Operating Margin: What Expansion and Compression Reveal

An operating margin is operating profit divided by revenue, and one income statement carries three of them. On Sankalp Industrial Systems Limited, an invented manufacturer, the same Rs 12,00,00,00,000 of revenue gives 24.0 per cent before depreciation, 20.0 per cent after it, and 15.00 per cent once the company's own assumed tax rate is applied. So a margin quoted without its numerator is not yet a measurement.

There is a fruit juice cart outside the gate of a college in a small town. The man who runs it will say, without being asked twice, that he keeps thirty paise on every rupee that crosses the counter. He is not lying. He is also not yet saying anything usable, and the reason is worth four minutes.

Thirty paise on the rupee, before what? Before the rent he pays the shopkeeper whose frontage he stands on? Before the instalment on the second mixer he bought last year? Before the money he hands his nephew who works the evening rush? Asked those three questions one after another, he gives thirty, then twenty two, then eighteen, then twelve, and every one of those four numbers is an honest answer to a slightly different question. The number never changed; the line it was measured after did.

The juice seller’s answer is the whole difficulty with an operating margin, scaled up. A large manufacturer's income statement has several places where it is reasonable to stop and call what is left over the operating profit. Each stopping point gives a real margin. Nobody is being dishonest. But two people can quote a margin for the same company in the same year and be four or five points apart, and the gap between them is not a disagreement about the business. The gap is a disagreement about where to stop.

What is an operating margin, and which one does somebody actually mean?

An operating margin puts a profit figure over revenue and expresses the result as a percentage. The division itself is uncontroversial. The trouble is the numerator. On a normal income statement there are three candidates, and all three get called the operating margin in ordinary conversation.

The first stops before depreciation and amortisation, and gives what is usually written as earnings before interest, tax, depreciation and amortisation (EBITDA) over revenue. The second continues past depreciation to earnings before interest and tax (EBIT). The third goes further still and takes the tax charge off as well, giving operating profit after tax. Notice what is not in any of them: interest. All three are struck above the financing line. Sitting above that line is what makes them operating margins rather than profit margins. A company's borrowing changes the profit that reaches its shareholders and does not change any of these three at all.

Every one of the three is a legitimate operating margin, and the only thing that makes a quoted margin useless is not saying which one it is. The habit worth building is small and it costs nothing: whenever a margin is written or read, the numerator travels with it in the same breath. An EBITDA margin. An EBIT margin. A post-tax operating margin. Three words instead of one, and the ambiguity disappears.

Which three margins does one income statement carry, and what separates them?

Take Sankalp Industrial Systems Limited, a manufacturer of industrial valves, precision castings and the aftermarketSupplying spares and servicing for equipment that was sold years ago and is still running, which is a different trade from selling the equipment. parts and service that go with them. Every figure below can be checked against every other figure without anybody having to trust a source.

Its last completed year, called Year 0 throughout, carried revenue of Rs 12,00,00,00,000. Work down from there.

Two items and only two stand between the top margin and the bottom one. Depreciation and amortisation of Rs 48,00,00,000 takes 24.0 per cent down to 20.0 per cent. The tax charge of Rs 60,00,00,000 takes 20.0 per cent down to 15.00 per cent. The 25.0 per cent rate is this invented company’s own assumed effective tax rateThe tax a company actually ends up bearing, read as a share of what it earned before tax, rather than any rate printed in a statute., stated as an assumption of the forecast and not as a fact about anywhere.

The spread between the highest and the lowest of the three is nine full percentage points on one company in one year. Nine points is wider than the entire gap between this company and most of the businesses it would ever be compared with. Nine points of spread is why the numerator matters more than almost any refinement made afterwards.

One more thing worth noticing in the table. Every one of those percentages was computed here from the rupee amounts, not carried across from a label. Recomputing is a small discipline and the single most reliable protection against a stale description sitting on top of a correct set of numbers. Divide the rupees; never trust the caption.

How much does a percentage point of margin cost in rupees?

Before going anywhere near expansion or compression, it is worth converting a margin into money once. After that every argument about margins becomes arithmetic that can be done standing up.

