Depreciation: why an asset's cost is spread over years

Buy a laptop for eighty thousand rupees and the money leaves the account immediately. It does not follow that eighty thousand rupees of cost belongs to that month.

The reasoning is about matching. A month's profit is meant to show what the month's activity was worth: revenue earned less the cost of earning it. A laptop used for three or four years contributes to earning revenue across all of them. Charging its entire cost to the month of purchase would report a terrible month followed by several artificially good ones, none of which describes the business accurately.

So the accounting treats the purchase as a transformation rather than a consumption. Cash became an asset of equivalent value, and no cost has yet been incurred. The asset then sits on the balance sheet and a portion of its value is charged to profit each period as it is used up. That charge is depreciation.

Which gives the two facts worth holding onto. Depreciation is the recognition of a cost already paid, spread across the periods that benefit from it. And it moves no money at all: the cash left when the asset was bought, and the depreciation entry each year is a bookkeeping recognition with no payment attached.

At purchase, the entry debits a fixed asset account and credits bank. Nothing reaches the profit-and-loss statement. This is the step most often got wrong in informal books, where the whole amount is debited to an expense head, which overstates costs now and understates them for the rest of the asset's life.

Each period, the entry debits depreciation expense and credits accumulated depreciation. The expense reduces profit. Accumulated depreciation is a balance that sits against the asset, reducing it, and it grows each year. The asset's original cost stays visible and the accumulated total against it grows, so the balance sheet shows both what was paid and how much has been written off. The difference between them is the carrying value, sometimes called the book value.

When the asset is eventually sold or scrapped, both the cost and its accumulated depreciation are removed, and the difference between the carrying value and whatever was received is a gain or a loss on disposal. Disposal is the step that gets skipped, and skipping it is why asset registers drift: a machine physically scrapped but never removed from the books keeps generating a depreciation charge every year, with arithmetic that is entirely correct and a result that is entirely wrong.

One consequence worth noting: carrying value is not market value and was never intended to be. It is unrecovered cost. An asset can be fully written down and still working, and it can have a carrying value well above what anyone would pay for it.

Two methods dominate. Straight line charges an equal amount each period, computed from the cost, the expected life and any expected residual value. Written down value charges a fixed percentage of the remaining carrying value each period, so the charge is largest in the first year and falls thereafter.

The difference is a judgement about how the asset is consumed. Straight line suits assets that deliver steady service across their life. The reducing method suits assets that lose most of their usefulness or value early, which is a reasonable description of vehicles and computing equipment.

Here is the part that surprises owners: a business commonly has two depreciation figures for the same asset, and both are correct. The figure in the accounts is determined by accounting requirements and reflects the useful life the business genuinely expects. The figure allowed as a deduction is determined by tax law, which prescribes its own rates and its own method and does not care what life you expected.

Under Indian income-tax law, depreciation is generally computed on the written down value of a block of assets rather than asset by asset, with rates set out in Appendix I to the Income-tax Rules. As at the time of writing, that appendix places general plant and machinery at fifteen per cent, furniture and fittings at ten per cent, and non-residential buildings at ten per cent, with separate higher-rate entries including computers. There are also rules for assets used part of a year. These rates are amended by notification, so the current appendix is the only source to rely on.

The whole mechanism depends on a prior decision: is this purchase an asset or an expense? Get that wrong and no amount of correct depreciation arithmetic helps.

The general test is whether the item will be used across more than one accounting period, rather than consumed in the current one, and whether it is held for use in the business rather than for resale. A delivery van is an asset. Fuel for it is an expense. A shelving system is an asset. The stock on the shelves is neither, it is inventory, which is a current asset accounted for differently. Stock held for resale is never a fixed asset regardless of how long it sits there.

Two boundary cases cause most of the difficulty. Repairs versus improvements: fixing a machine so it keeps working as before is generally a repair and an expense, while an upgrade extending its life or capacity is generally capital, and the distinction is genuinely contested. And small-value items: nobody maintains an asset register for a calculator, and where that line sits is a policy matter for the accounts and a question of specific provisions for tax.

The honest position is that classifying a purchase is a professional question more often than it looks. Costs incidental to acquisition, such as freight and installation, generally form part of the asset's cost rather than being separate expenses, which is another place informal books diverge from correct treatment. When in doubt about an item of any size, ask before entering it, because the entry sets a treatment that runs for years.

Depreciation has three practical consequences that reach beyond bookkeeping.

It changes reported profit, and it is one of the estimates inside a profit figure that people treat as a measurement. The charge depends on an assumed useful life and an assumed residual value, both chosen by someone. A longer assumed life produces a lower annual charge and a higher reported profit for exactly the same asset and the same cash. That is not manipulation, it is the nature of accrual accounting, and it is a reason to know what assumptions sit behind a profit figure before relying on it.

It separates profit from cash decisively. Depreciation reduces profit without moving money, while the purchase moved money without reducing profit. This is why a business can report a loss and have a healthy bank balance, or the reverse, and why reconciling profit to cash requires adding depreciation back and subtracting capital spending.

And it means the tax deduction for an asset is spread rather than immediate, which affects the cash cost of buying it. The tax benefit arrives across years, on the tax rules' schedule rather than yours.

What depreciation is not is a fund. There is no money accumulating anywhere to replace the asset. Accumulated depreciation is a record of cost already written off, and if replacement is going to need cash, that cash has to be planned for separately. Businesses that treat the depreciation charge as provision for renewal find at replacement time that the provision was an accounting entry all along.

An asset register that is properly maintained holds the cost of each asset, its purchase date, its accumulated depreciation, its carrying value and its disposal if it has gone. From that, the periodic charge is arithmetic, and arithmetic is reliable.

What the register cannot do is know about the physical world. It cannot know that a machine was scrapped, sold informally, stolen or lent to another site and never returned. Every one of those leaves the register unchanged and the depreciation charge continuing, correctly computed on an asset that does not exist. The only thing that detects it is a physical verification: somebody walking around with the register and confirming that each item is present. Nothing else does, and no system can substitute for it.

It also cannot tell you whether the useful life assumed is still realistic, or whether an asset has fallen in value faster than the schedule assumes, which is an impairment question and a judgement.

So the boundary for any accounting tool here is sharp. Software can hold the register, compute both the accounting charge and the tax figure on their different bases, apply the correct rate to the correct block, handle part-year rules, post the entries, and flag an asset past its assumed life still carrying a value. Those are the tedious and error-prone parts. What it cannot do is decide whether a purchase was capital, know an asset's real remaining life, or tell you the van was sold last year. Those need a person, and the last needs one with a clipboard.