Capital Expenditure: What Gets Capitalised and Why It Matters

Capital expenditure buys something that will serve for more than one period, so its cost sits on the balance sheet and reaches the income statement slowly through depreciation. Operating expenditure is consumed now and lands now. The test is whether the spend produces a resource the business controls that will bring future benefit, and the answer decides which statement the money appears in and when profit feels it.

Here is what sits underneath that. A business spends money on two entirely different kinds of thing, and it spends both out of the same bank account on the same day, so the bank statement cannot tell them apart. Some of the money buys the capacity to trade: a machine, a roof, a vehicle, a piece of software. The rest of the money is the trading itself: wages, paper, power, the courier bill. Accounting refuses to treat those two as the same event, and the rule it uses to separate them is the whole subject.

A household shows the identical shape, and it is much easier to feel there. A household buys a refrigerator for Rs 40,000 and pays an electricity bill of Rs 2,000. Both left the same account in the same week. But at the end of the month the electricity is gone and the refrigerator is standing in the kitchen, and it will still be standing there in eight years. Nobody would say the household is Rs 42,000 poorer in any meaningful sense; it is Rs 2,000 poorer and Rs 40,000 rearranged. Capital expenditureMoney spent to acquire, build or improve something that will serve the business for more than one accounting period. The cost becomes an asset first and an expense later. is the refrigerator. Operating expenditureMoney spent on running the business through the current period, consumed as it is incurred and charged against this period's profit in full. is the electricity bill.

What follows is the rule: a three-part test for the hard cases, the point at which costs stop attaching to a machine, the difference between a repair and an improvement, one business's Rs 12,00,000 of year two machinery traced through all three statements to the rupee, and the three things an outside reader can actually check when a capitalisation policy looks generous.

Capex vs Opex: what separates a cost that stays from a cost that goes?

A contrast is useless if either side is fuzzy, so define both properly first. Capital expenditure is money laid out to acquire, build or improve something that will still be doing work for the business after this period has closed. The spending creates or enlarges a resource. Operating expenditure is money laid out to keep the business running through this period, and once the period closes there is nothing left of it: the power was burned, the wages were earned, the paper was cut into notebooks and sold.

The line between the two is not about how much was spent, who approved it, or whether the money was borrowed or earned, but purely about whether something longer-lived came out of the spending. That is worth saying twice because almost every wrong classification comes from importing a different question. A Rs 40,000 machine part and a Rs 40,000 electricity bill are the same size and land in different boxes. A Rs 500 spanner and a Rs 5,00,000 machine can land in the same box. Size is a matter of practical policy, described in the notes. It is never the test itself.

Anjani Stationers Private Limited, an invented business that cuts and stitches school notebooks, faces the pairs in the ordinary run of a year. A second binding machine at Rs 9,00,000 sits opposite a service visit to the machine it already runs. A new roof on the shed sits opposite patching a leak in the roof it already has. A new stock-control software module at Rs 1,00,000 sits opposite the annual licence fee for the accounting software it renews every April. Each pair read left to right makes the difference audible: the left-hand item leaves something behind, the right-hand item keeps something going.

What is the test for capitalising a cost?

Three conditions decide it, and all three must hold at once. Does the business control the resource, meaning it can direct the use of the thing and stop others using it? Will the resource bring benefit beyond the current period? Can the cost of it be measured reliably, rather than estimated from a guess? Fail any one of the three and the whole amount is an expense of this year. To capitaliseTo record a cost as an asset on the balance sheet instead of charging it against this period's profit, so that it reaches the income statement later through depreciation or amortisation. a cost is to say that all three held.

The three conditions are joined by and, never by or. An obviously valuable and obviously long-lived thing can therefore still fail the test and be expensed in full. The classic case is the reputation a business builds with its customers. Reputation brings benefit for years. Reputation is worth real money to whoever might buy the business. But it cannot be measured reliably from any invoice, and a business does not control its customers, so not one rupee of it reaches the balance sheet. The exclusion is not an oversight. The third condition exists to keep unverifiable amounts out of an asset figure that a lender will one day rely on.

Run the test on the case in hand. Anjani Stationers bought a stock-control software module in year two for Rs 1,00,000. Control holds: the module is licensed to this business, sits on its servers, and nobody else can direct its use. Future benefit holds: it will run the stock records for four years, so the benefit plainly outlives the year in which it was bought. Reliable measurement holds: there is one supplier invoice for Rs 1,00,000, with no part of the amount estimated. All three hold, so the Rs 1,00,000 became an asset and reaches profit at Rs 25,000 a year for four years rather than as a single Rs 1,00,000 hit.

