Adjusted EBITDA: What Was Adjusted, and Who Chose It
Adjusted EBITDA is earnings before interest, tax, depreciation and amortisation (EBITDA) after management has removed items it says do not belong. No accounting standard defines it, so the company chooses both the starting figure and the removals. Sarvani Coatings Limited, an invented paint maker, reports Rs 446 crore, presents Rs 452 crore, and a two way adjustment gives Rs 448 crore. All three are correct arithmetic on different rules.
Two things sit underneath that answer, and both are covered separately. The first is the separation of a reported result into the part that should be expected again and the part that should not, together with the habit of running any adjustment in both directions instead of one. The second is the test that decides whether a label is honest: an item called one off earns that name by not coming back, and counting the years it appeared is how that is checked. The new element is the measure itself. Adjusted EBITDA has no definition anywhere, so the starting figure is a choice too, and the same person makes both choices at the same moment.
What is adjusted EBITDA, and how many decisions are hidden inside the phrase?
The phrase splits into two words that do two different jobs. The second word names a figure. The first word says that somebody has taken things out of it. Almost every argument about adjusted EBITDA is an argument about the first word. Which items came out, and did they deserve to? Very few arguments are about the second word. The odd part is that the second word names no line anybody can point to on a published statement.
A household makes the point. Somebody states that their normal monthly spending is Rs 42,000/-, after taking out the wedding they attended and the tyre they had to replace. Before the tyre can be argued about as genuinely unusual, there has to be agreement on what was in the Rs 42,000/- to begin with. Did it include the rent? The school fee paid once a term? The money sent home? The removals are the visible argument. The starting figure is the silent one, and it decides more.
EBITDA works the same way. EBITDA is assembled from published lines rather than read off one, and different assemblies of the same year give slightly different answers. So a company that presents an adjusted figure has already made a decision before it names a single add back. A reader who examines only the list of removals has accepted the starting figure without examining it, and the starting figure was chosen by the same people who chose the removals.
The good news is that both decisions leave a trail. The starting figure can be rebuilt from the published statement. The removals sit in the note to the accountsThe detailed disclosures printed after the main financial statements, where an item that is bundled into a single published line gets broken out and described., where an item that has been bundled into one published line gets described separately. Both are recoverable. Neither is recovered by reading the headline.
Which accounting standard defines adjusted EBITDA?
None. No standard defines the measure, and a great deal follows from that, so the answer earns its own section rather than a passing clause.
The statements that Sarvani Coatings Limited publishes are prepared under a named framework. In India that framework is Ind ASThe Indian Accounting Standards, the notified set of rules under which a listed Indian company prepares and presents its financial statements. The route to the notified standards themselves is in the block below., and the rules for recognising, measuring and disclosing each item in the reported statement live there. Revenue, employee cost, other expenses, depreciation: every one of those has a definition somebody can look up and an auditor can test against. Those definitions are what make the reported statement a shared object. Two readers can disagree about what it means and still agree about what it says.
EBITDA is not in that set. Neither is adjusted EBITDA, nor operating EBITDA, nor cash EBITDA, nor any of the other phrases that appear in results presentations. Each of those phrases is assembled outside the statement, by the company, for its own commentary. Because nothing defines the measure, two companies can publish the identical phrase meaning two different calculations, and neither of them has done anything irregular.
The last clause is the part readers skip. The absence of a definition is not a loophole somebody is exploiting. The absence is a plain feature of a measure that was never standardised in the first place. The consequence is narrower and more useful. The phrase carries no information on its own, so two companies cannot be compared on it until both reconciliations have been opened and the same things confirmed to have been done.
Where to look for the requirement in force
The definition and measurement of every line in the reported statement sit with the Institute of Chartered Accountants of India at icai.org and with the standards notified by the Ministry of Corporate Affairs at mca.gov.in. How a company may present a measure that no standard defines, and what it must show alongside it, is a disclosure question that belongs with the Securities and Exchange Board of India at sebi.gov.in, and the filed results themselves are published through the exchanges at nseindia.com and bseindia.com.
Two listed companies both publish a line called adjusted EBITDA for the same year. Are they publishing the same measure?
Who chooses the adjustments, and at what point in time?
The issuer chooses. Most readers already assume that much. The part worth slowing down for is when.
A financial year opens, the business runs for twelve months, the year closes, the numbers are assembled, and the result becomes known inside the company. Only after all of that does anybody sit down and decide which items will be described as not belonging. The adjustment policy is written when the outcome is already on the table.
The order of events is not an accusation but a structural fact, and a note can state it out loud without implying anything about anybody. A rule written after the result is known behaves differently from a rule written before it, and saying so alleges nothing at all.
