How to Interpret Operating Leverage in a Company's Results

Interpreting a leverage figure is a five step procedure. Establish the split it rests on. Compute it for two years rather than one. Test whether the change came from the numerator or the denominator. Check what the figure assumes about the year ahead. Write down what would invalidate it, and stop. Run in order, the procedure shows what a cost structure does to a revenue movement. The procedure cannot say whether that structure is wise.

Here is the join that fixes the order. Step one produces a written assumption, and everything after it is conditional on that assumption being visible. Step two produces a direction. A single figure never carries one. Step three produces the driver. A reader of the figure alone cannot see the driver at all. Step four produces an expiry condition. Step five turns all four into a sentence somebody else can check. Remove any one of them and whatever follows is left with nothing underneath it, and that is what makes these five an order of work rather than a menu.

How the fixed and variable split is estimated, how contribution is computed, what the leverage figure means about a revenue movement and why the ratio takes the shape it takes are each covered separately and taken as given here. Four things remain: the sequence, the one test that separates two completely different situations reporting the same number, the shelf life of the reading, and the point at which the work stops. The common failure in reading operating leverage is almost never a slip in the arithmetic; it is taking a figure that moved for one reason and reading it as though the other reason had moved it.

Why does the split come before the arithmetic?

Step one is to establish the fixed and variable split the figure rests on, and to write it down in the analysis where a reader can see it, not in a private working file. The reason is uncomfortable and worth stating flatly: no Indian company publishes a fixed and variable split in its statutory statements, so every leverage figure anybody quotes was built on an estimate somebody made. Two analysts working from the identical filing, splitting the same costs differently, will report different leverage figures, and both will be correct arithmetic. A leverage figure without its split attached therefore cannot be interpreted at all.

Take the household version first. Two people are asked to work out what share of a household's monthly spending is unavoidable. One counts rent, the school fee and the loan instalment. The other counts those and adds the internet connection, the milk delivery and the gym membership, on the ground that none of them has ever once been cancelled. Neither has made a mistake. The two have drawn the boundary in different places, and every figure that flows from the boundary moves with it. A single number handed over without saying where the boundary was drawn is something nobody can use.

Run step one on Anjani Stationers Private Limited and the output is a block of writing, not a number. All of the cost of materials consumed is treated as variable. Carriage outward and packing move with volume, so that slice of other operating costs, Rs 5,40,000 in year one and Rs 6,00,000 in year two, is treated as variable too. Everything else is treated as fixed: employee benefits, the rest of other operating costs, and depreciation. The split gives contribution of Rs 1,02,60,000 in year one and Rs 1,15,50,000 in year two, against fixed costs of Rs 49,60,000 and Rs 74,00,000. Each pair ties back to the published earnings before interest and tax (EBIT) of Rs 53,00,000 and Rs 41,50,000 exactly. Any split of the same total cost ties. The tie is a completeness check on the arithmetic, not evidence that the boundary was drawn in the right place.

Step one has one other output, and it is the one people skip. Sometimes the notes do not break the other expenses line into components that can be classified, and the split cannot be established. Then the output of step one is that the figure cannot be trusted, and the work stops there. Stopping there is a legitimate result of running the procedure, not a failure to run it. A reader who writes down operating leverage is 2.78, having quietly assumed a split nobody can see, has produced a figure with a hidden input, and hidden inputs are how a reading survives being wrong.

What does a second year add that one figure cannot?

Step two computes the figure for two years. Not for the latest year alone, for two. A single leverage figure is a fact about a moment. The scale has no natural landmarks, so 2.78 is not high or low against anything present in the accounts, and there is almost nothing a reader can do with it. Two figures are different in kind. Two figures carry a direction, and a direction is a statement about what happened to the business between one balance date and the next. Anjani Stationers reading 2.78 on its own supports no sentence at all. A reading of 1.94 rising to 2.78 supports one immediately: the cost structure changed materially inside twelve months.