Which costs attach to an asset beyond its purchase price?

A machine costs more to have than the number on the supplier's invoice, and several of those extra costs belong inside the asset rather than inside this year's expenses. The ones that attach are the directly attributable costsCosts that would not have been incurred but for getting this particular asset into the condition and location needed for it to operate as intended. Freight, installation and testing are the standard examples. of bringing the asset to the condition and location needed for it to operate as intended: the purchase price itself, import duties and any tax the business cannot claim back, freight and handling to get it to the shed, preparing the site it will stand on, installing it, and testing it until it runs properly.

Every one of those costs attaches only until the asset is ready for its intended use. The cut-off is readiness rather than first use. A machine that sits idle for two months after it was ready still stopped absorbing costs on the day it became ready. Read that boundary carefully. It is the one people slide past. Ready for intended useThe point at which an asset is in the condition and location needed to operate the way management intended. Everything spent to reach that point attaches to the asset; everything after it is an expense. is a state of the machine, not an event in the calendar. Once the trial runs produce an acceptable notebook, the machine is ready, and the electricity bill from the following week is an operating cost even though it is spent on the very same machine.

The list of what never attaches is just as firm. Training the operators does not attach. The machine works whether or not anyone has been taught, and the training benefits the people, whom the business does not control. Promoting the new notebook format does not attach. General administration does not attach. Anything at all incurred after the readiness point does not attach, however closely related it feels. Relocating a machine because the shed was rearranged is one example. The same instinct decides which costs attach to inventory: the question is always about getting the thing ready, never about who spent the money or how large the amount was.

When is a spend a repair and when is it an improvement?

The repair against improvement line is the hardest in practice. Both kinds of spending are aimed at the same machine, often by the same engineer, sometimes in the same visit. A repair restores an asset to the condition it was already in, so it is an expense of the year. An improvement, sometimes called a bettermentA spend that raises an asset's capacity, output quality or remaining useful life above what it had before, as distinct from a repair, which only restores what was already there., raises the asset's capacity, improves the quality of what it produces, or extends how long it will go on working. An improvement is capitalised and added to the asset's carrying amount.

The comparison is always against the asset's own previous condition, never against the condition of a brand new asset and never against the size of the bill. This is where the household picture helps again. Replacing four worn tyres on a car restores the car to how it drove before the tyres wore; the car is not better than it was when the tyres were new. Fitting a new engine that was not there before, or converting a two-seater into a load carrier, changes what the vehicle can do. The tyres are a repair even at Rs 40,000, and the conversion is an improvement even at Rs 15,000.

Take three cases from the shed and reason each one out. Replacing worn cutting blades on the original binding machinery puts output back where it was before the blades dulled, so it is a repair and is expensed. Fitting an automatic sheet feeder that lets the same machine handle more sheets an hour than it ever handled raises capacity, so it is an improvement and is capitalised, then depreciated over the machine's remaining useful life. Repainting the shed walls restores appearance and protects the surface, but the shed does not hold more, last longer or produce better notebooks, so it is a repair. Notice that the third case is the one where the amount can be surprisingly large and the answer is still repair.

Where does capital expenditure appear in the three statements?

One spend, three statements, three different dates. The Rs 12,00,000 that Anjani Stationers put into machinery in year two consists of a second binding machine at Rs 9,00,000 on a six-year useful life and cutting equipment at Rs 3,00,000 on a four-year useful life. Both were in service from the opening day of the year, so each takes a full twelve months of depreciation, and that full-year assumption is a simplification. Both are assumed to have nil residual value at the end of their lives.

The cash left in full on the day it was paid, the asset appeared in full on the same day, and the income statement will only feel the Rs 12,00,000 spread across the six years that follow. The lag in the income statement is why a business can be cash poor and profitable in the same twelve months. The investing section of the cash flow statement carries the whole Rs 12,00,000 as an outflow of year two. The balance sheet carries Rs 12,00,000 of gross additions, against which Rs 2,25,000 of depreciation is charged in the first year, leaving Rs 9,75,000 of the addition still standing as an asset at the year end. The income statement carries Rs 2,25,000, being Rs 1,50,000 on the binding machine and Rs 75,000 on the cutting equipment.

The pattern of the charge is worth seeing rather than describing. Both assets are still being written down in years two through five, so each of those years carries Rs 2,25,000. By year six the cutting equipment has finished its four-year life and only the binding machine is left, so years six and seven carry Rs 1,50,000 each. The six charges add to Rs 12,00,000 exactly. The total is the arithmetic guarantee underneath all of this: capitalising never removes a cost from profit, it only decides which years' profit carries it.

