Reported vs Adjusted Earnings: Who Defines Each One

Reported earnings follow an accounting framework and are audited against it. Adjusted earnings follow rules the company writes for itself and are not. The difference is not accuracy: it is who chose the definition and whether the same definition will be used next year. A reader compares reported with reported, adjusted with adjusted, and never one with the other.

Underneath both definitions sits a single question: who held the pen. One of these two numbers was produced under rules written outside the company by a body the company does not control, and checked by somebody who reports to the shareholders. The other was produced under rules the company wrote for its own purposes, and nobody outside the company approved them. Both can be useful. Only one of them was defined by a party with no interest in how it came out.

What exactly is a reported figure?

A reported figure is any line that appears in the financial statements themselves: revenue, earnings before interest, tax, depreciation and amortisation (EBITDA) as the ladder builds it, profit before tax, profit after tax, earnings per share. Its value is not a matter of preference. A reported figure is what the recognition and measurement rules of an accounting frameworkThe published body of rules a company must apply when it decides when an amount is recognised, at what value it is measured, and where it is shown. Set outside the company and applied by everyone reporting under it. produce when they are applied to the year that happened.

Two companies applying the same rules produce numbers that can honestly be set beside each other, and that is what makes a reported figure comparable at all. That is the whole reason a framework exists. Nobody wrote Ind AS to make accounting elegant. Ind AS was written so that a number produced in one company's ledger means roughly the same thing as the number with the same name in another company's ledger. A reader can then line up ten issuers and learn something from the comparison rather than just from the noise.

Consider how two street food carts could be compared if each of them counted its own takings on its own basis. One counts the money in the tin at the close of the day. The other counts every order placed, including the four plates on credit that may never be paid for. Both are sincere. The two numbers are not measuring the same thing, so neither can be compared with the other. A framework is society's answer to that problem, imposed from outside so that nobody has to trust anybody's private definition.

What is an adjusted figure, and who wrote its rules?

An adjusted figure starts from a reported one and takes amounts out of it. The company decides which amounts, on the argument that removing them describes the ongoing business better than the reported line does. For year three, the statements of Sarvani Coatings Limited, an invented coatings maker, carry an EBITDA line of Rs 446 crore. Alongside it the company puts forward an adjusted EBITDA of Rs 452 crore, arrived at by adding a Rs 6 crore restructuring charge back to that line. The step from one to the other is written down. The rule that decided a restructuring charge qualifies and something else does not is written down nowhere outside the company.

Households already do this. Asked what it spends in a normal month, a household will name a number that quietly leaves out the school fee paid once a year, the wedding gift given last April and the day the fridge died. The household's number is genuinely more useful for planning next month than the raw bank statement is. The household also defined that number, using a rule it made up on the spot, and it will define it slightly differently the next time the question is put.

An adjusted figure can be the more useful input for a forecast, and it is still a figure whose definition has to be read before it is used. If a plant was shut once and will not be shut again, a forecast built on the reported line carries a cost that will not recur. Removing it is the correct modelling choice. The point is only that the removal happened under somebody's rule, and that the analyst is entitled to read the rule before inheriting its answer.

Which of the two did the auditor actually examine?

Most readers place the audit boundary in the wrong spot, and it is worth drawing plainly. A statutory auditThe independent examination a company's accounts must undergo before they are published, carried out under a mandate set by law rather than by the company. The scope of an audit and the way it is carried out are taught in the accounting layer of this library. addresses the financial statements prepared under the framework. A measure the company invented for its own commentary is not one of those statements. The measure was not prepared under the framework. Since the rule behind it is the company's own, no external rule exists for an auditor to test it against.

Sitting outside the audit does not make an adjusted figure lawless. Reputable issuers reconcile it, meaning they show the arithmetic that runs from the audited line to their own. The weight in that sentence sits on two facts. The audited object is one end of the bridge. The company's measure is the other end. Neither of those two facts says anything by itself.

Which is exactly why the reconciliation between them is the most important object in the whole release. It is the only place where the number nobody outside the company defined is tied, item by item, to the number that was defined outside it. A reconciliation tableA short table that shows the steps from one published figure to another, naming each amount added or removed, so a reader can walk from the first number to the second without guessing. that lists three amounts by name and value has stated the definition. A sentence saying profit was adjusted for exceptional items has stated nothing at all.

So the two figures differ on more than one field at once, and it helps to see all of them together rather than arguing about which number is better. Four fields separate them, and accuracy is not one of the four.

Comparable with what, exactly?

The word comparability covers two completely different tests, and readers slide between them without noticing. Most confusion on this subject lives there. Ask it precisely and the two separate cleanly.

The first test runs down one company across years. Is Sarvani Coatings' adjusted figure for year three comparable with its adjusted figure for year two? Only if the definition held. If last year the company removed restructuring and a legal settlement, and this year it removes restructuring alone, the two numbers were built by two different rules and the change between them contains a change of method as well as a change of business.

The second test runs across companies in one year. Is Sarvani Coatings' adjusted EBITDA comparable with the adjusted EBITDA of Nandivarman Paints Limited? Only if both companies adjust for the same categories of item. Each wrote its own rule for its own reasons, so they rarely do. Two adjusted figures with the same name can be built on rules that overlap barely at all.