Compare it with something ordinary. A cricket team that agrees before the toss that rain stopping play means a rematch has made a rule. A team that decides, after losing, that the rain should count, has made a different kind of rule, even if the words are identical and even if everyone involved is completely sincere. The difference is not honesty but sequence.
When is the adjustment policy for a financial year chosen, relative to the result for that year?
What does a one sided adjustment look like when nobody intended one?
Here is the mechanism, and it needs no villain.
Two kinds of item can sit inside a single published expense line. One is a charge, something that made the year look worse. The other is a benefit, something that made it look better. A reader who sees costs rise wants to know why, and the person writing the commentary would rather explain a charge than be asked about it. So a charge attracts explanation. A benefit attracts no such pressure. Nobody writes in to ask why the number was good.
The charge gets named and quantified in the commentary, and a named figure is easy to add back. The benefit gets absorbed into a caption, and a figure nobody has quantified is easy to leave alone. The two kinds of item attract completely different amounts of explanation for perfectly ordinary reasons, so no intent is required for the asymmetry to appear, and the result is a reconciliation in which every row happens to point the same way.
The household version of this is familiar. A tea stall owner says a normal month brings in Rs 60,000/-, then adds that last month was worse than normal because the cart needed repairing, so really it is Rs 64,000/-. He is not lying. But the same month also carried a local festival that doubled a week of sales. Nobody ever asks a stall owner to explain a good week, so he does not mention it. The corrected figure is one sided, and he arrived at it honestly.
For Sarvani Coatings Limited in year three, the two items are these. A restructuring chargeA cost recognised when a business reorganises something about how it operates, such as closing or combining a site or a line. How such a cost is recognised and measured belongs to the accounting material routed above. of Rs 6 crore sits inside published other expenses of Rs 460 crore, or 1.30 per cent of that line. A provision write backA reversal of an amount a company had previously set aside for an expected cost. Because the amount is released, it reduces the expense line in the year of the reversal rather than adding to income. of Rs 4 crore also sits inside the same line, and it went the other way: it reduced other expenses and so lifted the same EBITDA line the charge had pushed down. Without it, other expenses for the year would have read Rs 464 crore.
Management adds back the charge and does not remove the write back. Both items are disclosed. Neither is on the face of the statement. The reconciliationThe table a company prints to show how it got from a figure in the published statement to the adjusted figure it prefers to talk about, one named item per row. that results is arithmetically perfect and structurally lopsided, and the shape of it is visible before a single number is checked.
How is adjusted EBITDA recomputed both ways, step by step?
Three steps, and they work on any issuer that prints a reconciliation at all.
Step one is the reported figure exactly as published. Not the company's adjusted figure, and not a figure rebuilt out of preference. The published one. Anybody following the working then starts from the same place.
Step two adds back every disclosed item that reduced it. Charges, one off costs, provisions taken, write downs. The company has already done this step and shown its working on it, so everybody performs it.
Step three removes every disclosed item that increased it. Write backs, one off gains, releases, recoveries. The reconciliation provided will not contain step three, so step three is usually the analyst's alone. The third step is the whole method, and a reader who performs only the second has recomputed nothing and has simply reproduced the presented figure by hand.
The limits of the method matter as much. The method passes no judgement on whether the charge was really unusual, a different test covered separately. The method holds no opinion on whether management should have removed the write back. The method is a mechanical instruction: apply the same treatment to items running in both directions, and whatever falls out is the answer.
Every disclosed charge in the reconciliation is added back, and the work stops there. Which figure has been produced?
Management adds back a Rs 6 crore charge for year three. If a Rs 4 crore benefit sits inside the same published line, what happens to the presented figure?
What are the three figures for Sarvani Coatings Limited in year three?
Every figure below belongs to the year ended 31 March, year three. Revenue for that year is Rs 2,415 crore and reported EBITDA is Rs 446 crore, a margin of 18.47 per cent on that revenue.
Adding back the Rs 6 crore restructuring charge gives the Rs 452 crore management presents, a margin of 18.72 per cent on the same revenue. Running the third step as well removes the Rs 4 crore write back and leaves Rs 448 crore, a margin of 18.55 per cent. All three figures are correct arithmetic on the same year, and they differ only in which rule was applied.
| Year three, ended 31 March | Figure | Margin on Rs 2,415 crore |
|---|---|---|
| Reported EBITDA, as published | Rs 446 crore | 18.47 per cent |
| Add back the restructuring charge disclosed in the notes | plus Rs 6 crore | |
| Adjusted EBITDA, as management presents it | Rs 452 crore | 18.72 per cent |
| Remove the provision write back disclosed in the same notes | less Rs 4 crore | |
| Adjusted EBITDA, run both ways | Rs 448 crore | 18.55 per cent |
The presented figure therefore sits Rs 4 crore above the one reached by running the adjustment both ways. The gap is 0.9 per cent of reported EBITDA for the year, and 0.17 of a margin pointOne percentage point of margin. A move from 18.55 per cent to 18.72 per cent is 0.17 of a point, and a point here is worth Rs 24.15 crore of EBITDA on the year three revenue base..
