Analysis

Three definitions in the squad cost rule, and what they change

The squad cost rule defines the cost of a player's registration, gate receipts and commercial revenue in its own terms. We had followed the clubs' accounts. This sets out what the rule says in each case, what we now count, how we estimate what the accounts do not give, a fourth correction to cup prize money, and what it does to every club's figure.

By Arthur van de Laak4 October 2026Updated 25 October 2026 · View updates15 min read

Figures as of 9 October 2026, Orange Engine release V2BV · Latest figures · What the engine does not capture

The squad cost ratio divides what a squad costs by the revenue the rule counts. Both halves are defined in Appendix 2 of the Premier League Handbook, and neither is taken straight from a club’s accounts. In three places we had used the accounts’ figure where the rule defines its own. This article sets out the three definitions, what we count now, and what moves. It also corrects a fourth input, cup prize money, which we had entered and never calculated.

Every club’s ratio rises. Bournemouth’s central estimate is now 85.2%, 0.2 points above the 85% line, in an assumption range of 79.8% to 87.4% that spans the line: the estimate does not reliably tell which side Bournemouth are on. No other club’s central estimate crosses a threshold.

1. What a player’s registration costs

The rule. Squad cost includes “Player Transfer Fees”, amortised over the contract, for at most five years where it was signed or extended on or after 12 December 2023, and the rule defines them as all amounts payable “in connection with a Transfer Agreement whether to a Transferor Club or otherwise” (Rules A.1.22(c) and A.1.43). Agents’ fees that a club capitalises are amortised in the same way (A.1.22(f)).

What we did. For players signed since a club’s last accounts, we charged the reported transfer fee over the contract.

What that left out. A club pays more than the fee. Two things are added to it:

• The transfer levy. Rule V.38 of the Handbook: on paying a transfer fee a club “shall forthwith pay to the League a levy equal to 4% of the sum paid”. It applies to transfers from abroad as well as from English clubs, and it is due in full when the player is registered, whatever the instalments. It funds the players’ pension scheme.
• Agents’ fees paid by the club on the signing.

Eleven of the twenty clubs say in their accounts that agents’ fees are part of the cost of a registration; eight name the levy as well. None says how much.

How we measured it. We read all twenty clubs’ 2024/25 accounts and compared what each club recorded as the cost of the players it bought with the fees reported for the same players. Where the accounts give a total for named signings, it is 11% to 17% above the reported fees: Chelsea record six signings at “an initial cost of £263.3m” against about £225m to £233m of reported fees, and Manchester United record £168.7m “including associated costs” against £152.5m. Across the ten clubs where the comparison is clean the middle reading is about 13%.

What we count now. Reported fee plus 12%, for every club. Four of the twelve percentage points are the levy, which is a rule. The rest is an estimate, and we give it a range of 6% to 20%. It is applied in three places: the yearly charge for players signed since the accounts, the book value that leaves when a player is sold, and the profit on that sale.

We use one rate because the accounts fix the level but cannot tell clubs apart: the same club gives different answers by different routes.

2. What counts as gate receipts

The rule. Gate receipts are “the gross revenue recognised from membership fees, general admission and corporate match attendance, from both season tickets and matchday tickets” for the men’s first team (Rule A.1.30). The rule then takes last season’s gate receipts per home match and multiplies by the home matches it counts for this season.

What we did. We took the “matchday” or “gate receipts” line of each club’s accounts.

What that left out. Clubs draw that line in different places. Newcastle’s includes corporate hospitality. Aston Villa’s is “the sale of match tickets” alone, with executive boxes elsewhere. Manchester City’s includes catering. Bournemouth print four separate lines, and we had added all four together, including shop sales and other income, which are not gate receipts.

What we count now. Tickets, memberships and match hospitality. Nothing else: not catering, not merchandise, not conferences or concerts, not prize money. Where a club’s accounts do not say what its matchday line contains, we use the line and say that it is not defined.

One club’s figure moves by a measurable amount. Bournemouth’s gate base is now its ticket income and its “hospitality and events” line, £11.4m, where we had £15.6m. The accounts do not split hospitality from events; on ticket income alone Bournemouth’s ratio would be about two points higher.

