Depreciation, amortisation and capitalisation — the quietest lever on profit

A telecom company spends ₹6,000 crore upgrading its network. The cash leaves the company — the cheques are written, the equipment is installed. Now the finance team faces a choice that the cash has already settled but the accounts have not. Does that ₹6,000 crore count as a cost of this year, charged in full against this year's profit? Or is it an asset — something the company now owns and will use for a decade — to be charged against profit a little at a time, ₹600 crore a year for ten years? illustrative

The cash is identical either way. But the reported profit is not. Charge the whole ₹6,000 crore now and a ₹3,000 crore profit becomes a ₹3,000 crore loss. Spread it over ten years and this year carries only ₹600 crore of it, so the company still reports a healthy profit. Same business, same cash, same year — and a swing of ₹5,400 crore in the number at the bottom of the profit-and-loss account, decided entirely by which of two boxes the spend is written into.

This module is about that choice, and the two slow charges it creates: depreciation, which spreads the cost of a physical asset, and amortisation, which spreads the cost of an intangible one. It is the quietest lever on reported profit, because nothing dramatic happens on the day it is pulled. There is no sale to point to, no cost that visibly rose or fell. A single estimate changes, and the profit moves with it.

Module 6 established that profit is an opinion — the sum of many judgements. This module takes the single largest of those judgements for an asset-heavy business and shows exactly how it works, because it is the one most often used to make a weak year look strong.

Start with the honest reason the choice exists at all. A company that buys a machine for ₹100 crore and uses it for ten years has not really suffered a ₹100 crore cost in year one. It has bought ten years of use. So accounting does something sensible: it puts the machine on the balance sheet as an asset and charges roughly ₹10 crore a year to profit for ten years. That yearly charge is depreciation. The act of putting the spend on the balance sheet in the first place, rather than expensing it immediately, is capitalisation. When the asset is not physical — a software platform, a broadcasting licence, a drug patent, a film library — the same yearly charge has a different name, amortisation, but it does the identical job.

The trouble is that two of the inputs to this machinery are estimates, not facts. The first is the decision of whether a given spend is even an asset. Is ₹6,000 crore of "network enhancement" a lasting asset, or is it really this year's running cost dressed up? The second is how long the asset lasts — its useful life — which sets how fast its cost hits profit. Both are chosen by management, within a range the auditor will accept. And both feed straight into reported profit. Stretch the useful life and every year's charge shrinks. Capitalise a cost that should have been expensed and this year's profit never sees it at all. Neither move requires the business to sell one extra unit. That is why this is the lever to understand: it moves the profit without moving the business, and it does so in near silence.

Follow a single ₹6,000 crore spend through the two routes it can take. The cash flow statement records the same ₹6,000 crore leaving either way — that never changes. What changes is the profit-and-loss account.

Route one — expense it now. The whole ₹6,000 crore is treated as a cost of this year. It lands on the P&L in full. If the business would otherwise have earned ₹3,000 crore before tax, it now reports a ₹3,000 crore loss. Nothing goes on the balance sheet.

Route two — capitalise it. The ₹6,000 crore is treated as an asset. It goes on the balance sheet at ₹6,000 crore, and only a slice is charged to this year's profit. If the asset is judged to last ten years, that slice is ₹600 crore — one-tenth. This year still reports a ₹2,400 crore profit. Next year another ₹600 crore is charged, and the asset on the balance sheet drops to ₹5,400 crore, then ₹4,800 crore, and so on until it is fully written off.

Notice what the capitalising route quietly does. It borrows profit from the future. The ₹5,400 crore it did not charge this year has not vanished — it will be charged over the next nine years, ₹600 crore at a time. If the asset genuinely earns for those nine years, that is fair: the cost is being matched to the income it helps produce. If the asset does not earn — if the "network enhancement" turns out to be routine maintenance, or the film flops, or the app is abandoned — then the company has simply moved a real cost off this year's profit and parked it on the balance sheet, where it will sit until someone forces a reckoning.

