When EBITDA Lies: Industrials & Manufacturing

Melrose Industries and the UK engineering peer set

Manufacturing EBITDA multiples are applied to a number that stops before the capex needed just to stand still, before the cash the profit has not yet become, and before the liabilities a buyer will take off the price. This paper reads one UK industrial’s own accounts through those gaps.

The industrials & manufacturing paper in the Bloor Finance series: a CFO-perspective reading of Melrose Industries PLC (LSE: MRO), the owner of GKN Aerospace, against three UK-listed engineering peers – The Weir Group, IMI and Spirax Group – on one basis: IFRS, sterling, years to 31 December 2025. It is written for boards and finance leaders who will one day hear their business valued at a multiple of EBITDA, and want to know what the buyer will do to the number first.

Prepared by Darren Sack · Research Director · Bloor Finance Desk · Bloor Research International · November 2026

The opening paper in this series argued that wherever an organisation changes how it operates, cost leaves the line EBITDA sees and reappears somewhere else – in depreciation, in working capital, in provisions, in ‘one-off’ adjustments. For industrials it put the problem in one sentence.

“Capital-heavy and automation-led, industrials face the GPU problem in a factory setting: AI-optimisation and robotics assets with genuinely short economic lives are frequently depreciated on the long-life schedules built for plant and buildings.”

“The gap flatters current operating profit and defers a reinvestment wave that is as real as payroll.”

Re-verifying those lines held them as written. The standard does not prevent the mismatch; it requires it to be reviewed. IAS 16 asks for useful lives to be reviewed at least at each financial year-end, names technical and commercial obsolescence as factors, and says an asset’s useful life may be shorter than its economic life. The installed base is large: the International Federation of Robotics counted 542,000 industrial robots installed in 2024 and an operational stock of 4,664,000. Its own working assumption for a robot’s service life has been 12 years, and a pilot study it ran with UNECE found nearer 15 – a plant-like schedule for an asset class whose software and vision systems age much faster (directional, and dated).

Paper 1 named no industrial company. This one does. Melrose is the anchor because its own disclosures carry all five of the lies in Pallabi Mohanty’s keyword brief for the sector: capex below depreciation, a large balance of revenue recognised but not yet billed, three consecutive years of restructuring treated as adjusting, provisions for loss-making contracts and warranties, and a defined-benefit pension history. Melrose is not a weak case. It sets out each of these in its own notes, and its cash-flow table is as clear a bridge from EBITDA to cash as any company in this paper publishes.

Maintenance capex is the capital spending needed to keep existing capacity working – replacing worn tooling, rebuilding furnaces, refreshing the control software – as distinct from growth capex, which adds capacity. EBITDA ignores both by construction: it is struck before depreciation, so the cost of the assets that produce the profit never reaches it. In a software business that matters little. In a factory it is the difference between profit and cash.

No company in this paper discloses a split between maintenance and growth capex, so the honest proxy is capex against depreciation. A ratio persistently below 1.0x means the asset base is being consumed faster than it is replaced, which flatters cash today and builds a reinvestment bill for later.

Melrose states the ratio itself. Capital expenditure in 2025 was £94m (2024: £123m), which the company says was 0.9x the depreciation of owned assets (2024: 1.1x). Against total depreciation and the amortisation of software and development costs – £138m (2024: £142m) – it was about 0.7x. Plant and equipment is depreciated over 3 to 15 years, with lives reviewed annually, and £139m of assets were still under construction at the year-end.

Capex below depreciation is not a saving. It is a loan from the next owner of the factory.

The ratio has to be read with the business. An aerospace structures business that has just finished a multi-year restructuring and closed sites should spend less on owned capacity for a period; Melrose also sold and leased back £28m of land and buildings in 2025. That is the point: EBITDA does not tell the reader which case they are in. The capex ratio, the age of the asset base and the useful-life policy do.

The capex vs opex line decides how much of this year’s spend EBITDA ever sees. Development costs that meet the six tests in IAS 38 must be capitalised and amortised; everything else is expensed. Once capitalised, the spend has left EBITDA permanently – it returns only as amortisation, below the line – and the amortisation period must be reviewed at least annually.

Melrose’s capitalised development is now running down. Additions were £7m in 2025 (2024: £13m) while amortisation charged within adjusted operating profit was £32m (2024: £38m), leaving a net book value of £184m (2024: £213m). Melrose keeps that amortisation inside its adjusted results, because it treats software and development as ‘trading in nature’. The £252m of amortisation on intangibles acquired in business combinations is excluded. Both choices are disclosed; only one of them is in EBITDA.