Watch the bar that does not move.
Reported EBITDA for year three stays at Rs 446 crore, revenue stays at Rs 2,415 crore, and the restructuring charge added back stays at Rs 6 crore. The control moves one thing only: how large the favourable item is that management does not remove, from Rs 0 crore to Rs 12 crore. The presented figure does not depend on that item at all, so it stays fixed at Rs 452 crore and 18.72 per cent at every setting. The figure run both ways is Rs 452 crore less whatever the control says. Start at the disclosed Rs 4 crore, then push it and watch which bar responds.
With Rs 4 crore of favourable item left in, the presented figure reads Rs 452 crore at 18.72 per cent while the figure run both ways reads Rs 448 crore at 18.55 per cent, a gap of Rs 4 crore and 0.17 of a margin point, and enterprise value of Rs 11,592 crore over that two way figure is 25.9 times.
What is the difference worth here, and why is the size not the finding?
Set the Rs 4 crore beside a year three reported EBITDA of Rs 446 crore and it comes to 0.9 per cent. As a margin the gap is 0.17 of a point. On this revenue base of Rs 2,415 crore that is the difference between 18.55 and 18.72 per cent. Nobody would build a case on that alone, and pretending otherwise would be dishonest.
But read the two facts side by side rather than one after the other. The amount is small and the method is not. Applied to a year carrying a large favourable item, the same method would produce a large error, and there is no way of knowing in advance which kind of year is on the table.
The control above is built for exactly that reason. Rs 12 crore is not an outlandish size for a provision release at a company of this scale. Move the control there and the figure run both ways falls to Rs 440 crore at 18.22 per cent. The presented figure has not shifted a paisa. The gap becomes half a margin point. The rule that produced it never changed. Only the size of the item it happens to ignore did.
So the honest sentence is not that Sarvani Coatings Limited has overstated anything by a meaningful amount in year three. The honest sentence is that the presented figure is built on a rule insensitive to favourable items, and that the cost of the rule in any given year is whatever the favourable items happen to add up to.
The gap is Rs 4 crore on reported EBITDA of Rs 446 crore for year three. Is that worth writing about?
Enterprise value is held at Rs 11,592 crore. How much does the choice between the three EBITDA figures move the multiple?
What does the choice of figure do to a multiple built on it?
A multiple is a division, and a division has two sides. Change the bottom and the answer changes, whether or not anything happened to the business.
For Sarvani Coatings Limited the top of that fraction is fixed here. There are 24.00 crore shares in issue and the illustrative price is Rs 486/- as at 28 August 2026, giving a market capitalisation of Rs 11,664 crore. The company is in net cash: cash and investments of Rs 312 crore against borrowings of Rs 240 crore, the company holds net cash of Rs 72 crore and its net debtBorrowings less cash and liquid investments. When the cash is larger than the borrowings the figure is negative, and the company is described as being in net cash. is therefore minus Rs 72 crore. Net cash makes enterprise valueThe market value of the equity plus net debt, meant to represent what the whole operating business is being valued at rather than just the shares. Building it up is covered in the valuation material. Rs 11,592 crore, below the market capitalisation rather than above it.
Now divide that one fixed number by each of the three EBITDA figures for year three and see what happens.
| Denominator used | EBITDA, year three | Enterprise value to EBITDA |
|---|---|---|
| Reported, as published | Rs 446 crore | 26.0 times |
| Adjusted, as management presents it | Rs 452 crore | 25.6 times |
| Adjusted, run both ways | Rs 448 crore | 25.9 times |
| The whole spread, on an unchanged enterprise value of Rs 11,592 crore | Rs 6 crore | 0.35 of a turn |
About a third of a turn, or 1.33 per cent of the reported multiple, produced entirely by the choice of denominatorThe number on the bottom of a fraction. In a multiple it is the earnings figure being divided into, and changing it changes the answer even when the number on top has not moved. and by nothing that happened inside the paint business. Put it another way. The spread is exactly what a real Rs 6 crore swing in EBITDA would have done, and Rs 6 crore is 1.35 per cent of the year three figure. The rule moved the multiple as far as a genuine change in trading would have.