The second figure is what licenses the two words in the middle of that sentence. Changed materially. The word materially is available because there are two points on the same scale, computed on the same split, by the same person, from the same kind of statement. None of that comparability exists between one analyst's figure and somebody else's figure for another business, and it exists least of all between a computed figure and a remembered sense of what is normal. The second year is the only comparison the analyst actually controls. Step two is short. The whole of it is the instruction to compute the comparativeThe prior period column presented alongside the current period in a set of accounts, so that each figure can be read against the same figure a year earlier. year on exactly the split documented at step one, not on a different one.

Did the numerator move, or did the denominator?

Step three is the reason this reading is a procedure rather than a summary of an arithmetic result. The leverage figure is contribution over EBIT. The figure rises when the numeratorThe figure on the top of a fraction, above the dividing line. rises, and it rises just as obediently when the denominatorThe figure on the bottom of a fraction, below the dividing line, which the top is divided by. falls. The two rises are not variations of one story. A business whose contribution grew is a business that sold more or held its prices. A business whose EBIT fell is a business that did worse. The same rise from 1.94 to 2.78 describes both of them, and the figure by itself cannot tell which of the two is in hand.

Here is the household version, and it is a familiar conversation. A friend says the share of their income committed to fixed monthly outgoings has gone from a third to half. The first question, before anything else is said, is whether the outgoings rose or the income fell. Half an income committed after a promotion and a bigger flat is a different life from half an income committed after a pay cut with the same flat. The ratio is identical. The two situations are not remotely alike, and nobody would dream of responding to the ratio without asking. Step three is that question, written into the procedure so it does not depend on whether it happens to occur to anyone.

Running it is arithmetic on two numbers already in hand. Contribution went from Rs 1,02,60,000 to Rs 1,15,50,000, a rise of Rs 12,90,000 or 12.6 per cent. EBIT went from Rs 53,00,000 to Rs 41,50,000, a fall of Rs 11,50,000 or 21.7 per cent. Both moved, and they moved in opposite directions, so both pushed the reading up. To say which did more, each moves one at a time. Holding EBIT at year one's Rs 53,00,000 and letting contribution rise alone: the reading goes 1.94 to 2.18, a rise of 0.24. Holding contribution at year one's Rs 1,02,60,000 and letting EBIT fall alone: the reading goes 1.94 to 2.47, a rise of 0.54. The two together give 2.78, a rise of 0.85, of which 0.07 is the part that only exists because both moved at once. The denominator effect is more than twice the numerator effect, so most of Anjani Stationers' rise in leverage is EBIT falling rather than contribution growing, and step three's output is that sentence and no more of it than that.

Notice carefully what step three has not said. Step three has not said the fall is bad, that the business is in trouble, or that the leverage figure is a warning. Step three has said which of two inputs moved. A smaller claim than most readers want is the claim the arithmetic supports, and the gap between the two is where most misreadings of this figure live.

The readings are worth writing down. At nought decimal places both paths read 3. At one, both read 2.8. Leverage figures are quoted to two places in practice everywhere, and at two both read 2.78. The first difference appears at three decimal places, where path one reads 2.783 and path two reads 2.780, a difference of three thousandths. Three thousandths of a ratio is the only visible trace of a gap of Rs 12,90,000 in contribution and Rs 4,60,000 in EBIT. No achievable precision in the ratio recovers the situation underneath it. The instinct when two numbers look the same is to measure them more finely. Here measuring more finely produces a difference so small that anybody would round it away, while the two businesses it describes are not alike in any respect that matters.

How long does a leverage reading stay true?

Step four asks what the reading assumes about the year ahead, and the answer is that it assumes the fixed costs stay fixed. The assumption is not a technicality; it is the entire shelf life of the figure. So step four is a search through what has already been disclosed for anything that will change the fixed base: a lease signed, a plant or warehouse commissioned, staff hired, an acquisition completed, an asset put into use partway through the year. A leverage reading has a shelf life measured in disclosed commitments rather than in months, so a figure computed on year two's accounts can already be void on the day it is computed if the notes say something has been signed.

