Earnings Quality: How to Tell Durable Profit From Reported Profit

Earnings quality asks one question: is this year's profit a sound base for expecting next year's? Profit scores high when cash has already confirmed it, when the trading that produced it repeats, and when few estimates sit inside it. Quality is low when the profit arrived once, or rests heavily on judgement. The word is not a verdict on honesty: an entirely honest business with lumpy trading reports low quality earnings.

The subject starts on a street rather than in a statement. Two shops a hundred metres apart both closed last year with a profit of about Rs 6,00,000. The first sells exercise books to the same four schools every June, and has done for eleven years. The second sells the same goods but made most of its money by letting out its empty godown for one wedding season. The owner needs the space back, so that season is not happening again. Both figures are correct. Both were counted honestly. Only one of them says anything at all about next year, and no amount of care in the counting changes which one.

The whole subject sits in that one street. The three statements, the ratios and their honest limits, how revenue is recognised and where the cost side of a business sits are all already in hand. The new habit is asking a second question about a number already accepted as correct, and writing down what that question returns without turning it into a verdict about anybody.

What does earnings quality actually mean, and why is it not a word about honesty?

Quality here carries a narrow and slightly unfortunate meaning: durable, and well evidenced. Durable means the trading that produced the profit is the kind of trading that happens again. Well evidenced means somebody outside the business has already confirmed a large part of it, most often by paying. Neither half of that is a statement about whether the accounts are truthful.

Earnings quality measures how much weight a forecast can safely rest on this year's number, not whether this year's number is true. Those are separate questions answered by separate work. Whether the number is true is what an audit is for. Whether the number will repeat is the question here. A completely clean set of accounts can still be asked it. The wedding godown was reported perfectly, and it still says nothing about next June.

Left alone, the misreading survives and does real damage, so kill it now. A business with genuinely lumpy trading, a boat builder taking two orders a year, a contractor finishing one large project, a stationer selling most of its year in ten weeks around a school session, reports low quality earnings while doing absolutely nothing wrong. Its profit is hard to forecast because its trading is hard to forecast. Lumpy trading is a description of the business, not an allegation about its accounts. The reverse trap is just as real. None of these tests is designed to find a misstatement, so a business can pass every one of them and still be misstated.

Treating the phrase as an accusation is the commonest misuse of this whole subject, and the cost of that misuse falls on ordinary businesses that happen to trade in an uneven shape. A supplier who reads a seasonal stationer's uneven profit as evidence of something being hidden will tighten terms on a business that has done nothing but sell notebooks in June. The word describes a shape, not a character.

What are the three durability tests, and what does each one ask?

Three questions do most of the work, and each one is answered from a different place in the accounts. The first is quickest and the third is the one people forget, so ask them in that order.

The first test asks whether the profit is backed by cash. Profit is a measurement; cash is a fact somebody else confirmed by paying. Cash conversionOperating cash flow divided by profit after tax, for the same period. A figure above one means the business collected more cash from trading than it reported as profit. puts operating cash flow over profit after tax and asks how much of the reported profit has already turned into money in the bank. Anjani Stationers Private Limited, an invented stationer, reported profit after tax of Rs 30,00,000 against operating cash flow of Rs 36,30,000. The ratio is 1.21 times. More cash arrived than profit was reported, and cash ahead of profit is the reassuring direction.

The second test asks whether the trading repeats. Look at what actually produced the profit and ask whether that thing happens again. Anjani Stationers sold Rs 2,70,00,000 of notebooks and exercise books against Rs 2,40,00,000 the year before, up 12.5 per cent, at a gross margin of 45.0 per cent in both years. Selling notebooks to schools is exactly the kind of trading that repeats, and the gross margin holding steady across two years says the core of the business behaved the same way twice. One part of year two is genuinely new: Chitra Binding Works was bought at the start of that year for Rs 21,00,000, so year two carries a binding operation that year one did not, and a transaction cost that is charged once and never again.