How we estimate last season’s receipts, which most clubs have not yet filed, is in How we estimate matchday revenue.

3. What counts as commercial revenue

The rule. Commercial revenue is counted on three different bases (Rule A.1.19):

• sponsorship, advertising and match catering at their full value;
• merchandise at gross profit: sales less the cost of the goods;
• conferences, non-matchday use of the stadium and other operating income at net profit.

What we did. We took each club’s commercial line in full.

What that left out. A club that runs its own shops records the full price of every shirt as revenue. Under the rule only the profit on the shirt counts. Where retail is run by a partner, the club records only its share, and nothing needs taking off.

What we count now. For each club, the commercial line less the cost of the merchandise in it.

• Three clubs file the cost of the goods: Arsenal (£30.9m), Tottenham (£17.2m) and Bournemouth (£1.2m). We use it.
• Chelsea’s merchandising company files its own accounts, with one cost figure for merchandise and stadium tours together. We use it, as set out below.
• Four clubs’ retail is run by a partner or sits outside the club company: Manchester City, Everton, Sunderland and Manchester United. Nothing is taken off.
• Three clubs file their merchandise sales and not the cost: Nottingham Forest, Ipswich and Hull. We take off 45% of sales, on a gross margin of 55%.
• Leeds files its merchandise sales too, and moved its shops to a partner in July 2025. The rule counts merchandise at gross profit either way, so from 9 October 2026, after this article was written, we take off 45% of its 2024/25 merchandise sales, as at the three clubs above: £11.1m, in a range of £8.6m to £16.0m. It is an estimate, not a filed figure.
• Eight clubs file neither. We estimate it: 17% of commercial revenue is merchandise, on a 55% margin.

For a promoted club the rule can take a three-year average of commercial revenue in place of last season’s figure. Where it does, as at Ipswich, the cost of the goods comes off each of the three seasons.

The margin. Only two clubs’ accounts give both merchandise sales and their cost, each for two years: Tottenham (56% in both) and Bournemouth (52%, then 35%). The mean of the four readings is about 50% and the median about 54%. Our 55% is a chosen central value near the median. It is not an average, and we give it a range of 35% to 65%, which takes in all four.

A third club’s accounts, read for this note, are not a like-for-like reading. Chelsea’s merchandising company made a gross margin of 29% on merchandise and stadium tours together. Its cost of sales covers the tours as well as the goods, so it does not give a margin on merchandise. We use Chelsea’s own figures for Chelsea only, and keep 55%, in a range of 35% to 65%, for the clubs we estimate.

The share. Four clubs print a merchandise line: it is 9% of commercial revenue at Bournemouth, 14% at Tottenham, 21% at Ipswich and 26% at Nottingham Forest. We use 17%, with a range of 9% to 26%. Both the margin and the share are thin, and the estimates built on them are the least certain figures in this note.

Newcastle. Newcastle took its shops in-house in May 2024. The filing indicates retail costs rose materially, but does not disclose the absolute amount. The model therefore uses the league-wide merchandise method; the true figure could be higher. That gives £8.3m, in a range of £3.4m to £18.3m.

Chelsea. Until October 2023 Chelsea’s megastore was run by a Nike company. Since then Chelsea has run it through its own subsidiary, Chelsea FC Merchandising Limited, whose accounts for 2024/25 show merchandise sales of £27.2m, stadium tour income of £5.3m and a cost of sales of £23.0m for the two together. The accounts do not split that cost. The merchandise cannot have cost more than £23.0m, and not less than £17.7m unless the tours ran at a loss. We take £19.2m, the cost shared in proportion to sales, with those two figures as its range.

What is still not counted properly. Conferences, concerts and other non-match income should count at net profit. No club publishes that profit, so we still count them in full. That overstates revenue, so the ratios here are if anything still slightly low. It matters most where the stadium earns most on non-match days.

A fourth correction: cup prize money

The first three corrections are definitions. The fourth is an input: a figure we had entered for each club and never calculated.

What the rule counts. Revenue includes a three-year average of what a club received from the FA Cup and the EFL Cup (Rule A.1.29(f)). For the test in March 2027 the three seasons are 2023/24, 2024/25 and 2025/26.

