Football finance, from the outside inThe Orange LedgerFor professionals

Chelsea · Financial analysis

Chelsea's sales count twice: how one player sale enters both PSR and SCR

When Noni Madueke left Chelsea for Arsenal, the profit went into two sets of rules at once: the last of the old loss limit, and the first of the squad cost rule. What each one counts is different, and on their public accounts, before deductions the rule allows that are not public, Chelsea have almost no room left under the old loss limit.

By Updated · View updates10 min read

Figures in this article: release V2CO, 11 November 2026 · Latest figures · What the engine does not capture

Bar chart of Chelsea's losses on public accounts, each against a £105m limit: window to 2023/24 £39.6m; window to 2024/25 £180.3m; the two filed seasons of the final window, 2023/24 and 2024/25, £104.7m.
On Chelsea's public accounts the windows to 2023/24 and 2024/25 show losses of £39.6m and £180.3m, and the final window's two filed seasons £104.7m, against a £105m limit. Each figure is the loss on public accounts at Chelsea FC Holdings Limited, not the PSR loss. The deductions the rule allows are not fully public and would reduce it; a lower value accepted by the League for the women's team sale would increase it. Chelsea FC Holdings Limited accounts; Orange Engine release V2CM. Orange Ledger.

Latest note, 7 November 2026: editorial, no figure changed. The correction note is at the end.

On 18 July 2025 Noni Madueke left Chelsea for Arsenal for a reported £48.5m. Arsenal paid once. Chelsea will count that sale in two sets of rules, in two different ways, and the same is true of every player they sold between 2 February 2024 and 30 June 2026.

The old rule, the Profitability and Sustainability Rules (PSR), limits a club’s losses over three years. The Premier League replaced it from 5 June 2026, but one last calculation covers 2023/24, 2024/25 and 2025/26: clubs had to submit it to the League by 31 October 2026, and its result is not public (Rule E.48). The new one, the squad cost ratio (SCR), limits what a club spends on its squad as a share of its revenue, and profit on player sales counts towards that revenue.

PSR takes a sale’s profit whole, in the year it is booked, and Madueke’s falls in 2025/26, the last PSR year. SCR takes the same profit a third at a time: it adds up the profit or loss on player sales in the last six transfer windows and divides by three (Rules A.1.13 and A.1.53). So a third of it counts at the test of 1 March 2027, a third at 1 March 2028, and nothing at 1 March 2029, when his window has dropped out.

Take a sale in the summer 2025 window that made a profit of £15m. The figure is an illustration, not Madueke’s profit, which is not public. In PSR the whole £15m counts in 2025/26, the last year of the final window. In SCR the average adds £5m to revenue at the test of 1 March 2027 and £5m again at 1 March 2028. At 1 March 2029 it adds nothing.

What that profit was, the public record does not say. Chelsea’s accounts for 2024/25 give one figure for the 15 players sold after that year end, up to the signing of the accounts on 14 October 2025, Madueke among them: £31.8m in total.

Madueke’s move to Arsenal made me ask a simple question: where does that profit count? The answer turned out to be: twice.

In PSR: the whole profit, in 2025/26

The windows so far

These are Chelsea’s windows on their public accounts, at Chelsea FC Holdings Limited, the company that holds the club. Each figure is the loss on public accounts, not the PSR loss itself, and the two can differ either way. The deductions the rule allows for academy, women’s football and community spending are not fully public, and they would make the PSR loss smaller. A lower value accepted by the League for the women’s team sale, described below, would make it larger. On two cases where a Commission published a figure, Everton’s window ending 2021/22 and Nottingham Forest’s ending 2022/23, this measure came out about £45m and £21m higher than the published loss even after the estimates.

Window ending Loss on public accounts, £m Limit, £m
2023/24 39.6 105.0
2024/25 180.3 105.0

Our estimates of the allowed deductions take the second window’s loss to £147.5m, against the same £105m limit. The two windows differ by one season each way. The second gains 2024/25, a loss of £262.4m before tax, and loses 2021/22. Which company is read matters a great deal: at the club company alone, Chelsea Football Club Limited, the same two windows are £319.5m and £460.1m. The difference is mostly the sales inside the group described below.

What the final window needs

The last calculation keeps 2023/24 and 2024/25 and adds 2025/26. Chelsea’s accounts for 2025/26 are not filed. The two filed seasons, after the add-backs for depreciation and intangible assets that the public accounts allow us to count, come to £143.6m and a loss of £248.3m: together a loss of £104.7m against a limit of £105m.

That leaves £0.3m. It is what remains of the £105m after the two filed years alone, on public accounts: before any further deduction the rule allows, before any adjustment by the League, and before 2025/26, whose result is not yet known. It is not confirmed PSR headroom. The deductions for academy, women’s football and community spending are not public, and each pound of them adds to the room. Our estimate for 2023/24 alone is £19.9m, which would make it £20.3m, and we hold none for 2024/25 or 2025/26. The women’s team sale, described below, counts here at the value Chelsea filed, and a lower value set by the League would take room away. On the public accounts, then, Chelsea’s 2025/26 could show a loss before tax of no more than about £14.5m, with add-backs like 2024/25’s, for the window to end inside the limit.

