the gain is computed on: proceeds, allowable cost and the basic formula

The starting point is straightforward. The chargeable gain — the excess of proceeds over the allowable cost of the shares disposed of — is computed as follows:

Unindexed gain = Proceeds − Allowable cost − Incidental costs of disposal

Chargeable gain = Unindexed gain − Indexation allowance (where applicable)

Proceeds are the consideration received — the gross sale price before broker commission. Where the transaction is not at arm’s length, market value is substituted (TCGA 1992 s.17).

Allowable cost is defined in TCGA 1992 s.38. Section 38(1) restricts it to acquisition consideration plus incidental acquisition costs, enhancement expenditure reflected in the asset at disposal and incidental disposal costs. Section 38(2) defines “incidental costs” as amounts “wholly and exclusively incurred for the purposes of the acquisition or … the disposal” — in practice, broker commission, SDRT (0.5% on most UK share acquisitions under Finance Act 1986 s.87) and any transaction-specific exchange or custodian fee. A general portfolio management fee covering oversight across the whole portfolio is not “wholly and exclusively” for any particular acquisition; it is a revenue management expense, not capital. That distinction — confirmed by the Supreme Court in HMRC v Centrica Overseas Holdings Ltd [2024] UKSC 25 — shapes how every cost line on a broker statement is classified.

Incidental costs of disposal (broker commission on the sale; no SDRT on a disposal) reduce proceeds before the gain is computed.

Indexation allowance — a relief for inflation between acquisition date and disposal date — is frozen at December 2017 and is covered in detail below.

Share matching: the fixed order that decides which cost is used

Share matching — the process of pairing a disposal against specific earlier acquisitions to determine which cost basis applies — follows a fixed statutory order for UK companies (TCGA 1992 ss.105 and 107; HMRC, CG51600 and CG51615):

1. Same-day acquisitions (s.105(1)) — any shares of the same class bought on the same day as the disposal.
2. Acquisitions in the previous nine days (s.107(3)) — the “ten-day rule”; where multiple acquisitions fall in the window, they are matched FIFO (s.107(4)(b)).
3. The s.104 pool — the averaged-cost pool of all shares of that class acquired since 1 April 1982 and not already matched under steps 1 or 2.

The order is fixed. It is not optional, and it is not a choice the company or its adviser makes.

The company-specific rule: nine days back, not thirty days forward

This is the most error-prone area of UK share pooling, and the brief is worth stating plainly.

There is a “bed-and-breakfast” rule for individuals and trustees under TCGA 1992 s.106A that matches a CGT disposal against acquisitions in the following 30 days. That rule’s heading reads “Identification of securities: capital gains tax.” It does not apply to companies.

For companies, the analogous rule is s.107(3), which matches a disposal against acquisitions in the previous nine days. The directions are opposite, the time windows are different and the legal basis is entirely separate. Much commentary on share matching was written with individual CGT in mind; the company-specific regime diverges materially.

The mental model for the company rule: a disposal looks backwards over the previous nine days for any acquisitions to consume first; an acquisition looks forwards over the next nine days for a disposal that might consume it. If no disposal occurs within nine days, the acquisition enters the s.104 pool. HMRC’s framing at CG51615:

“A company acquires shares which would otherwise create or be added to a Section 104 holding. Within ten days it makes a disposal of shares of the same class in the same company. In that case the disposal is identified against the acquisition within the previous nine days, TCGA92/S107(3), on a first in/first out basis if they are made at different times in the ten day period, TCGA92/S107(4)(b).”

The s.104 pool: how it works

The s.104 pool is the statutory vehicle for averaging the cost of a company’s shareholding in a given class of security. TCGA 1992 s.104(1) provides:

“Any number of securities of the same class acquired by the same person in the same capacity shall for the purposes of this Act … be regarded as indistinguishable parts of a single asset growing or diminishing on the occasions on which additional securities of the same class are acquired or some of the securities of that class are disposed of.”

The pool tracks two state variables per security: pool quantity and pool cost (sterling). Every acquisition of the same class adds to both. Every disposal reduces both, proportionately.