The word doing the work there is disclosed. Step four does not speculate about what a business might do next; it reads what is already recorded. A commitmentAn obligation a business has already entered into that will require payment in a future period, described in the notes rather than shown as a figure on the face of the statements. disclosed in the notes is a fact about the future that has already happened in the sense that matters. So is an asset commissioned in the last month of the year. Next year that asset carries a full year of depreciation against the fraction it carried this year. Neither requires a forecast. Both end the reading.

Run it on Anjani Stationers and the first thing that surfaces is flagged by the balance sheet rather than by any cost line. Chitra Binding Works was bought at the start of the year, so an investment of Rs 21,00,000 sits in year two where year one carried nothing. Be precise about what the acquisition can and cannot do to the fixed base. It is easy to get wrong. The accounts are standalone figures, the parent company by itself. The wages and running costs of the binding operation have therefore never formed part of the Rs 74,00,000. The acquisition reaches this base by two narrower routes instead. The first is the cost of doing the transaction, the professional fees and the stamp duty of buying a business. Transaction cost is expensed as incurred, so it sits inside year two's other expenses and inside no part of year one's comparative. The second is the relationship. Work Anjani Stationers used to do in its own shed, or used to buy from somebody unconnected, may now be bought from an operation that became a related partyA person or business connected closely enough to the reporting business that their dealings are disclosed separately, because the terms may not be the terms a stranger would have got. in April.

Both of those routes end the reading, and they end it in different ways. A transaction cost expensed once sits inside the fixed base this year and will not be there next year, so the base is already known to be overstated for the purpose the reading puts it to, by an amount the face of the statement does not give. The related party route is subtler and matters more on this figure than on any other built on the same split. Work moved out of the shed and bought in from a connected operation may arrive on entirely different terms: a fixed retainer behaves as a fixed cost, a price per ream behaves as a variable one, and the same physical work therefore lands on either side of the split depending on a contract nobody outside can read. A contract of that kind changes the split itself with nothing whatever happening to the business's activity, and the split is what the whole reading rests on.

Two more surface without leaving the ladder. Depreciation and amortisation went from Rs 5,00,000 to Rs 12,00,000, and depreciation is a fixed cost that does not fall when revenue falls, so a base that has just grown Rs 7,00,000 on this line is a base with a floor under it. And other operating costs treated as fixed went from Rs 8,60,000 to Rs 20,00,000. The split attributes part of that rise to new warehouse costs, and a warehouse taken during the year sits in year two's figures for part of a year and in year three's for all of it. Every one of these says the same thing. The year two reading of 2.78 is a reading of a base that has not finished moving.

When does the reading stop, and what goes out?

Step five writes down what would invalidate the reading, names the specific disclosure that would change it, and stops. The stopping rule has four conditions and they are met together, not one at a time: the split is documented, the direction is established, the driver is identified, and the expiry conditions are named. If all four are met the work is finished whatever it found. If any one is missing it is not finished however confident the output sounds. A rule of the form stop when it feels understood can never be checked by somebody else. The rule is written this way so that it can be.

The output is a conditional sentence rather than a number, and the sentence is the part that disappoints people. For Anjani Stationers it reads roughly like this. On a split that treats materials and a Rs 6,00,000 slice of other operating costs as variable and everything else as fixed, the reading moved from 1.94 to 2.78 between the two years, and most of that movement is EBIT falling rather than contribution growing. The reading assumes the Rs 74,00,000 fixed base holds, and it will not hold once the one-time cost of buying the binding operation drops out of it, nor if what is bought from that operation as a related party changes in scale or in terms, nor if the assets behind the Rs 12,00,000 depreciation charge are added to, nor if any lease or capital commitment in the notes is exercised. Change the split and the reading changes with it. The whole paragraph is the output. Every clause in it is either an arithmetic result or a named condition, and there is no sentence in it about the business.

The output names its own author. Change the split and the reading changes: that clause is why step one insisted the split be written where a reader can see it, and it is what makes the difference between a figure that can be checked and a figure that can only be believed. A reader who disagrees with the analyst's classification of the warehouse cost can recompute from that same block and see exactly how much of the reading the disagreement moves. Recomputing from the block is the whole practical value of running the five steps in order rather than reporting the ratio, and it is worth more than the ratio ever was.