That reckoning has a name: impairment. When a capitalised asset can no longer earn back the value it is carried at, the difference must be written down, and the write-down hits profit in the year it is finally taken. Impairment is the delayed bill for capitalising too eagerly. A balance sheet full of intangibles "under development" is a balance sheet full of costs that have not yet been charged to anyone — and one honest year can bring them all due at once.

So there are three separate judgement points, and each moves profit on its own. First, capitalise or expense — does the spend go on the balance sheet at all? Second, the useful life — over how many years is a capitalised cost spread? Third, impairment — when is a capitalised asset admitted to be worth less than its carrying value? A management team that wants a better profit this year can lean on any of the three, and none of them will look like anything on the face of the P&L.

The naive reading is that capitalising a cost is aggressive and expensing it is honest. That rule is too simple, and it inverts depending on what the spend actually buys. Here are four sectors where the same judgement carries a completely different weight.

The point of laying them side by side is that the accounting entry is the same in every box — a cost moved from the P&L to the balance sheet — while its honesty runs from plainly correct to plainly aggressive. You cannot judge the choice by the choice. You have to ask what the spend bought and whether that thing will still be earning over the life the company assumed. For telecom, the answer is almost always yes. For an early consumer platform capitalising its staff cost, the answer is often no, and the same entry that is prudent for the network becomes the tell.

Take the telecom composite. It reports revenue of ₹89,000 crore and EBITDA of ₹39,000 crore — a fat 44% margin that looks, on that line alone, like a wonderful business. Then the two slow charges arrive. Depreciation on the network takes ₹21,000 crore. Amortisation of spectrum — the licence to use the airwaves, an intangible — takes another ₹8,000 crore. Between them they consume three-quarters of the EBITDA, and what looked like a 44%-margin machine reports an operating profit of ₹10,000 crore and, after ₹7,000 crore of interest, a pre-tax profit of just ₹3,000 crore. illustrative

This is the first thing to read live: for a capex-heavy business, EBITDA and profit live in different worlds, and the distance between them is depreciation and amortisation. A telecom, a cement maker, a utility — each can show a handsome EBITDA margin and a thin bottom line, and the gap is not a trick. It is the real, slow cost of the assets the business runs on.

Now watch the lever. Against that ₹3,000 crore pre-tax profit, the ₹6,000 crore network spend from the opening is enormous. Expense it and the year is a ₹3,000 crore loss. Capitalise it over ten years and the year is a ₹2,400 crore profit. The single choice is larger than the entire reported profit of the business. This is why, for an asset-heavy company, the depreciation and amortisation policy in the notes is not fine print — it is one of the most important pages in the annual report. Two telecoms with identical networks and identical cash flows can report very different profits purely because one assumes its equipment lasts eight years and the other assumes twelve. The one assuming twelve reports higher profit every year, right up until its older equipment has to be replaced sooner than its accounts implied.

The habit to build: whenever a company owns a lot of long-lived assets, find three numbers before you trust its profit. The depreciation-and-amortisation charge as a share of EBITDA, so you know how much of the operating profit is really the slow return of past capex. The assumed useful lives, which you compare against peers — an outlier life is an outlier profit. And the movement in "capital work in progress" and "intangibles under development", which is where costs can sit un-charged. None of these is on the face of the P&L. All of them decide what the face of the P&L says.

Pick a spend type and toggle it between "expense now" and "capitalise and depreciate". Watch two numbers move: the amount charged to this year's profit, and the amount that lands on the balance sheet instead. When you capitalise, the tool draws the trail of small charges that will hit profit in every future year until the asset is written off.

The presets are the real judgements from the four sectors above. The telecom network is the case where capitalising is plainly right and the useful life is the live question. The SaaS development cost is the case where the identical entry is most often the profit-flatterer. Switch between them and notice that the arithmetic is exactly the same — only the honesty of it changes with the business you point it at.

Reading the depreciation and capitalisation policy tells you how a company is choosing to report the cost of its assets. It does not, on its own, tell you whether the underlying assets are any good. A company can depreciate honestly over conservative lives and still own a network that customers are leaving, or a plant making a product nobody wants. The policy is about the timing of the charge, not the value of the thing being charged.