On an acquisition: pool quantity increases by the acquired quantity; pool cost increases by (acquisition consideration + s.38 incidental costs). That is the whole entry for post-January 2018 acquisitions. No indexation logic applies.

On a disposal: the cost relieved is pool cost × (disposed quantity / pool quantity) — the proportionate quantity rule. Pool quantity and pool cost both reduce by the disposed amounts. The remaining pool’s average cost per share is unchanged. There is no FIFO or LIFO within the pool; it is a single averaged-cost asset. A part disposal — the s.122 framework concept for any disposal that reduces a holding without extinguishing it — simply takes the proportionate slice and leaves the rest.

A key implication: the pool operates per-company, per-security-class, per-capacity. For most Portive clients, that resolves to one pool per ISIN. Ordinary and preference shares of the same company form different pools. Different ETF share classes under different ISINs are different pools.

Worked example: pool baseline (Example 5.1 from the research specification)

This example applies the pool mechanics with no same-day or nine-day matching. All acquisitions are post-January 2018, so indexation is nil throughout.

A UK Ltd company makes three transactions in ABC plc:

• Day 1: Buy 100 shares at £10.00 each, broker commission £5, SDRT £5 (£10 of s.38 incidental costs)
• Day 30: Buy 100 shares at £12.00 each, broker commission £5, SDRT £6 (£11 of s.38 incidental costs)
• Day 60: Sell 100 shares at £14.00 each, broker commission £7

Day 30 is well outside the ten-day window from Day 1; Day 60 is well outside the ten-day window from both earlier dates. No same-day or nine-day matching applies.

Pool state after Day 1:

Pool state after Day 30:

Disposal on Day 60:

Pool state after Day 60:

The remaining 100 shares carry an average cost of £11.105 per share — the correct base cost for any future disposal from the pool.

This £282.50 feeds into the company’s chargeable-gains total for the period and from there into taxable total profits, which is then taxed at the applicable CT rate. The figure that determines which rate applies is augmented profits — defined in CTA 2010 s.18L as taxable total profits plus exempt distributions received from non-group companies (broadly, non-group dividend income exempt under CTA 2009 Part 9A). The rate question — main rate, small profits rate, marginal relief or the close-investment-holding-company (CIHC) overlay — is governed by CTA 2010 Parts 3 and 3A and is out of scope here. For illustration only, if the company is a CIHC (always 25% under CTA 2010 s.18N, no marginal relief), the CT on this gain alone is £282.50 × 25% = £70.63.

The indexation freeze: operational implications

Indexation allowance — the inflation-uplift relief on acquisition costs — was frozen for companies at December 2017 by FA 2018 s.26. Section 26 amended TCGA 1992 ss.53, 54, 110 and 114, substituting “December 2017” for the variable date references and inserting s.53(1B) and s.54(1B) to prevent indexation for costs incurred after 1 January 2018.

The practical split:

• Post-January 2018 acquisitions: no indexation, ever. Acquisition cost goes into the pool at its sterling amount with no further calculation.
• Pre-January 2018 acquisitions still held: the indexed rise factor is frozen at (RPI_Dec-2017 − RPI_{acquisition month}) / RPI_{acquisition month}, using ONS RPI (CHAW series). The factor never changes. For the pool, TCGA 1992 s.110 governs how the indexed rise is computed and added to pool cost at each “operative event” (acquisition or disposal).
• Indexation cannot create or increase a loss (s.53(1)(b) and s.53(2A)). If the unindexed result is a loss, indexation is zero. If full indexation would push a gain below nil, indexation is restricted to the gain amount.

FA 2008 had already abolished indexation for individuals (Schedule 2 paragraphs 77–83, with effect from 6 April 2008). TCGA 1992 Part 2 Chapter 4 now begins “this Chapter applies only for the purposes of corporation tax.” So indexation — both its availability and its freeze — is a company-specific feature of the chargeable-gains computation.

A note on timing for pre-2018 pools: under s.110 the indexed rise is computed at an operative event (an acquisition or disposal), not on the date the freeze took effect. A pool with a pre-2018 acquisition and no activity through to 2024 computes the indexed rise — capped at December 2017 — at the next operative event, whenever that occurs.