A multiple must always name the figure it was divided by, and a multiple quoted without one cannot be compared with anything. Two vegetable sellers can both quote a price per kilo, and if one of them weighs with the packet on the scale and the other without, the two prices are not comparable however carefully they are read. The unit has to be stated before the number means anything.
A comparison table lists five companies with an enterprise value to EBITDA multiple against each, and names no denominator anywhere. Which fault does the table carry?
Who actually does this, and what do they do with it
Covenant headroom is usually expressed against an earnings measure, and a borrower who defines that measure generously has quietly bought itself room the lender did not intend to give. So a lender sizing a facility runs the two way recomputation before anything else. The lender writes its own definition into the agreement for exactly this reason.
An analyst building a comparison rebuilds every row on the reported figure first, then adds a separate column for each company showing what the two way adjustment did to it. The second column is often more interesting than the multiples. A company whose adjustments run one way every year is saying something about how it presents itself, and that is a question to carry forward rather than a conclusion to reach.
A long term holder of the shares mostly wants to know whether the base they are compounding on is the reported one or a decorated one. The Rs 4 crore in year three barely registers. The habit, repeated across five years, changes the base they think they are starting from, and that is what they are watching for.
The failure: a multiple written into a table with its rule left behind
An analyst takes adjusted EBITDA of Rs 452 crore straight from the release, divides the enterprise value of Rs 11,592 crore by it to get 25.6 times, and drops that figure into a comparison table beside multiples that other people built on reported EBITDA. Nothing in the table records which figure each row was divided by. The analyst has done no arithmetic wrong.
The error compounds sideways, so the damage is worse than a single wrong cell. The row now sits 0.35 of a turn below where it would sit on the reported figure, and a reader takes that gap as a statement about the business rather than about an adjustment policy. Every ranking built from the table inherits it. Every average inherits it. And the mismatch is largest exactly where the adjustments are largest. A reader needs the comparison to hold most at exactly that point.
The fix is a label and a rebuild: every multiple carries the figure it was divided by in its own row, and every row gets rebuilt on the same rule before any of them is compared with another. The rebuild costs an hour, and it is the difference between a table and a mess.
What is asked at the end of this, and what is refused as a conclusion?
Three figures now stand for the year ended 31 March, year three, along with a recomputation that can be shown to anybody. The list of what gets written down next is short, and the list of what gets left out is longer.
The question is specific. Why does the adjustment run in one direction. Not whether management is honest, not whether the presentation is aggressive, not what it says about the culture of the place. Just the direction, asked plainly, of a company that publishes a reconciliation with two disclosed items in the same line and removes one of them. The question has answers, and the record is where to look for them: the same reconciliation in earlier years, whether favourable items have ever been removed, and what the company says when somebody asks.
The work ends with the recomputed figure of Rs 448 crore and that question. A verdict on Sarvani Coatings Limited or on its presentation is not something this method produces.
The restraint is not politeness. A one sided adjustment is consistent with several explanations, including an entirely unremarkable one in which nobody thought about it at all, and the published statements do not separate them. Writing an unsupportable verdict is how a reputation for careful work gets spent. Writing the number and the question is how it gets built.
The figure has been recomputed, Rs 448 crore reached, and the adjustment found to run one way. Which sentence gets written?
Where do the rules behind these figures come from?
How a restructuring charge or a provision write back is recognised and measured belongs to the accounting material and is covered there. How a company may present a measure that no standard defines rests with the securities regulator and the two exchanges. Every rupee above was built so that a reader can recompute it from the working shown.
The routes, and what each one settles
| Body | What is looked up there | Site |
|---|---|---|
| Institute of Chartered Accountants of India | How an item inside the reported starting figure is recognised, measured and disclosed under Ind AS | icai.org |
| Ministry of Corporate Affairs | The notified accounting standards and the Companies Act requirements the reported statements are prepared against | mca.gov.in |
| Securities and Exchange Board of India | What a listed issuer must show alongside a measure that no accounting standard defines | sebi.gov.in |
| National Stock Exchange of India | Where a filed set of results and the release that accompanies it is published | nseindia.com |
| BSE Limited, formerly the Bombay Stock Exchange | The second filing route for the same results and the same release | bseindia.com |
Sarvani Coatings Limited, Nandivarman Paints Limited, Kesaria Surface Solutions Limited, Thottam Chemicals Limited, Ravindra Setlur and Meghna Iyer are invented.
Educational material. Not advice on any investment, tax, budget or market position.