The third test asks how many estimatesA figure in a set of accounts that nobody can look up, so somebody has to judge it. How long a machine will last, how much of a debt will go unpaid, and how long a lease will really run are all estimates. the profit rests on, and this is the one that surprises people. The instinctive reaction is that a good set of accounts should rest on none. Every set of accounts rests on several. Anjani Stationers' profit sits on a provision for debts that may not be collected, on the useful lives chosen for its assets, on the term assumed for its lease, and on the formula chosen to value what sits in its warehouse. Estimates are always present. The question is how many the profit depends on, and how far the answer moves when one of them moves.

What is an accrual, and why does the accrual component sit at the centre of this?

Profit and cash differ for exactly one reason, and it has a name. An accrualThe act of recording something in the accounts when it happens rather than when the money moves. A sale on credit is recorded when the goods go out, months before the customer pays. records an effect when it happens rather than when the money moves. A notebook order despatched in March is revenue in March even though the school pays in July. A machine bought once is charged a little at a time across the years it works. A business measured only by its bank statement would look absurd, so every one of those decisions pushes profit and cash apart on purpose.

Now turn that round. If accruals are the only reason profit differs from cash, then the gap between profit and cash is precisely the part of profit that cash has not yet confirmed. The gap has a name too. The accrual componentProfit after tax minus operating cash flow, for the same period. The subtraction isolates the part of reported profit that has not yet been confirmed by money arriving. is profit after tax less operating cash flow, and it is the single most useful number in this whole subject because it needs two figures and a subtraction and nothing else.

Do it on Anjani Stationers. Profit after tax was Rs 30,00,000. Operating cash flow was Rs 36,30,000. Rs 30,00,000 less Rs 36,30,000 is minus Rs 6,30,000. The accrual component is negative. Negative accruals mean cash arrived ahead of profit. The business collected more from trading than it wrote down as earnings, and on this test that is the reassuring direction. A large positive accrual component would be the shape worth asking about, because it says the profit exists mostly in balances that somebody still has to pay.

Two cautions belong here immediately, and they are not decoration. First, the sign is a direction and not a grade. A business collecting hard on old debts, or running down its stock, produces negative accruals in a year for reasons that say nothing about the durability of its profit. Second, one year is one reading. Whether the accrual component is negative once or negative for four years running is a completely different fact, and Anjani Stationers has published two years, not four.

Where does the cash conversion test stop working?

Cash is the strongest single test of the three. Its limits need naming exactly because of that strength. Three of them matter.

The first limit is that cash cannot detect a misstatement that moves cash as well. If money genuinely arrived, the cash flow statement records that it arrived. Whether it arrived for the reason described is a different question, and the cash line has no view on it. The second limit is that the test says nothing at all about whether the cash repeats. A one-off receipt is still cash. The godown let out for a wedding season converted to cash beautifully and will not happen again, so the first test passes it and the second test is the one that catches it.

The third limit is the one that catches careful readers, and it deserves the arithmetic. A conversion ratio is a net result, and a net result can be produced by two small movements or by two enormous movements that nearly cancel. Take Anjani Stationers apart. Two large adjustments sit inside its year two bridge from profit to operating cash, pointing opposite ways. On one side, non-cash charges of Rs 18,00,000 are added back: the published depreciation and amortisation of Rs 12,00,000, plus the Rs 6,00,000 charged in the year for debts that may not be collected. Neither of those two sums left the bank, so both come back on the way from profit to cash. On the other side, the working capital cycleThe money tied up in stock and in unpaid customer bills, less the money the business itself has not yet paid its suppliers. When the cycle lengthens it consumes cash. consumed a net Rs 17,00,000, being Rs 17,00,000 more owed by customers and Rs 9,00,000 more sitting in the warehouse, against Rs 7,00,000 of extra credit taken from suppliers and Rs 2,00,000 more taken in advance from customers. The two nearly cancel, and the Rs 5,30,000 that carries cash the rest of the way past profit is timing on interest and on tax.

Read the middle rows again. Two movements of Rs 18,00,000 and Rs 17,00,000 very nearly cancel, and what survives from them into the ratio is a net of Rs 1,00,000, with a further Rs 5,30,000 arriving from timing on interest and on tax. A conversion of 1.21 times produced by two enormous offsetting movements is a different situation from the same 1.21 times produced by two small ones, and the ratio on its own cannot tell which of the two is in front of the reader. Nothing in that sentence says either situation is wrong. Both are ordinary. A business that is growing, stocking ahead of a season and carrying a customer group on long terms will always show large gross movements, and the fact that they offset is arithmetic rather than intent.