The peers show how much policy matters. Weir charges development expenditure to the income statement unless it relates to a new product or technology and meets the criteria for capitalisation. IMI and Spirax both capitalise development that meets the criteria. None of the four discloses capitalised development as a share of total R&D, and the amount capitalised in the year is n/d for Weir, IMI and Spirax in the documents read for this paper.

Cash conversion ratio is the share of profit that turns into cash, and the first thing to know is that there is no single definition. Each of the three peers that reports one divides by adjusted operating profit, not EBITDA, and each uses a different numerator:

– Weir: free operating cash flow – after net capex and lease payments – as a percentage of adjusted operating profit. 92% in 2025 (2024: 102%), against a target range of 90–100%.

– IMI: adjusted operating cash flow as a percentage of adjusted operating profit. 96% in 2025 (2024: 92%).

– Spirax: adjusted cash from operations divided by adjusted operating profit. 89% in 2025 (2024: 87%).

Melrose publishes no cash conversion ratio. It publishes something more useful: a bridge from adjusted operating profit to free cash flow, line by line. In 2025 adjusted operating profit was £647m and adjusted EBITDA £785m. Free cash flow before interest and tax was £252m – about 32% of adjusted EBITDA (2024: £23m, about 3%). After interest and tax it was £125m (2024: an outflow of £74m).

£785m of EBITDA, £252m of cash before interest and tax. The largest single item between them is £324m of revenue that has been earned and not yet billed.

That item is the anchor’s specific lie. On its engine risk and revenue sharing partnerships (RRSPs), Melrose recognises revenue as variable consideration where it is not yet due for collection, and holds it as a contract asset called ‘unbilled work done’. The build in 2025 absorbed £324m of cash (2024: £274m). Contract assets stood at £1,732m at the year-end – £1,490m of it non-current – against £1,454m a year earlier. The accounting is constrained, audited and disclosed; Melrose says all 19 of its RRSP programmes are set to be cash positive in 2028. But for now it is profit in EBITDA that is not yet cash, and a buyer pays for the cash.

The rest of the bridge is the normal cost of standing still in manufacturing: £94m of capex, £31m of lease payments, £22m of pension contributions, £31m of restructuring cash, £68m of payments on the Pratt & Whitney GTF powder-metal issue, and £11m of profit released from loss-making-contract provisions that brought no cash at all. Working capital outside unbilled work done was a £33m inflow. The working-capital trend is the peers’ version of the same lie. Weir’s conversion fell ten points in a year, which it attributes to higher inventory supporting its order book and its Performance Excellence programme; working capital rose to 22.4% of sales from 20.7%. Spirax’s working capital was 21.8% of sales (2024: 21.9%). A working-capital peg in a sale is set from exactly these numbers – which is why a business that builds stock in the year before a sale has quietly moved cash from the buyer’s side of the table to its own.

Standard costing values production at a planned cost per unit – materials, labour and an absorbed share of fixed factory overhead – and books the differences as variances. Its trap is overhead absorption. Every unit built absorbs fixed overhead into inventory, so a factory that builds more than it sells moves cost out of the income statement and onto the balance sheet. Margin rises. Cash falls. EBITDA sees only the margin.

IAS 2 limits the effect without removing it. The allocation of fixed production overheads to the costs of conversion is based on the normal capacity of the production facilities; unallocated overheads from low production or idle plant are expensed when incurred; and in periods of abnormally high production the amount allocated to each unit is reduced, so inventory is not measured above cost. ‘Normal capacity’ is a judgement, and the judgement moves margin.

The peers’ disclosures show where the judgement sits. IMI’s Audit Committee says it ‘reviewed the judgements applied to standard costing valuations’ and to excess and obsolete stock; inventories were £396.5m after £62.5m of provisions, and inventory valuation was a key audit matter for its external auditor. Spirax includes an appropriate share of production overheads based on normal operating capacity. Weir includes related production overheads in cost. Melrose held £542m of inventory (2024: £528m), of which £221m was work in progress, and inventory absorbed £32m of cash in 2025 (2024: £71m).

When inventory grows faster than sales, ask how much of the margin is overhead that has been parked in the warehouse.

None of the four discloses its absorption rate, its volume variances or the share of fixed overhead in closing inventory. Those three numbers are the honest gauge, and they are n/d across the peer set.

An order book is only as good as its definition, its margin and its conversion. Some companies report the full contract value, some only firm orders, some only the next twelve months; few disclose the margin in the backlog or how much of it is loss-making.

The peer set shows the range. IMI reports a closing order book for its Process Automation division of £875m (2024: £857m), with order intake of £1,071m split between aftermarket and new construction. Weir and Spirax both cite a strong order book in their reviews without disclosing its value in the documents read. Melrose cites a backlog of more than 15,000 aircraft across Airbus and Boeing – its customers’ order book, not its own – and discloses no group order book figure.