Why does capitalising a cost raise this year's profit?

Because the decision moves an expense from this year to later years, and profit is measured a year at a time. Nothing else is going on. Anjani Stationers capitalised Rs 13,00,000 of spending in year two, being Rs 12,00,000 of machinery and Rs 1,00,000 of software. Had every rupee of it been treated as an expense as it was incurred, Rs 13,00,000 would have landed at once and the related depreciation and amortisation of Rs 2,50,000 would never have arisen. Earnings before interest and tax (EBIT) would have been Rs 41,50,000 plus Rs 2,50,000 less Rs 13,00,000. The sum is Rs 31,00,000, an EBIT margin of 11.5 per cent against the published 15.4 per cent.

In either world the bank was drained by identical amounts on identical dates, so profit differs by Rs 10,50,000 between the two treatments and cash differs by nothing at all. The cash flow statement rearranges rather than changing: operating cash flow would have been Rs 23,30,000 instead of Rs 36,30,000, investing would have been minus Rs 21,00,000 instead of minus Rs 34,00,000, financing is untouched at minus Rs 4,30,000, and the net movement in cash stays at minus Rs 2,00,000 in both. Free cash flow, being operating cash flow less the cash the business laid out on long-lived items, is Rs 23,30,000 in both worlds as well. Under the published treatment it is Rs 36,30,000 less Rs 13,00,000. Under the other, the Rs 13,00,000 has already gone out through the operating line, so it is Rs 23,30,000 less nothing further. Expensing the spend does not stop it being cash laid out on long-lived items. The pairing of a Rs 10,50,000 profit difference with a nil cash difference is the reason experienced readers reach for the cash flow statement when a profit figure surprises them.

One further figure is worth naming here so that it never gets mistaken for free cash flow. Operating cash flow plus the whole investing section comes to Rs 2,30,000, and that measure is not free cash flow. The investing section also carries the Rs 21,00,000 paid for the holding in Chitra Binding Works, and that holding bought a share of another business rather than capacity to make notebooks. Free cash flow subtracts what a business spent on the long-lived assets it runs itself, and buying another business is a different decision answered by different tools. The two measures therefore differ by exactly Rs 21,00,000, the whole of the purchase. Call the second one what it is, cash left after everything in the investing section, and keep the name free cash flow for the measure that leaves acquisitions out.

The panel below takes an illustrative Rs 4,00,000 that sits inside the published year two spend and moves it between fully expensed and fully capitalised, one rupee at a time. Four things are worth watching at once: earnings before interest and tax move, the asset base moves with them in the same direction, and both free cash flow and the net movement in cash refuse to move at all. The panel opens on the published treatment, so its first reading is the year as reported.

The readings that matter are these. At the published treatment, the whole Rs 4,00,000 capitalised on a six-year life, EBIT is Rs 41,50,000 and net property, plant and equipment closes at Rs 36,00,000. Drag the slider to nothing capitalised and EBIT falls to Rs 38,16,667 while net property, plant and equipment falls to Rs 32,66,667. Switch the useful life from six years to four with the whole amount still capitalised and EBIT falls to Rs 41,16,667 on an estimate alone, with no invoice anywhere changed. Across every one of those positions the net movement in cash reads minus Rs 2,00,000 and free cash flow reads Rs 23,30,000, the first because operating cash flow and the investing outflow move by exactly the same amount in opposite directions, and the second because the cash laid out on long-lived items is subtracted explicitly and so cannot hide inside either treatment. Both identities hold at every slider position rather than only at the ends.

What Are Capitalised Costs and Why Do They Matter?

Capitalised costs are simply the amounts a business has recorded as assets rather than as expenses. Capitalised costs matter to a reader for three reasons. Only the third is the one people rush to, so take them in order. The first is arithmetic: the decision moves profit between years without moving a rupee of cash, so two businesses trading identically can report different profits. The second is that the boundary genuinely involves judgement, most of all in the repair against improvement call and in the useful life applied afterwards. The third is that a business under pressure to show profit faces a temptation at exactly that boundary.

The third reason is a risk to check rather than an accusation to make. The same pattern is produced by a business that has genuinely bought a lot of long-lived equipment, and nothing in a set of published figures separates the two. This matters more than it sounds. An outside reader cannot see the engineer's report on whether the sheet feeder raised capacity. An outside reader can look in three specific places, form a question, and ask it. Anyone who converts the arithmetic into a claim that somebody arranged something has gone past what the evidence supports, and has usually also gone past what is fair to the people involved.