What we had done. We had entered one figure for each club, with no calculation behind it. Every one was too high, by a factor of 1.3 to 16.

How we found it. We audited every figure we had typed in without a calculation. The cup figure was one of them, and at Bournemouth it decided on which side of the 85% line the central estimate fell.

What we count now. Each club’s cup ties over the three seasons, 337 ties in all, with what each one paid: the FA’s published prize fund for the round, and the FA’s published broadcast fee where the tie was chosen for live television. The EFL publishes no prize money for its cup, so we use the figures reported in the press. They are small.

Bournemouth’s figure falls from £2.5m to £0.3m, Chelsea’s from £8.0m to £1.9m and Coventry’s from £0.8m to £0.6m. Less cup income means less revenue under the rule, so this raises every club’s ratio, by between about 0.1 and 1 percentage point.

What is still not known. Three things.

• The FA published no broadcast fees for 2024/25, and none in any season for semi-finals and finals. For a 2024/25 tie that was televised we use the average of the fees for the same round in the seasons either side. For semi-finals and finals the central figure counts no broadcast fee.
• The EFL publishes neither prize money nor broadcast fees.
• The rule does not say whether a club’s share of the gate at an away tie or a Wembley tie is a “distribution”. The text can be read either way, and the League’s manual is not public. We treat it the same way for all twenty clubs: it is in the range, and it is not in the central figure.

So each club’s cup figure has a low and a high. The low counts only what the FA has published. The high adds an allowance for every fee that is not published and a ceiling on gate shares. Less cup income means a higher ratio, so the evidence-only figure is the top of the ratio’s range.

The figures are provisional. They will be replaced when the unpublished amounts are published, or when the League’s manual says what a distribution is.

How far out these could be

The ranges printed beside each club’s ratio now include all four: the 6% to 20% on registration costs; the merchandise margin and the merchandise share, each varied over its own range; the weights and growth rate of the matchday estimate that we set out on 3 October; and the cup figure, drawn between its low and its high.

Last season’s matchday income, where a club has not yet filed it, already carried a general allowance of 7% either way: the one we give a figure not yet filed that has no bounds of its own. We have kept that allowance in full and added the matchday method’s own ranges on top. This is deliberately cautious, and it probably counts some of the same uncertainty twice. We will reset it by comparing these estimates with the filed figures once enough clubs’ 2025/26 accounts are out.

Together these widen the ranges by up to 0.8 points. The wage estimate is still by far the largest uncertainty at every club.

What moves

No club’s central estimate crosses 115%. One is now on the 85% line: Bournemouth’s central estimate is 85.2%, 0.2 points above it, the assumption range is 79.8% to 87.4%, which spans the line. The estimate does not reliably tell which side Bournemouth are on. Its range is wider than it would otherwise be, on purpose: three more of its inputs rested on judgement, and their uncertainty is now in it. On the three definitions alone it was 84.5%; the cup figure is the rest. Two more clubs are within two points of 85%: Newcastle at 84.2% and Brentford at 83.9%, each with an assumption range that spans the line. Chelsea moves six points, to 94.0%: about three from the cost of its merchandise, two from the cost of its registrations and one from cup prize money. Ipswich moves a little over five points, to 94.1%: about three and a half from the cost of its registrations, one and a half from the three-year average of its commercial revenue and under half a point from cup prize money.

Each club’s article carries a dated note with its own figures and what moved them.

What we got wrong

In all three places we used a figure from the accounts where the rule sets its own definition, and in all three the rule’s text was available to us. We found the first by reading the rule on agents’ fees again, and the other two by reading the definitions beside it. A fourth error was simpler: at one club we had added four revenue lines together and called the sum matchday income.

The cup figure was a different kind of error. There was no definition to misread: we had typed in a figure for each club and never calculated it, and every one was too high. We found it only by auditing every figure of that kind.

Before publishing this we read every filed figure it newly depends on against the page it comes from, a second time and by a second reader.

When these stop being estimates

Each club’s 2025/26 accounts, due by spring 2027, will give the cost of the players it signed that year, last season’s matchday income and, for clubs that file them, merchandise figures. We will report how close each estimate was.