Reported figures win on comparability precisely because nobody at either company chose their rules. That is the trade. The tailoring that makes an adjusted figure a better forecasting input is given up, and in exchange comes a number whose definition nobody had to negotiate. When the two tests pull in different directions, the reported basis is the one that survives contact with a second company.

Where does each figure actually sit in the document?

Open any results filing and the two figures are physically in different places. The adjusted numbers are near the top: in the headline of the earnings releaseThe short document a listed company puts out alongside its filed results, carrying a headline, management commentary and usually a slide deck. The release is lodged with the exchanges., in the first paragraph of management commentary, on slide three of the deck. The reported numbers are further in: in the statements themselves, and the amounts that explain them are deeper still, in a note to the accountsThe supporting disclosure standing behind a line in the statements, setting out what it is made of and the judgements inside it. The framework fixes what has to appear in a note, and that is taught elsewhere in this library. that most readers never open.

Prominence and definitional authority run in opposite directions inside a release, so the higher a figure sits, the more likely it is that the company chose the rules that produced it. This is not a conspiracy and there is nothing to allege about anyone. A company puts forward the measure it thinks describes the year best, and that measure is usually the one it built. The reader's job is simply to know that height in the release is not evidence of anything, and to travel downwards.

Sarvani Coatings' own year three is a clean illustration of the geography. The Rs 6 crore restructuring charge and the Rs 9 crore insurance claim are both described in the notes, not on the face of the statements. So is the Rs 4 crore provision write backMoney set aside in an earlier year and no longer needed, taken back into the accounts. The expense line falls in the year that happens. The mechanism sits in the accounting layer of this library.. Every one of those amounts is disclosed. The disclosure just sits at the bottom of the document rather than the top.

What happens when the definition moves between years?

Here is the failure mode that costs readers the most, and it is quiet enough that it usually goes unnoticed for years. A company adjusts each year for the items that year's commentary describes as unusual. Adjusting for what a year's commentary calls unusual is a perfectly reasonable thing to do. But the year is different every year, so the items are different every year, and so is the rule. The result is a series of adjusted numbers that looks like one series and is actually several, laid end to end.

The only defence a reader has is to read this year's reconciliation beside last year's and check that the same categories of item are being removed. Not the same amounts, which will obviously differ. The same categories. If restructuring came out both years, that is a rule holding. If restructuring came out last year and a legal settlement came out this year and neither year removed the other, that is two rules wearing one label.

The household again. The cost of a normal month draws one answer in March, when the school fee is fresh in mind and gets excluded, and a different answer in September, when the fee is forgotten and the wedding gift is the unusual item instead. Nobody lied in either month. The trouble is only that there are now two numbers built by two rules, and putting them on a chart produces a trend that is partly a picture of the household's own memory.

How is a company that adjusts compared with one that does not?

The mismatch comes up constantly and there is only one honest answer. Sarvani Coatings publishes an adjusted EBITDA. Suppose Kesaria Surface Solutions Limited publishes nothing but the statements. One company then has two figures and the other has one, and a comparison is still to be made.

The reported basis is the only one both of them publish, so rebuild both on it. It costs the tailoring, and the note should say so. The gain is a comparison where both sides were measured the same way by somebody outside both companies.

One alternative gets proposed often and is worse than it looks. Inventing adjustments for the company that does not make any would put both sides on an adjusted footing. The requirement behind that is worth spelling out. The analyst would be deciding which of Kesaria's costs are unusual, using a rule Kesaria never wrote, applied from outside with less information than its own finance team holds. Inventing them is not levelling the field. The analyst would be writing another company's accounting policy for it, and then comparing that company against an outsider's draft of itself.

What if last year carries no adjusted figure at all?

The missing prior year is the ordinary case and gets almost no attention. A company starts presenting an adjusted measure this year. Only the reported figure was published then, so the prior year comparative in the same document is the reported one. A growth rate is wanted, and what is available is one adjusted number and one reported number.

The honest answer is that no adjusted growth rate can be computed at all, and the correct output is a reported growth rate plus a written note of what was adjusted in the current year. Anything else pairs two different rules across a time gap and calls the difference growth. The note states it plainly: reported against reported gives this, the company also presented a figure of that, and no like for like adjusted comparison exists on this record. A reader who sees that sentence trusts everything around it a good deal more.

What do three different rules do to one year?

The three EBITDA figures and their reconciliation are built under the EBITDA bridge, covered separately. Setting them beside each other raises a different question: who defined each of these, and what may each be compared with.

Start at the EBITDA line for year three, where revenue is Rs 2,415 crore. Under the framework the line reads Rs 446 crore, or 18.47 per cent. Under the company's own rule it reads Rs 452 crore, or 18.72 per cent, the charge having gone back in. A test run in both directions at once takes the charge out, and because the Rs 4 crore write back pushed the line the opposite way, it takes the write back out too. On that basis the line reads Rs 448 crore, or 18.55 per cent. Three numbers, one year, three rules.

