EBITDA: What It Deliberately Leaves Out

Four letters, four deletions, one purpose

Read the acronym as an instruction list. Take earnings, then put back the interest, because that reflects how the company borrowed rather than how it trades. Put back the tax, because that reflects where it is incorporated. Put back depreciation and amortisation, because those reflect what it bought years ago and how quickly its accountants chose to write it off. What remains is meant to be the operating engine, unclothed by financing, geography and accounting history.

The measure was built for a specific job, and it does that job honestly. If you are comparing two supermarket chains, one of which leases its stores and carries no debt while the other borrowed heavily to buy its freeholds, their net incomes are not comparable and their EBITDAs roughly are. The same holds when a buyer wants to know what a target earns before deciding how to finance it, since the acquirer's own capital structure will replace the seller's. In both cases the additions back are not a cosmetic trick — they are the point of the exercise.

One company, walked all the way down

Start at $50 million of revenue. Cost of goods sold takes $20 million, leaving $30 million of gross profit, a 60.0% gross margin. Operating expenses before depreciation take another $17 million, so EBITDA is $13 million and the EBITDA margin is 13 ÷ 50 = 26.0%. That is the number that goes in the press release, and everything below it is where the company actually lives.

Depreciation and amortisation cost $6 million, so operating profit is $7 million and the operating margin is 14.0% — the first margin has just halved. Interest of $2 million brings pre-tax profit to $5 million. Tax at 25% is $1.25 million, leaving net income of $3.75 million and a net margin of 7.5%. From 26.0% to 7.5% is a fall of more than two thirds, and every step of it was a real payment or a real consumption of an asset. EBITDA is 13 ÷ 3.75 = 3.47 times net income, a gap of $9.25 million a year.

Same headline, 2.4 times the profit

Now build a second company with the same revenue, the same cost of goods and the same operating expenses, so it reports exactly the same $13 million of EBITDA at exactly the same 26.0% margin. The only differences are that it rents rather than owns, so its depreciation is $1 million instead of $6 million, and that it carries no debt, so it pays no interest. Its operating profit is $12 million, a 24.0% margin. Tax at 25% takes $3 million, leaving $9 million of net income and an 18.0% net margin.

Nine million against three point seven five. From an identical EBITDA, one company earns 2.4 times what the other does, and no reader of the headline number would know. Notice where the difference sits: depreciation is 12.0% of revenue in the first company and 2.0% in the second, and that ten-point gap is almost the whole story. EBITDA did not hide this deliberately — it was designed to remove exactly that difference. The mistake is not the measure, it is treating a measure built to erase a distinction as if it preserved one.

Depreciation is a delayed cost, not a non-cash fiction

The standard defence of adding depreciation back is that it is a non-cash charge. That is true of the accounting entry and misleading about the economics. Cash did leave the business — it left when the asset was bought — and depreciation is simply the accountant's attempt to match that outflow to the years in which the asset earns. Calling it non-cash is like calling a mortgage payment non-cash because you already bought the house. The money moved; the recognition is what was spread.

The distinction matters most when the asset has to be replaced. A one-off purchase that is never repeated genuinely is a past event, and adding its depreciation back tells you something useful about the ongoing business. But a fleet of delivery vans, a bottling line or a network of base stations wears out on a schedule and gets bought again. For those, the depreciation charge is not a memory of an old outflow but an estimate of a recurring one, and adding it back does not remove a distortion — it creates one.

The maintenance-capex floor EBITDA never shows

Give the capital-intensive company a concrete asset base: $60 million of gross property, plant and equipment with a ten-year useful life. Straight-line, that is exactly the $6 million annual charge already in the accounts. In a steady state where the fleet is replaced as it wears out, cash capital expenditure is also $6 million a year, and the arithmetic closes neatly: EBITDA of $13 million, less $6 million of capex, less $2 million of interest, less $1.25 million of tax, leaves $3.75 million — exactly the reported net income. That is not a coincidence. When depreciation equals replacement spending, net income is the cash the owners can actually take.