The accounting/tax parallel: why the two tracks must be kept separate

A listed equity held by a UK company under FRS 102 is normally measured at fair value through profit or loss. The accounting carrying amount moves every period. The s.104 pool cost is sterling, fixed until an operative event, and wholly independent of fair-value movements. The gain at disposal is proceeds less pooled cost — not proceeds less carrying amount.

This is the central distinction with the loan-relationships regime, which does follow the accounts (CTA 2009 s.307). Chargeable gains do not. The structural mismatch — fair-value movements in the accounts producing timing differences that never appear in the tax computation until a real disposal — is the subject of the FRS 102 vs corporation-tax reconciliation and is explained there at length.

For the sterling chargeable gain at disposal on a non-sterling holding, the same pooled-cost logic applies: acquisition cost is translated to sterling at the spot rate on the acquisition date; disposal proceeds are translated to sterling at the spot rate on the disposal date. The FX effect is embedded in the gain, not separately computed. (HMRC’s authority for this is CG78310.) The FX treatment of monetary assets such as bonds and gilts is structurally different — for those instruments, the loan-relationships regime applies and FX runs through the accounts rather than the chargeable-gain pool.

Corporate actions: what happens to the pool

Corporate actions change a holding without being ordinary acquisitions or disposals. The TCGA treatment depends on the nature of the action; the pool consequences follow from that classification.

HMRC’s operational reference for share reorganisations is at CG51700 onwards, and for company reconstructions at CG52500 onwards.

The rule of thumb: if the action is a “reorganisation” within s.126/s.127 (bonus issue, split, scrip), there is no disposal and the pool absorbs the change in quantity without a change in cost. If there is a cash element (rights sold nil-paid, partial takeover consideration, tender offer), a disposal arises on that element.

What this means for the journal and the broker statement

Two disciplines follow directly from the mechanics above.

First, every cost component on a buy confirmation must be classified before it touches the pool. Broker commission and SDRT on an acquisition are s.38 incidental costs — capital, into the pool. A periodic custody or portfolio-management fee is not; it is a revenue management expense. Mis-classifying a recurring fee as capital inflates the pool cost and understates gains.

Second, the pool and the carrying amount are different numbers, always. Under FRS 102 the investment account moves with fair value. The s.104 pool does not. The only time they could match is on the day of an acquisition at cost, before any fair-value remeasurement. After that, they diverge. Any process that conflates the two — feeding the accounting book value into a gains computation, for instance — will produce wrong answers.

Applied piece A1 works through a more complex pool scenario — two buys, one sale, and the journal entries that result — with annotated workings. Applied piece A9 covers listed equity options and the s.144 interaction with the pool.

---

Sources

Legislation

• Taxation of Chargeable Gains Act 1992 — s.1(2), s.2A, s.17, s.38, s.53, s.54, s.104, s.105, s.107, s.110, s.115, s.122, s.126, s.127, s.135 (legislation.gov.uk)
• Finance Act 2018 s.26 (indexation freeze at December 2017) (legislation.gov.uk)
• Finance Act 2019 Schedule 1 paragraph 3 (s.2A substitution) (legislation.gov.uk)
• Finance Act 1986 s.87 (SDRT principal charge) (legislation.gov.uk)
• Corporation Tax Act 2009 s.2, s.4, s.307 (legislation.gov.uk)
• Corporation Tax Act 2010 Parts 3, 3A — including s.18N (CIHC rate) and s.18L (augmented profits) (legislation.gov.uk)

HMRC manuals

• HMRC, CG51600 (company share matching — the ordering rule)
• HMRC, CG51615 (the ten-day rule for companies; no-month-overlap note)
• HMRC, CG51700 (share reorganisations)
• HMRC, CG78310 (foreign-currency disposals — per-leg-at-spot translation)

Case law

• HMRC v Centrica Overseas Holdings Ltd [2024] UKSC 25 (capital/revenue classification of fees)

---

Last reviewed: 2026-05-20.