Carried down to the profit line, the same three rules start to bite. Reported profit before tax is Rs 371 crore. Taking out the Rs 9 crore insurance claim sitting inside other income and putting back the Rs 6 crore charge removes exactly the two amounts the notes call a non-recurring itemAn amount a company does not expect to see again in the ordinary course of its business. Whether a given item genuinely qualifies, and how often the label gets used, is set out under non-recurring items.: Rs 368 crore. Now instead carry down the both directions test from the line above, so the Rs 4 crore write back comes out as well: Rs 364 crore. Both are internally consistent. The two figures are consistent with different rules, and the Rs 368 crore figure is not the symmetric one.

The per share column is worked on the 24.00 crore shares in issue and the published effective tax rateTax charged for the year expressed as a share of profit before tax. The result is what a company actually paid on that profit, rather than the headline rate in the statute. of 25.1 per cent. Taking tax at that rate turns Rs 368 crore into roughly Rs 275.6 crore, or roughly Rs 11.48/- for each share, and turns Rs 364 crore into roughly Rs 272.6 crore, or roughly Rs 11.36/-. Against the reported Rs 11.58/-, the underlying figure is lower by about Rs 0.10/-, roughly 0.9 per cent, and the two way figure by about Rs 0.22/-, roughly 1.9 per cent. The spread is small in a quiet year, and that is exactly why the rule behind each figure has to travel with it: nothing about the number itself reveals which of the three is in hand.

What happens to a growth rate when the two bases are mixed?

Everything above is definitional until a change is computed. In year two the statements carry an EBITDA line of Rs 340 crore against revenue of Rs 2,120 crore, a margin of 16.04 per cent. Nothing is adjusted anywhere near it. Setting the year three figures against that shows what happens.

The framework figure over the framework figure, Rs 446 crore on Rs 340 crore, gives a rate of 31.18 per cent. The company's own figure over that same framework figure, Rs 452 crore on Rs 340 crore, gives 32.94 per cent. The second one looks like a slightly better year. Nothing was ever adjusted in the year underneath, so the higher rate is neither a better year nor an adjusted growth rate. The rate is a mixed basis one: one rule on top, another rule below, and the distance between the two rules entered in the accounts of the analysis as growth.

The gap is 1.76 of a percentage pointThe plain gap between two percentages. Going 31 to 33 is a gap of two points. Saying it rose by two per cent instead would mean something quite different and much smaller., and every bit of it was manufactured by a change of definition rather than by anything that happened in the business. Written to one decimal the two rates read 31.2 and 32.9, and subtracting those gives 1.7, so the two decimal figures are the ones to subtract: the gap is small enough that the rounding moves it.

Every pair is worked in full above: Rs 446 crore beside Rs 452 crore beside Rs 448 crore, Rs 371 crore beside Rs 368 crore beside Rs 364 crore, three per share figures and two growth rates. The one quantity a reader might want to vary is the size of an item left sitting inside the line, and the EBITDA bridge, covered separately, moves exactly that. The distinction at work here is definitional rather than numeric, and the honest finding at the end of it, that no adjusted growth rate exists on this record at all, is a statement rather than a setting.

The reported growth rate of 31.2 per cent is what gets published, and beside it a sentence naming what was adjusted this year and stating that the prior year carries no adjustment. The sentence is not a hedge but the finding. A reader who wants the adjusted series has to be told that the record cannot produce one. The alternative is that they assume it was computed and quietly trust it.

Who actually has to make this call, and when?

The covenant that will be tested in eighteen months is written against defined lines, and tested by somebody who was not in the room when the adjustment was argued, so a lender sizing a facility works almost entirely on the reported basis. If a borrower's presentation leads with an adjusted figure, the credit note carries the reported one and mentions the adjusted one as commentary. Carrying the reported one is not scepticism about the borrower. The covenant will be tested against a definition, and only one of the two figures has one that outlives the meeting.

A plant closure that will not recur genuinely should not sit in next year's model, so an equity analyst goes the other way and starts a forecast from the adjusted basis. But the disciplined ones write their own adjustment schedule rather than inheriting the company's, apply it to every year in the model, and publish the reconciliation from the reported line so a reader can see exactly what was taken out. Meghna Iyer, working through Sarvani Coatings' year three, would carry both Rs 368 crore and Rs 364 crore in her file and name the test behind each rather than choosing one and going quiet.

A household investor reading a results release has the simplest version of the same job and the least time for it. The work is to find the reconciliation, count the items, and check whether the amounts removed go in one direction or both. The check takes a minute and catches most of what is worth catching. If everything removed happens to make the year look better, that is not wrongdoing found, it is a question worth carrying into the next set of accounts.

And when the company voice is on the call, as Ravindra Setlur would be for Sarvani Coatings, the useful question is never whether the adjusted number is right. The question worth asking is which items the rule covers and whether the same rule will be applied next year. An answer that names the categories is worth more than an answer that defends the figure.

Sarvani Coatings Limited, Kesaria Surface Solutions Limited, Thottam Chemicals Limited, Nandivarman Paints Limited, Meghna Iyer and Ravindra Setlur are invented.
Educational material. Not advice on any investment, tax, budget or market position.