Now allow the obvious: replacement costs rise. If the equipment costs 3% more each year, the machine you bought for $60 million costs 60 × 1.03¹⁰ = $80.63 million to replace after its ten-year life, which is $8.06 million a year set aside rather than $6 million — 34.4% above the book charge. Rerun the same steady state with that figure and the owners are left with 13 − 8.06 − 2 − 1.25 = $1.69 million, roughly 45% of the reported net income and 13% of the EBITDA everyone quoted. Nothing was mismeasured. The accounts recorded historical cost faithfully; the business simply has to buy at tomorrow's prices.

When EBITDA is the right number, and what to put beside it

Use it where the comparison it enables is the comparison you want. Judging whether a factory manager improved throughput and cost control this year, comparing two divisions financed by the same parent, or valuing a target you intend to refinance are all fair uses, because in each the financing and the historical asset ledger are noise. Lenders use it for the same reason when they set leverage covenants — they are sizing debt against the cash the operations throw off before that same debt is served, and every borrower is measured the same way.

The discipline is to never let it travel alone. Put three numbers beside it and most of the distortion disappears: depreciation as a percentage of revenue, which tells you how asset-heavy the business is; capital expenditure as a percentage of depreciation, which tells you whether the asset base is being maintained, harvested or expanded; and operating cash flow, which tells you whether the profit turned into money. In the example above, depreciation is 12.0% of revenue and capex equals depreciation exactly — a business standing still. Read those alongside a 26.0% EBITDA margin and you have the whole picture. Read the 26.0% alone and you have a headline.

Frequently asked questions

Is EBITDA the same thing as operating cash flow?

No, and treating it as a proxy is the most common way it misleads. Operating cash flow also reflects the change in working capital, the interest actually paid and the tax actually paid. A company growing fast can post a rising EBITDA while cash goes backwards, because every extra sale ties up money in inventory and receivables before it collects. In the example above, EBITDA is $13 million while the cash left after replacement spending, interest and tax is $3.75 million — and that is with working capital held flat. If you want the cash question answered, read the cash flow statement; EBITDA is a profit measure with two costs removed, not a cash measure.

Why do lenders rely on EBITDA if it overstates profit?

Because for their specific question it is the right measure, not a flattering one. A lender wants to know how much cash the operations produce before that cash is used to pay them, so subtracting interest first would be circular. Tax is deducted after interest, so it moves with the loan too. What lenders do not do is stop there: leverage covenants are paired with interest-cover tests, maintenance capex is scrutinised in diligence, and the definition of EBITDA in a credit agreement often runs to a page of permitted and forbidden add-backs precisely because the borrower has an incentive to inflate it. The lender's use is disciplined by a contract; a press release is not.

What is "adjusted EBITDA" and should I trust it?

Adjusted EBITDA is EBITDA with further items added back at the preparer's discretion — restructuring costs, share-based payment, legal settlements, integration expenses, sometimes rent. Because neither IFRS nor US GAAP defines it, there is no standard list, and the same company can change its own definition between years. That does not make it worthless: excluding a genuine one-off, such as a single factory fire, improves comparability. It becomes a warning sign when the same category of "exceptional" cost appears every year, because a restructuring charge in each of five consecutive years is an operating expense with a euphemism attached. Read the reconciliation table, count how many years each add-back has appeared, and treat anything recurring as a cost.

Is EBITDA allowed in published accounts?

It is not a line item in any income statement prepared under IFRS or US GAAP, because neither framework defines it. It is permitted as a supplementary measure, subject to rules that exist precisely because it can mislead. In the United States, Regulation G and Item 10(e) of Regulation S-K require a non-GAAP measure to be reconciled to the most directly comparable GAAP figure and forbid giving it greater prominence. In Europe, the ESMA guidelines on alternative performance measures require a definition, a reconciliation and consistency between periods. If a company presents EBITDA without a reconciliation, that is itself the finding.

How do I sanity-check an EBITDA margin in thirty seconds?

Divide depreciation by revenue and see how much of the margin is being added back. In the capital-intensive example that ratio is 6 ÷ 50 = 12.0%, so nearly half of the 26.0% EBITDA margin is depreciation; in the asset-light one it is 1 ÷ 50 = 2.0% and the two margins nearly coincide. Then divide capital expenditure by depreciation: a ratio near 1.0 means the asset base is being held level, well below 1.0 means the company is running its assets down and will pay for it later, and well above 1.0 means it is expanding. Those two ratios take seconds and tell you whether the EBITDA margin is close to the truth or a long way from it.