How to Build an Earnings-Quality Checklist, and what does a finished one look like?

A checklist here is a list of questions, each attached to the disclosure that answers it. The named source is what makes the list usable. A question with no named source is a wish, and it is the reason most checklists collapse into vague impressions after four lines. Eight questions cover the ground, and every one of them is answered from a document a reader can actually open.

Now the part that decides whether the checklist is worth building. Running it produces a profileA set of separate readings kept separate, each with its own answer and its own limits, rather than combined into a single figure or verdict.: eight independent readings, each with its own answer, some reassuring, some raising a question, and some honestly recording that the answer was not available. A checklist produces a profile and never a score, and anybody who reduces eight readings to one number out of ten has thrown away everything the exercise was built to produce.

Why so strongly? Because the readings are not the same kind of thing and do not add. A cash conversion reading and an estimate reading answer different questions from different documents with different reliability, and there is no exchange rate between them. The third state on the list is the most valuable one: not established. A score destroys it. Two of the eight lines above will often come back unanswered for a business whose full filing is not to hand, and an unanswered line is information. An unanswered line names exactly what to ask for next. Averaged into a number, that line silently becomes a middling result. A middling result is the one thing it certainly is not.

Ind AS 8 and Ind AS 1: which documents govern estimates and presentation in India?

Durability and evidence are ideas rather than local rules, so everything above holds wherever the accounts were prepared. The named documents that carry the requirements behind three of the checklist lines, and every condition inside them, belong in one marked place.

What does the checklist say about Anjani Stationers?

All eight lines are run on the published figures and what each one returns is written down. Tidying the result while writing it is where readers go wrong.

Leaving out the ordinary explanation is how a reading becomes an accusation. Put the ordinary explanation beside each questioning line, and give it the same weight. The receivables gap is largely one customer. The Sunrise Public School group takes about 171 days to pay against about 110 days for everybody else, and holding Sunrise at its prior 144 days would have left the overall figure at 120 days rather than 128, so eight of the nine extra days come from one school group paying exactly on the terms it agreed. The inventory build is a stationer stocking ahead of a school session, bought while paper was rising through the year to Rs 220/- before easing back to Rs 200/-. The provision tripled because the ageing of the book genuinely worsened, and the not-yet-due bucket was identical in both years, so every rupee of the growth sat in balances that were already overdue. The margin fall is Rs 25,00,000 of extra cost below the gross line, itemised as Rs 6,00,000 of employee cost, Rs 12,00,000 of other operating cost and Rs 7,00,000 of depreciation on assets that were bought. Every rupee of it sits below the gross line, and the gross margin never moved.

One detail deserves saying out loud. Raising the provision made the reported profit smaller, not larger. The estimate that moved furthest moved in the direction that costs the business its own reported earnings. That is worth holding on to, because the mental shortcut running underneath most misreadings is that a moving estimate is a flattering one, and here it is the opposite.

Now the close, and it is deliberately unresolved. Anjani Stationers is a business whose profit is well backed by cash and whose estimates moved a long way in one year. Both of those are true at the same time. Neither cancels the other, and the honest output of the checklist is exactly the mixed profile written above rather than a verdict in either direction. A mixed profile is not a failure of the method; it is the normal and correct result of running the method on an ordinary business. What a reader would want next is nameable and finite: a third year of cash conversion, the ageing table for the year after this one to see whether the provision was too high or not high enough, the note giving the cost of the acquisition, the tax reconciliation, and the audit report. The five requests are the whole of the next step, and the mixed profile is where the reading stops rather than a conclusion about Anjani Stationers.

Because a reading that lives only inside a panel is invisible to anyone who cannot run it, here are the settings that matter. At the default, eight questions worked with the notes open, the profile is three reassuring, three questioning and two not established. The estimate, the non-recurring item, the policy question, the tax comparison and the audit report all live in the notes rather than on the face. Switch to the face of the statements only and five of the eight lines go unanswered. A reader left with two green lines and almost nothing else is exactly the position the failure below describes. Switch to the cash conversion line only and the panel returns a single reassuring reading out of eight. Switch to the receivables line only and it returns a single questioning reading out of eight. The last two settings produce opposite impressions from the same published accounts. Nothing shows more clearly that a single line is not a profile.