So here are the three checks, all of them available from outside the business. First, read the capitalisation policy in the notes: what the business says it capitalises, over what lives, and whether the wording changed from last year. A policy that changed without a stated reason is worth a question on its own. Second, put the additions column of the fixed asset scheduleThe note to the accounts that shows, for each class of asset, the opening gross amount, additions and disposals during the year, the depreciation charged, and the closing net amount. The schedule is where the movement behind the balance sheet figure is visible. beside the investing outflow in the cash flow statement and see whether they tell the same story. Third, ask whether additions are growing faster than the business is. Anjani Stationers' total assets grew 35.3 per cent while revenue grew 12.5 per cent. The gap has several ordinary answers, one of which is a business building capacity it has not filled yet.

What did Anjani Stationers actually spend in year two?

The published figures, exactly as they stand, carry the real weight of the case. Anjani Stationers Private Limited reported an investing cash outflow of Rs 34,00,000 in year two, and it is made of three things that answer different questions. Rs 12,00,000 went into property, plant and equipment, being the binding machine at Rs 9,00,000 and the cutting equipment at Rs 3,00,000. Rs 1,00,000 went into software, being the stock-control module. And Rs 21,00,000 went into a 70 per cent holding in Chitra Binding Works. A holding is not capital expenditure in the ordinary sense at all. The money bought a share of another business rather than a machine that makes notebooks.

The single largest line in the investing section bought no operating capacity at all. Anyone reading the Rs 34,00,000 as this business's capital spend has overstated it by Rs 21,00,000 and misdescribed what the money did. Capital expenditure intensity, meaning capital spend over revenue, is Rs 13,00,000 over Rs 2,70,00,000, which is 4.8 per cent counting both machines and software, or 4.4 per cent counting property, plant and equipment alone. The two are nearly half a percentage point apart on the same year, so anyone quoting either figure has to say which one they used.

Now the part that matters most. Anjani Stationers also took a warehouse on a four-year lease in year two and recognised a right-of-use asset of Rs 7,00,000, with a matching lease liability of Rs 7,00,000. The right-of-use asset entered the books without a single rupee of cash changing hands on the day it arrived, so it appears in no investing line anywhere. Gross property, plant and equipment went from Rs 45,00,000 to Rs 64,00,000, an increase of Rs 19,00,000. The cash flow statement records Rs 12,00,000 of payments for property, plant and equipment. Both numbers are correct. The two figures answer two different questions.

Who reads a capital spend figure, and what do they do with it?

Put the mechanism aside here. One fixed asset note is opened by three readers inside a single week, and none of the three has come to admire the arithmetic.

A lender reads capital spend to work out how much of next year's cash is already committed, an analyst reads it to separate a business building capacity from a business whose asset base is quietly ageing, and Vaidehi Rao, the finance controller, reads it to check that what the schedule says was bought matches what is actually bolted to the shed floor. Take them one at a time. The lender wants not the amount already spent but the amount still owed. A machine bought outright is paid for. A right-of-use asset of Rs 7,00,000 carries a lease liability that will take cash out over four years, Rs 2,00,000 of it within twelve months. A lender who read Rs 19,00,000 of additions as Rs 19,00,000 of settled spending would have missed a committed outflow sitting in plain view on the balance sheet.

The analyst's use is comparative. Capital spend against the depreciation and amortisation charge gives a rough sense of whether a business is replacing what it consumes, and Anjani Stationers reads 1.08 times on the cash basis and 1.67 times on the additions basis. Whether a spend is worthwhile is a different question with different tools, covered under capital budgeting, so neither number says whether the level of spend is right. The ratio does support a question about direction, and the analyst pairs it with the age of the base. Accumulated depreciation over gross block moved from 37.8 per cent to 43.8 per cent. The base therefore aged even though a brand new right-of-use asset with no accumulated depreciation behind it had just joined the gross block and would by itself have made the base look younger.

Vaidehi Rao's use is the most concrete of the three, and it is the one that catches errors. She can walk the shed with the schedule in hand. A capitalised amount with nothing physical to point at is a real problem, and so is a machine standing in the corner with no line against it. Neither is visible from outside, and an outside reader therefore has to work with policy, disclosure and ratios instead. Capital spend shows what a business bought and when it paid; it never shows whether the purchase was a good idea, and no figure of that kind can be pushed that far.

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