One percentage point of margin on Rs 12,00,00,00,000 of revenue is one per cent of that revenue, or Rs 12,00,00,000. Not approximately. Exactly. So when somebody says the margin should be half a point better, they have said Rs 6,00,00,000, and when somebody worries about a hundred and fifty basis pointsThe unit people reach for because saying a quarter of a percentage point out loud is clumsy. Each one is a hundredth of a point, so a hundred and fifty of them come to one and a half. of pressure on the margin, they have said Rs 18,00,00,000.

Holding that one figure, Rs 12,00,00,000 for each point, sizes any margin argument on this company in a single step without opening anything. Rs 12,00,00,000 a point is the most useful single conversion in this guide.

The conversion has one condition attached and it is easy to forget: it is a point of margin on this revenue. Move to a company with half the revenue and a point is worth half as much. A point is not a unit of money. A point is a unit of money per rupee of sales, and that is why the same margin move means completely different things at two different companies.

At a fixed revenue the link between a margin and the rupees it produces is a straight line, so sweeping the margin across a range would only redraw what one multiplication already settles.

What would a moving margin actually establish?

Now the question the title asks. If a margin goes up, what has been learned?

Less than most people assume, and the honest way to see it is to take the same revenue and put three different margins against it. Two other margins for this same invented company sit in the underlying record under different sets of assumptions: 25.0 per cent and 22.5 per cent. Both of those come from cases that also move revenue, the discount rate and the long-run growth rate together, so neither margin can be read as a valuation on its own. Only the margin is borrowed, applied to the Year 0 revenue that all three cases share.

The whole range is two and a half percentage points wide, and on this revenue that is Rs 30,00,00,000 of EBITDA. Two and a half points of margin sounds like a detail and moves the operating profit by more than a tenth of itself. The asymmetry between the two numbers is why margins get argued about so hard: revenue is the big number, but the margin is the small number sitting on top of it, and small numbers on top of big numbers have enormous leverage.

Here is the part that disappoints people. Knowing that the margin moved from 22.5 to 25.0 per cent establishes that the rupees changed. The move does not establish why, it does not establish whether it will stay, and it does not establish whether anybody did anything well. Cause, persistence and credit are three separate questions, and the margin line answers none of them. A margin move gives a size, not an explanation.

In this forecast the margin does not move, and that is a finding

Everybody expects a growing manufacturer to get more profitable. Rent, the plant manager's salary, the cost of running a quality laboratory: none of that rises when another ten per cent of valves goes through the same shop, so the profit share ought to climb. The expectation of expansion is so strong that a reader will supply it unless it is contradicted.

The forecast below says otherwise, in the plainest possible way.

Revenue rises by half over the five years. The margin sits at 24.0 per cent in all six of them. Depreciation is held at 4.0 per cent of revenue too, so earnings before interest and tax also come out at a flat 20.0 per cent, and the after-tax operating margin at a flat 15.00 per cent, in every single year.

Every rupee of extra profit in this forecast is bought with a rupee of extra revenue at an unchanged rate, and none of it comes from the business getting better at anything. The construction has a consequence worth holding on to: the model has assumed operating leverage away entirely. Revenue up ten per cent lifts operating profit ten per cent, no more. Whether a real valve manufacturer would behave that way is a fair question, and one this forecast leaves open.

Saying this out loud matters. A model whose growth comes entirely from added volume at a constant rate is a completely different object from a model that quietly assumes the business improves, and the two produce very different answers while looking almost identical. A reader has a right to know which one is in front of them, and the only way to give them that is to draw the flat line and name it.

What makes a margin expand, and can the line itself tell which?

Suppose a company's margin went from 22.0 to 24.0 per cent over two years. Broadly three things could have done that, and all three are worth holding together because they are routinely confused with each other.

The first is the fixed cost effect. Revenue grew, a chunk of the cost base did not, so the profit share rose. Nobody negotiated anything; the arithmetic did it. The uncomfortable feature of this one is that it runs backwards just as fast. A company whose margin expanded on the way up will watch it compress on the way down, and the second movement will surprise people who read the first as an achievement.

The second is mix. The company sold more of the profitable thing and less of the thin thing, and every individual product line stayed exactly where it was. Nothing improved. The composition changed. A mix effect persists for as long as the composition does, and a composition may hold for years or for one unusually large order.

The third is price and cost. The company raised prices without losing volume, or bought its steel cheaper, or stopped doing something expensive. Price and cost is the one people mean when they say a margin expanded, and it is the least common of the three.