Melrose’s honest substitute is in the provisions note. It held £31m of provisions for loss-making contracts at the end of 2025 (2024: £28m), to be utilised over periods of up to 15 years. In 2025 £11m of that provision was utilised, which the company says ‘benefitted adjusted operating profit’, and £14m was charged on a net basis, mostly as an adjusting item. A contract that was loss-making when it was provided for adds profit to the adjusted line as it is delivered.

A long order book can be a long list of losses. The test is the margin in the backlog, not its length.

Warranty provisions, pension deficits and restructuring costs share one feature: they are real cash obligations that EBITDA never sees. Warranty and restructuring charges are often reported as adjusting items; pension deficits sit on the balance sheet; the cash goes out later. A buyer treats them as debt-like items and takes them off the price.

At the end of 2025 Melrose carried £147m of provisions (2024: £184m), including £19m for warranty, £31m for loss-making contracts, £14m for restructuring and £43m for environmental and litigation matters. Its net retirement-benefit deficit was £27m (2024: £59m), after it completed buy-outs of the GKN UK Group Pension Scheme (Number 4) and the US Consolidated Pension Plan; it paid £22m of pension contributions. Lease obligations were £330m.

Restructuring is the item the series keeps returning to. Melrose excluded £149m of restructuring costs from adjusted operating profit in 2023, £111m in 2024 and £34m in 2025 – £294m in three years – and says the multi-year programme is now complete. Statutory operating profit was £57m in 2023, a £4m loss in 2024 and £600m in 2025, against adjusted operating profit of £390m, £540m and £647m.

“Restructuring is already running – sizeable cuts announced across industrial-goods and chemicals, with automation, tariffs and AI variously cited – which raises the same attribution question the series keeps returning to: how much is genuine operating leverage, and how much is a cost problem wearing an AI label?”

The evidence supports the line as worded. Dow’s January 2026 announcement of about 4,500 role reductions said the plan was ‘in part’ enabled by AI and automation, while its 8-K did not cite AI. Bosch cited tariffs and a €2.5bn annual cost gap for 13,000 German job cuts; ZF and thyssenkrupp Steel, and LyondellBasell’s $1.1bn savings programme, cited the crisis, demand and energy. Demand and energy costs, not AI, dominate the stated reasons. Siemens, reporting under IFRS in euros, cut 6,000 roles, 5,600 of them in Digital Industries, citing weak demand. These are contrasts, outside the UK peer table.

The regulators have been clear for a decade. The FRC’s 2016 thematic review found companies where ‘restructuring costs of a roughly comparable magnitude occur every year’; its 2021 review found 17 of 20 companies (85%) adjusted for restructuring, and said such costs should not be called ‘one-off’. ESMA’s guidelines say issuers should not mislabel items as non-recurring, infrequent or unusual. The peers show both responses. IMI stopped recording restructuring as an adjusting item in 2025, after £54.7m in 2024; Weir excluded £45.2m for its Performance Excellence programme (2024: £35.7m); Spirax excluded £37.0m of Group restructuring costs (2024: nil).

A buyer will apply its multiple to a different number. Advisers describe private industrial valuations as a multiple applied to ‘underlying sustainable EBITDA’ – and the work of due diligence is deciding what is sustainable. In manufacturing that usually means five adjustments:

– Deduct maintenance capex. The buyer’s proxy is EBITDA minus capex. For Melrose: £785m less £94m is about £691m (2024: £682m less £123m, about £559m) – and the buyer will ask whether £94m, at 0.9x owned-asset depreciation, is enough.

– Discount profit that is not yet cash. £324m of unbilled work done in 2025 is earned revenue with collection deferred; a buyer will value the contract asset separately, with its own risk, rather than pay the EBITDA multiple on it.

– Set a working-capital peg on a normalised level. Completion accounts adjust the price for actual working capital against a ‘normal’ target; inventory built ahead of a sale, or receivables factored early – Melrose and others use non-recourse receivables programmes – move the peg.

– Take debt-like items off the price: pension deficits, restructuring and loss-making-contract provisions, warranty claims above the normal run-rate, capex commitments, environmental and legal provisions, and leases.

– Put recurring ‘adjusting items’ back. Three years of restructuring is a cost of running the business, and a buyer will treat it as one.

Bloor quotes no market multiple in this paper. No published UK manufacturing multiple found for this paper met the series’ sourcing standard, and the adjustments above matter more than the multiple: on Melrose’s own 2025 bridge they take £785m of EBITDA to £252m of cash before interest and tax. Melrose’s 2026 guidance is £700–750m of adjusted operating profit and £150–200m of free cash flow after interest and tax.