How much player-sale profit that takes depends on the rest of 2025/26, which is not yet public. As an assumption only, we hold 2024/25 flat, which gives a loss before player sales of £306.2m after the same add-backs. On that assumption the window needs £305.8m of player-sale profit in 2025/26.

What we can count comes to £48.0m. The filed £31.8m covers the sales up to 14 October 2025. Our estimates put the sales after it at £16.2m; they include Raheem Sterling’s departure at a loss and Marc Cucurella’s sale in June 2026.

On the assumed base, that £48.0m would end the window at a loss of £362.8m on public accounts, against a £105m limit, before the academy, women’s football and community deductions. It shows the size of the gap on a year we have assumed rather than read, not the result. Two items would make the gap smaller, pound for pound. 2024/25, the year held flat, carried £50.2m of legal and regulatory charges, which may not recur. And the base leaves out the £17.0m reported for Enzo Maresca’s release, which falls in 2025/26.

Sales inside the group, which count in PSR only

Not everything Chelsea have sold reaches both sets of rules. In 2022/23 Chelsea FC Holdings booked a gain of £76.3m on selling the hotels and car park at Stamford Bridge to Blueco 22 Properties, and in 2023/24 £198.7m on selling Chelsea Football Club Women Limited to Blueco 22 Midco, both companies under the same ownership. Both are in the profit before tax of those years, as filed, so both are in the PSR figures above. Without the women’s team sale, the two filed seasons of the final window would be a loss of £303.4m rather than £104.7m.

Neither reaches the average of player-sale profit that SCR adds to revenue. That average counts only disposals of registrations of players and the release of contracted managers (Rule A.1.29(g)), and profit or loss on disposing of investments is left out of commercial revenue (Rule A.1.19(e)).

The rules on deals with related companies apply as well. Chelsea’s accounts say the hotels’ consideration was subject to Premier League assessment under its associated party transactions rules, and that the women’s team consideration was subject to the League’s fair market value assessment. The rules as they stand are Rules A.1.28 and E.49 onwards of the 2026/27 Handbook; the versions in force at each sale differ. For the hotels, Chelsea’s accounts to 30 June 2024 record the League’s final determination of 13 July 2024, which took £6.0m off the £76.3m; Chelsea booked that in 2024/25, so it is in the figures above. Whether the League adjusted the women’s team sale is not public. These figures take it as filed: a lower value accepted by the League would make the PSR loss larger.

In SCR: a third at a time, test by test

The test looks at six transfer windows, from the first working day after the seventh window back (Rule A.1.53; the two summer 2025 windows count as one). Each sale therefore counts at a third for as long as its window stays in. The test of 1 March 2027 reaches back to 2 February 2024, which is why that date matters.

Period Profit, £m 1 March 2027 1 March 2028 1 March 2029
February to June 2024, our estimate 73.8 in out out
2024/25, filed total 57.9 in £20.0m in out
July to 14 October 2025, filed total 31.8 in in out
15 October 2025 to 2 February 2026 -11.4 in in out
June 2026 27.6 in in in
Compensation for Maresca, reported June 2026 17.0 in in in
2026/27, sales to 2 September 2026 168.1 in in in
Average counted, a third of the total 121.6 84.4 70.9
Ratio, with costs as they stand 91.3% 97.5% 100.0%

The line for these ratios is 85%, the green threshold. A club above it at a March test is notified of a breach and retested on its accounts, and from 2027/28 it pays a levy if it is still above. Points are deducted only above the red threshold, which is 85% plus the club’s allowance: 115% at the start, moving as the allowance is cut or rebuilt.

The two filed totals are not listed sale by sale. The 2028 test’s period starts on 4 February 2025, and we place £20.0m of 2024/25’s total after it, from three sales: Bashir Humphreys’s, Marcus Bettinelli’s, whose fee was not disclosed, and Kepa Arrizabalaga’s, which we date at the year end. We count the compensation reported for Enzo Maresca’s move to Manchester City beside player sales, under Rule A.1.29(g) on the release of a contracted manager; that is our reading of a confidential settlement, and without it the 2027 ratio would be about 92.2%. It is dated here by the report; if it belongs to his resignation in December 2025 it leaves the 2029 test. The ratios hold costs where they stand and add no new business, so they show what the expiry does, not a forecast.