What can a reader never establish from published figures?

Three things sit permanently outside what a filing can report, and being clear about them is the difference between a useful reader and a dangerous one.

The first is whether an estimate was made in good faith. The provision moved from Rs 3,00,000 to Rs 9,00,000, and the Rs 6,00,000 charge splits into Rs 2,23,000 explained by the ageing and Rs 3,77,000 of judgement. No document shows what was in the mind of the person who signed it. Vaidehi Rao, as finance controller, holds reasons that are not in the accounts and never will be. The second is whether a transaction had a business purpose. Chitra Binding Works invoiced Anjani Stationers Rs 8,00,000 for binding work during the year, of which Rs 1,50,000 was unpaid at the year end, and there is separately an Rs 8,00,000 guarantee over Chitra's borrowing that is disclosed and not recognised. The disclosure reports that those things happened and at what amount. The disclosure does not report why they were structured that way, and the two Rs 8,00,000 figures are different things that must not be added together.

The third is the hardest to accept. A reader cannot establish from published figures whether a well-executed misstatement exists at all. A misstatement built carefully enough to survive an audit is, by construction, built to look ordinary in exactly the disclosures a reader has, so published figures are the wrong instrument for that question. Finding one takes access a reader does not have: the ledgers, the contracts, the correspondence, and the power to ask somebody a question and require an answer.

Naming that limit is not defeatism and not modesty. A filing is a summary prepared by the business, checked to a standard, and published to a format. And the cost of pretending otherwise falls on people who did nothing wrong. A reader who converts three questioning lines into a private conviction will treat an ordinary business as suspect, will say so to somebody, and will be wrong. The cost of a false accusation is paid twice, once by an honest business that loses terms or a customer over an ordinary lengthening of its collection cycle, and once by the reader, whose credibility does not survive being confidently wrong in public.

Who uses an earnings quality reading, and what do they do with it?

Three people open the same set of accounts in the same week for three different reasons, and none of them is doing what the others are doing.

A lender reads the cash line to size the repayment, an analyst reads the estimate lines to size the forecast error, and Vaidehi Rao reads the whole profile to know which questions are coming. Watch each of them. The lender's question is narrow and practical: will the money be there when the instalment falls due. Operating cash flow of Rs 36,30,000 less Rs 13,00,000 spent on assets leaves Rs 23,30,000 of free cash. An instalment is paid in money, so the lender sizes the obligation against the free cash rather than against the Rs 30,00,000 of reported profit. The lender also notices that cash on hand fell Rs 7,00,000 to Rs 5,00,000 during a year that generated Rs 23,30,000 of free cash, and asks where it went. The lender asks the business out loud, and that is what a question is for.

The analyst's use is different. An analyst is building a forecast, and the estimate lines tell them how wrong that forecast can be for reasons that have nothing to do with trading. If Rs 3,77,000 of a Rs 6,00,000 provision charge is judgement, then a forecast of next year's profit carries at least that much uncertainty from one line alone, before a single notebook is sold. The analyst's output is not a verdict about the provision. The output is a range, and the width of the range is what the estimate reading bought them.

And Vaidehi Rao uses it in reverse. Every questioning line on the profile is a question a bank, a customer or a supplier will ask her, and the answer to each of them is ordinary. She runs the same eight questions on her own accounts before anybody else does. Receivables grew faster than revenue because one school group pays on longer terms. Inventory grew because the season is coming. The provision tripled because the ageing worsened and the estimate caught up. Margins fell below the gross line because of Rs 25,00,000 of cost she can itemise. Running the checklist against one's own accounts first is how a finance controller turns a set of awkward-looking divergences into four prepared sentences, and it is the single most useful move available to anybody who prepares accounts rather than reads them.

Anjani Stationers Private Limited, Chitra Binding Works, the Sunrise Public School group and Vaidehi Rao are invented.
Educational material. Not advice on any investment, tax, budget or market position.