Nor does it tell you the cash story. Depreciation and amortisation are non-cash charges — they lower profit but move no money — so a company with heavy D&A can be far more cash-generative than its thin profit suggests. That is precisely why the cash flow statement adds depreciation back. The capitalisation choice moves reported profit around; it does not create or destroy a rupee of cash. To judge the business you still have to go to the cash flow statement and ask whether the capex being capitalised is actually earning a return, which is a question this module sets up but the return-on-capital modules answer.

And it cannot, by itself, prove intent. A stretched useful life might be an honest reassessment — equipment really can last longer than first assumed — or it might be a reach for this year's profit. The policy shows you the lever has been pulled and by how much. Whether it was pulled in good faith is something you infer from the pattern over several years, from how the company's lives compare with its peers, and from whether the capitalised assets ever turn into the earnings that were promised for them.

How it comes up. When profit rises faster than the business seems to justify, a sharp analyst goes looking for the accounting reason. On an asset-heavy company the first place to look is the depreciation policy. The question sounds like this: "Your depreciation charge fell this year even though gross block went up. Did the useful-life assumptions change, and what was the profit impact?" What the analyst is really asking is whether the profit growth is operational or just a change of estimate.

A good answer, verbatim-style.

"Yes, we reassessed the life of our core network equipment from eight to ten years, based on a technical review showing the kit is lasting longer in the field. The impact was about ₹1,900 crore lower depreciation this year, and we've disclosed it in note 3. Stripping that out, underlying profit before tax grew about 6%, not the 30% the headline shows. The new lives are in line with what our tower and equipment vendors quote and with global peers."

A cause, a quantified impact, the location of the disclosure, and — crucially — the underlying growth with the estimate change removed. It lets you separate the accounting from the business.

An evasive answer, verbatim-style.

"We periodically review our asset lives in line with the accounting standards and our auditors are fully comfortable with the treatment. Depreciation reflects the useful economic life of the assets. Overall it's been a strong year of profit growth driven by operating leverage across the network."

This is not a strawman — it is the fluent, standard-sounding answer real managements give, and every sentence in it is technically true. What makes it evasive is what it omits. It never says the lives were changed, by how much, or what the profit impact was. It waves at "operating leverage" to explain a jump that was partly an estimate. And it hides behind the auditor's comfort, which confirms the treatment is permitted, not that the profit growth is real.

The follow-up nobody asks. "Can you give the exact rupee impact of the life change on this year's profit before tax, and the underlying growth excluding it?" That question turns the vague answer into a number you can hold. Watch what happens when it is not asked. If "we review our lives periodically" is allowed to stand unquantified, the reader is left crediting the business for a profit jump that a single estimate delivered. The silence is the tell — either the impact is embarrassingly large, or the analysts who would have pressed have stopped doing the work.

The first trap is treating EBITDA as if it were profit. For an asset-heavy business, the whole point of the depreciation and amortisation charge is that the assets really do cost money as they wear out and go obsolete. A telecom's 44% EBITDA margin is genuine, but it is not the return to shareholders — the network has to be renewed, and the D&A charge is the honest, if slow, admission of that. Anyone who values a capex-heavy company on EBITDA, ignoring the D&A, is pretending the assets are free. They are not; they are merely paid for in a different year.

The second trap is admiring rising profit that came from a falling depreciation charge. If a company owns more assets each year, its depreciation should generally rise. When depreciation falls while the asset base grows, the most common cause is a stretched useful life — a legal, disclosed, entirely reversible reach for this year's profit. It flatters every year until the assets wear out faster than the accounts assumed, and then it reverses in a lump. The face of the P&L will never flag it. Only the note on depreciation policy, read year over year, will.

The third trap is trusting a balance sheet swelling with "intangibles under development" or "capital work in progress". These are the waiting rooms for capitalised cost. A spend parked there has not been charged to profit and is not yet being amortised either — it is a cost in suspended animation. Sometimes that is perfectly proper: a plant genuinely being built, a platform genuinely being written. But it is also exactly where a company hides this year's running costs to protect this year's profit, and the bill arrives later as an impairment. When those line items grow far faster than the business, the question is not "what a lot of investment" but "how much of this is really this year's expense, waiting to come due?"