The sales of July to 14 October 2025 count only through the filed £31.8m. The largest reported fees among them were Madueke’s £48.5m, Christopher Nkunku’s £32.0m and Joao Felix’s £26.2m. The other sales of £10m of profit or loss or more:

Date Player Fee, £m Profit or loss, £m SCR tests
13 Apr 2024 Lewis Hall 28.0 26.6 2027
28 Jun 2024 Ian Maatsen 37.5 35.6 2027
30 Jun 2024 Omari Hutchinson 20.0 19.0 2027
21 May 2025 Bashir Humphreys 14.0 13.3 2027, 2028
28 Jan 2026 Raheem Sterling 0.0 -15.2 2027, 2028
15 Jun 2026 Marc Cucurella 47.4 24.4 2027, 2028, 2029
6 Jul 2026 Tyrique George 18.0 17.1 2027, 2028, 2029
13 Jul 2026 Andrey Santos 48.0 35.4 2027, 2028, 2029
9 Aug 2026 Trevoh Chalobah 25.7 24.4 2027, 2028, 2029
27 Aug 2026 Liam Delap 45.0 17.3 2027, 2028, 2029
28 Aug 2026 Nicolas Jackson 47.5 22.6 2027, 2028, 2029
1 Sep 2026 Marc Guiu 14.6 10.9 2027, 2028, 2029
2 Sep 2026 Enzo Fernandez 125.0 49.3 2027, 2028, 2029

Fees are as reported, and each sale’s profit is our estimate: the accounts give each year’s total, not each sale’s. A sale booked from 2 February 2024 to 30 June 2026 counts in the last PSR window and in SCR. One booked from July 2026 counts in SCR alone.

Try it: one profit, both rulebooks

The slider needs JavaScript. Its starting point gives the figures in the sections above.

Set Chelsea’s 2025/26 player-sale profit and see the final PSR result and the three SCR tests move together. Figures read from filed accounts, figures from our model and figures we have assumed are marked differently, because they do not deserve the same trust. For the 2029 test the slider counts 0.575 of each pound added or taken away, our calculation: £27.6m of today’s £48.0m was booked in June 2026, inside that test’s period.

The assumed base behind the slider is 2024/25’s loss before tax of £262.4m, of which £57.9m was profit on player sales: £320.4m before them, and £306.2m after the same add-backs. Of the £48.0m starting point, £16.2m is our estimate, with Cucurella’s sale dated at its completion, 15 June 2026. If Chelsea booked it in 2026/27, the later 2025/26 sales fall from £16.2m to about -£8.2m.

My view is that one sale should count once. Letting the same profit rescue the last PSR test and then prop up the first squad cost tests is a gap in the handover between two rulebooks, not a reward for good trading.

Chelsea are trading their way through the rules, not out of them. Each sale buys a season of headroom, and the next test always needs another one.

Sources and caveats.

PSR. From the filed accounts of Chelsea FC Holdings Limited (02536231) and Chelsea Football Club Limited (01965149), each read twice. The loss on public accounts is profit before tax with depreciation and intangible add-backs, not the PSR loss: the deductions the rule allows would reduce it, and a lower accepted value for the women's team sale would increase it. No figure here is a finding under any rule.

SCR. The Orange Engine, release V2CM. Fees are as reported, and each sale's profit is our estimate, except the sales from July to 14 October 2025, where the accounts' total is used (Chelsea FC Holdings accounts to 30 June 2025, note 30). We had modelled £82.4m for those sales and have used the filed figure since 31 October. The ratio of 91.3% is the published estimate for the test of 1 March 2027; the 2028 and 2029 ratios hold costs flat. The figures follow our corrections to the Chelsea estimate of 31 October, 1 November and 3 November 2026, which the notes on Chelsea's article set out.

Assumed. The 2025/26 base is 2024/25 held flat. It is an illustration, not a forecast.

Rules. Premier League Handbook 2026/27: Rule E.48; Appendix 2, Rules A.1.13, A.1.17, A.1.19, A.1.29, A.1.31, A.1.48 and A.1.53; Rules A.1.28 and E.49 onwards.

Independent estimate, not an official calculation, and not advice.

Corrections and updates (2)

5 November 2026 · Correction · no figure changed

Correction, 5 November 2026. Wording changed to state our figures against the limit without a verdict: the table’s column measuring each window from the limit is removed, and two amounts measured from the limit are given as the losses they come from (£42.5m as a loss of £147.5m, £257.8m as a loss of £362.8m), in the text and in the slider; no figure changed.

7 November 2026 · Editorial · no figure changed

Editorial, 7 November 2026. Wording and order only; no figure changed. The calculator had described its PSR result as before any deductions; depreciation and the amortisation of intangible assets other than players are already left out, and only the academy, women’s football and community deductions are not taken off. The same words in the text are corrected. The £0.3m left under the limit is now qualified where it is stated, a worked example shows how £15m of profit counts in each rulebook, and the 85% line is explained as the green threshold, with points deducted only above the red threshold. The calculator now runs to £350m and marks the point, about £305.8m, where the final window reaches the limit on the assumed base. The view near the end is now marked as the author’s. The sales inside the group now sit in the PSR section, and the headings follow the two rulebooks in turn. The loss on public accounts is no longer called a ceiling on the PSR loss: the deductions the rule allows would reduce the PSR loss, and a lower accepted value for the women’s team sale would increase it. The £0.3m is described as what remains after the two filed years alone, not confirmed headroom. The chart at the top is redrawn with the same figures and the same correction.