All three produce exactly the same movement on the margin line, and nothing in a margin figure distinguishes them. Telling them apart requires the revenue composition, some sense of which costs moved and which did not, and the prices. A margin on its own has none of that inside it.

The distinction is not a small point of technique. Naming a cause that cannot be seen is the difference between reporting a number and asserting a story about it, and the story is what people act on. If the record does not separate the three, the honest sentence is that the margin rose two points and the cause is not established. Such a sentence reads as weaker writing and is stronger analysis.

How far apart can the margins inside a single company sit?

Everything so far has treated this company as one thing with one margin. Sankalp is not one thing, and neither is any manufacturer of its size. Sankalp Industrial Systems Limited reports three segmentsA slice of a company that gets its own revenue and profit line in the results, usually because the work inside it is different from the work beside it., and their margins are not remotely alike.

The revenue foots exactly: Rs 6,00,00,00,000 and Rs 4,20,00,00,000 and Rs 1,80,00,00,000 make the Rs 12,00,00,00,000 shown at the top of this guide. So the three divisions between them are the whole company on the revenue line.

Look at the margins. The three division margins run from 23.0 to 30.0 per cent, a spread of seven percentage points inside one set of accounts. And the consolidatedParent and controlled companies added together and presented as one business, with the internal trading between them stripped out. figure, 24.0 per cent, sits below two of the three. The single number on the front of the results is not the middle of the business; it is a revenue-weighted result, and the biggest division pulls it hardest.

Revenue weighting is why the biggest division dominates. Industrial Valves is half the revenue, so its 23.0 per cent gets half the vote, and the 30.0 per cent business gets fifteen per cent of the vote because it is fifteen per cent of the revenue. Taking 24.0 per cent as a description of the work this company does describes a fourth business that none of the three divisions matches.

A 30.0 per cent division is not better run than a 23.0 per cent one. Different work carries different margins, and the reason sits in the work itself rather than in how well anybody performs it. Why an aftermarket business tends to carry the margin it does is a question about industry structure rather than about measurement.

Why the divisions never add up to the whole company

There is an arithmetic problem sitting in that table. The three divisions produce Rs 2,97,00,00,000 of EBITDA between them. The consolidated EBITDA for the same year is Rs 2,88,00,00,000. Rs 9,00,00,000 has gone somewhere.

The Rs 9,00,00,000 has not gone anywhere. The money was never in a division at all. Head office is what it pays for: the chief executive, the group finance team, the audit fee, the insurance policy that covers all three businesses, the building where the board meets. None of it belongs to valves rather than castings, and rather than push it into the divisions on some invented basis, it sits outside them.

Think of a caterer with two counters at a wedding, one selling chaat and one selling sweets. Each counter's takings and each counter's costs are easy to attribute. The van that brought both counters, and the man sitting at the entrance doing the billing for both, belong to neither. Totting up the two counters and calling the result the evening's profit forgets the van.

As a margin, the same fact looks like this.

The head office of this invented company costs exactly three quarters of a percentage point of consolidated margin, and that is the entire reason a segment table never foots to the reported figure. Once that gap has been seen it is hard not to look for it every time, and it is usually the most interesting line in a segment table. No division owns the cost, so nobody is directly accountable for it.

One warning that belongs here rather than anywhere else. The line between what sits in a division and what sits at the centre is a reporting choice, not a law of nature, and choices of that kind get revisited. If a company moves Rs 3,00,00,000 of cost out of the centre and into the divisions, every division margin falls, the consolidated margin does not move by a single basis point, and nothing whatever has happened to the business.

The consolidated margin can move while every division stands still

Go back to the wedding caterer for a second, because this is where the two counters earn their keep. Chaat sells hard and returns very little on each plate. Sweets sell slower and return a great deal on each box. The caterer changes nothing at all: same recipes, same prices, same staff. One evening the sweets counter simply has a longer queue than usual. His margin on the evening goes up. He did not get better at anything.

Now run the same thing on the invented company, on the locked figures, as an illustration and not as a forecast. Take Rs 60,00,00,000 of revenue out of Industrial Valves and put it into Aftermarket Parts and Service. Every division keeps its own margin exactly where it was.