Read on one basis – UK-listed, IFRS, sterling, years to 31 December 2025 – the peer set shows the same pattern in different strengths. Every company’s adjusted operating profit is above its statutory figure; every company that reports cash conversion defines it differently; and every company carries provisions and pension positions a buyer would deduct. The table does not rank the companies; it shows which honest gauges each one discloses.

Honest gauges – UK industrials peer set (FY2025; IFRS; £m)

All figures are from each company’s 2025 annual report (ESEF filing), years to 31 December 2025, 2024 in brackets. Capex ÷ D&A is Bloor arithmetic: capex as reported, divided by depreciation (including right-of-use assets) plus amortisation of software and development, excluding amortisation of acquired intangibles; Spirax discloses net capex only. n/d means not disclosed in the documents read and is never zero. Siemens (IFRS, euros) and Dow (US GAAP, dollars) are contrasts in section 6 and are not in the table.

The spread in the cash conversion column is the point of the table. Three companies report 89–96%, on three definitions, each against adjusted operating profit; the one company that reports no ratio converts about a third of its adjusted EBITDA into cash before interest and tax, because of a contract asset the others do not carry. None of the four discloses maintenance capex, absorption rates or order-book margin.

Not a different number, but the same numbers with the capex, the cash and the liabilities put back. The industrial edition of the honest-gauge dashboard, applied to the anchor:

• EBITDA minus capex, with capex ÷ depreciation. Melrose: c.£691m; 0.9x owned-asset depreciation as stated, c.0.7x on total D&A.
• Maintenance and growth capex, split. n/d across the peer set.
• Free cash flow ÷ EBITDA on your own published definition, with the bridge. Melrose: c.32% before interest and tax; the bridge is in its annual report.
• Working capital as a percentage of sales, against the peg a buyer would set. Weir 22.4% (20.7%); Spirax 21.8% (21.9%).
• Fixed overhead in closing inventory and volume variances. n/d across the peer set.
• Order book on a stated definition, with margin and the loss-making share. IMI Process Automation £875m; Melrose loss-making contract provisions £31m.
• Debt-like items in one schedule: pension deficit, restructuring, warranty, loss-making contracts, environmental and legal, leases. Melrose: £27m pension, £147m provisions, £330m leases.
• Useful lives of automation, vision and AI assets, separately from plant. n/d across the peer set.

Watch for the remaining When EBITDA Lies industry papers as they publish. For boards and owners who want an early diagnostic read of their honest gauges before a sale process or a results season, the Bloor Finance Desk is available for a confidential conversation.

Analysis only; not investment advice.

Every figure in this paper comes from a named, dated document. The principal ones are listed below with links, and a full Source Note gives the origin, pinpoint and status of each number. Annual-report figures were read from each company’s ESEF filing; page references are to the printed annual report.

• Melrose Industries PLC, Annual Report 2025 (notes 11, 14, 16, 17, 21; CFO’s review; APMs)
• Melrose Industries PLC, Annual Report 2024 (reconciliation of statutory to adjusted results)
• Melrose Industries PLC, AGM Trading Update, 29 April 2026 (2026 guidance)
• The Weir Group PLC, Annual Report and Financial Statements 2025 (notes 3, 6; financial review)
• IMI plc, Annual Report 2025 (notes 4, 5; financial review; Audit Committee report)
• Spirax Group plc, Annual Report 2025 (CFO’s review; notes; APMs)
• Standards and regulatory guidance
• IFRS Foundation, IAS 16 Property, Plant and Equipment (¶51, ¶56–57); IAS 38 Intangible Assets (¶57, ¶104); IAS 2 Inventories (¶13)
• FRC, Corporate Reporting Thematic Review: Alternative Performance Measures, 25 November 2016 and 7 October 2021
• ESMA, Guidelines on Alternative Performance Measures (ESMA/2015/1415en)
• Industry and market sources
• International Federation of Robotics, World Robotics 2025 press release, 25 September 2025; World Robotics 2017
• Dow Inc., Form 8-K and press release, 29 January 2026; Bosch (Reuters, 25 September 2025); ZF, 1 October 2025; thyssenkrupp Steel (Eurometal, 4 December 2025); LyondellBasell Q3 2025 8-K; Siemens (Smart Industry, 19 March 2025) (contrasts, attributed and directional)
• RSM UK, What drives the value of a privately owned industrials business, 9 September 2026; RSM, Debt free, cash free transactions: the devil is in the detail (directional)

Note: For references without a supplied direct report URL, the links lead to the publisher’s website rather than the specific publication.

Print Article for this blog Download here