Who runs this reading, and what do they do with it?

Three people run these five steps on the same accounts in the same month, and the fifth step means something different to each of them. A lender runs the reading to work out how far revenue can move before the interest stops being covered, an analyst runs it to know which parts of a forecast rest on an estimate rather than a figure, and Vaidehi Rao, the finance controller inside Anjani Stationers, runs it to find out what a reader outside would conclude from what has been published.

The lender has the clearest use, so follow the lender first. The fixed base of Rs 74,00,000 is the part of the cost structure that will still be there in a bad year, and finance cost of Rs 3,50,000 sits below it. So the lender reads step four hardest of all: any disclosed lease or capital commitment moves the number that has to be covered before interest is reached. The lender reads step three next. A leverage figure that rose because contribution grew and one that rose because EBIT fell describe very different exposures on the same loan. The lender has an advantage nobody else has: it can write and ask for the split rather than estimating it. The conditional sentence at step five is then a list of questions the lender can actually get answered.

The analyst cannot ask. For her the value of step five is that it marks the boundary of her own model: the Rs 6,00,000 of carriage and packing she treated as variable is an assumption, and if she had put it with the fixed costs her whole reading would shift. Next year she will need to know whether a movement in the figure came from the business or from her having reclassified something in between. The paragraph, not 2.78, is what she carries forward. Written readings compare across years. Remembered ones do not.

Vaidehi Rao's use is the strangest, and it is the most useful of the three. She already knows the split, exactly, from the ledgers, so nothing in steps one to three tells her anything she did not know. She is testing whether an outside reader who runs these five steps on the published statements would land somewhere defensible or somewhere wrong. If the fixed base looks like it grew 49 per cent because the business overspent, when part of it is the one-time cost of buying the binding operation and part is work that has simply moved out of her own shed and is now bought from a related party, then she can put a note in front of a reader now instead of arguing the point a year late. Same five steps in all three hands. The difference is whether the person running them can do anything about the conditions at the end.

What can this reading never conclude?

Four sentences look like the natural end of this work and none of them is available from it. The first is that the business is risky, and a ratio of two estimated figures cannot establish that. The second is that the leverage figure is too high, and nobody has supplied a standard to judge it against. The third is that the cost structureThe mix of costs a business carries, described by how much of the total stays the same when volume moves and how much moves with it. should change, and a decision of that kind needs inputs the statements do not hold. The fourth is anything at all about the borrowing. Borrowing is a different leverage entirely, computed on different lines, answering a different question. The procedure describes one mechanical property of a cost structure, what a revenue movement does to a profit figure, and nothing beyond that.

The reason to say this bluntly is that the last of the four is the easiest to slide into without noticing. A reading of 2.78 followed by a sentence about the finance cost of Rs 3,50,000 reads as one continuous thought and is two unrelated ones. Operating leverage is about the shape of the operating cost base. The amount a business borrowed, on what terms and against what security, is a separate subject and a separate arithmetic, and joining the two in a paragraph does not join them in fact. Keeping them apart costs nothing and is the difference between a reading a colleague can check and a paragraph that sounds knowledgeable.

So what is the last line of this reading?

Not a number and not a verdict. Anjani Stationers read 1.94 in year one and 2.78 in year two, on a split that no statement discloses and that an analyst estimated; most of the movement is EBIT falling rather than contribution growing; the fixed base behind it rose 49 per cent in a single year while revenue rose 12.5 per cent; and the reading holds only until the next disclosed commitment. The paragraph is the finished output, and it is longer than 2.78 for one reason. Every clause in it is a condition somebody else can check, and the bare figure withheld exactly that.

One thing above the rest separates a reading of operating leverage from a summary of it. Hold on to it. The figure went up. Going up is not the finding. The finding is that the profit figure underneath it got smaller, and the figure went up mostly for that reason. The margin ladder had already told anybody who read it as much, so the leverage figure added a description of the cost structure and not a second finding. Getting to that sentence takes five steps in order, and the fifth of them is the discipline to write down what would make it wrong.

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