The shortcut is worth learning because it removes the table. Rs 60,00,00,000 of revenue moved across a gap of seven percentage points, so segment EBITDA rises by seven per cent of Rs 60,00,00,000, or Rs 4,20,00,000. Added to the Rs 2,97,00,00,000 the divisions produced before, that gives Rs 3,01,20,00,000, exactly as the table says.

Head office cost has not changed. Take off the same Rs 9,00,00,000 and consolidated EBITDA becomes Rs 2,92,20,00,000. Nothing was sold that was not being sold before; it was only sold by a different part of the same company, so revenue is still Rs 12,00,00,00,000. Divide, and the consolidated margin is 24.35 per cent against 24.00 per cent before.

The consolidated margin rose thirty five basis points and not one division did anything differently. Run the shift the other way, moving Rs 60,00,00,000 from aftermarket into valves, and the consolidated margin compresses by the same thirty five basis points with nothing deteriorating anywhere.

If one sentence here is worth carrying away, that is the one. A rising consolidated margin is not evidence about operations until the composition underneath it has been examined. A shift in composition is a statement about which parts grew, and growth in the parts is a completely different question from quality in the parts.

The same dispersion shows up across companies, and it is not much wider

Six invented industrial businesses sit in the underlying record as a set of comparable companiesOther businesses put beside this one because somebody judged them to be built the same way, which is a judgement rather than a fact.. Their EBITDA margins run 20.0, 21.0, 23.0, 24.5, 26.0 and 30.0 per cent. Sankalp Industrial Systems Limited, at 24.0 per cent, sits a little above the middle of them.

Now put the two dispersions side by side. The comparison is the point. The six outside companies span ten percentage points, from 20.0 to 30.0. The three divisions inside this one company span seven, from 23.0 to 30.0. One company's internal spread is seventy per cent as wide as the entire spread across the six businesses it would be compared with.

The consequence is uncomfortable for anybody who reaches for a peer margin as a benchmark. If a competitor reports 26.0 per cent against this company's 24.0, one clean explanation is that the competitor sells a larger share of the aftermarket kind of work and a smaller share of the valve kind. Neither business is being run better. The two businesses are differently composed, and the margin line is reporting the composition.

So a margin comparison across companies carries an assumption that almost nobody states: that the businesses are built the same way, with comparable mixes of work and comparable costs sitting inside the same lines of the accounts. Sometimes that assumption holds. When it does not, the comparison is measuring the wrong thing very precisely. Which businesses belong in such a set, and what a multiple built on them would imply, are covered separately.

The limit: a margin says nothing about the capital it took to earn it

One last thing, and it is the boundary of the whole measure. A margin measures how much of each rupee of sales survives to the operating profit line. The margin says nothing at all about how many rupees had to be sunk into the business to produce that rupee of sales.

Two shops on the same street make the point faster than any formula. A jeweller keeps a great deal on each sale and turns his stock perhaps twice a year. A vegetable seller keeps very little on each sale and turns his stock every single day. Ask which one has the better margin and the jeweller wins easily. Ask which one earns more on the money tied up in the shop and the question becomes genuinely hard.

The finance version needs one extra term. Capital turnoverRupees of revenue per rupee of money tied up in the business. A workshop that sweats its plant twice as hard turns capital twice as fast. is revenue divided by invested capital, and a return on invested capitalProfit after tax from operations set against every rupee sunk into the operation, borrowings and shareholders' money together. is the after-tax operating margin multiplied by that turnover. The margin is one of the two axes and it is silent about the other.

On this invented company the two happen to coincide, and the coincidence is worth showing precisely because it is a coincidence. Invested capital at Year 0 is Rs 12,00,00,00,000 against revenue of Rs 12,00,00,00,000, so capital turns exactly 1.00 times. Operating profit after tax is Rs 1,80,00,00,000, a 15.00 per cent margin. Multiplying 15.00 per cent by 1.00 lands back on 15.00 per cent, and that is also what each rupee of invested capital earns here.

The equality is a property of these particular locked figures and not a rule, and it holds only because one rupee of capital happens to carry exactly one rupee of revenue here. A business at the same 15.00 per cent margin turning its capital twice would earn 30.00 per cent. A business at half the margin, 7.50 per cent, turning twice, would earn the same 15.00 per cent as this one on a margin half the size. How those returns are defined and which measure answers which question is settled separately. The limit is the part that belongs here: the margin is one axis of a two-axis question and cannot answer it alone.