Depreciation and Amortisation
5 steps · one page
In short
A company buys a machine for a million dollars and it will produce goods for a decade. Charging the whole cost against the year it was bought would make that year look terrible and the next nine look artificially good — and neither picture would describe the business.
Canonical data. Figures tie to Wexford Instruments (USD millions). The policy comparison uses a standalone illustrative asset.
Depreciation is the answer: spreading the cost of a long-lived asset across the periods it helps earn revenue. It is the matching principle applied to things that last, and it is the largest non-cash item in most companies' accounts.
Three things depreciation is not
Each of these misconceptions is common enough to cause real misreadings. It is not a cash outflow. The cash left when the asset was purchased, which was an investing activity in an earlier year. The annual charge moves no money at all, which is why it is added straight back at the top of the cash-flow statement and why Wexford's operating cash flow exceeds its profit. It is not a measure of what the asset is worth. Book value is original cost less accumulated charges — an arithmetic result of a policy decision, not an appraisal. A fully depreciated machine carried at zero can be running productively and would fetch a real price; a recently bought asset carried near cost can be worthless. And accumulated depreciation is not a fund. No money has been set aside to replace anything. It is a contra-account recording how much of the original cost has been charged, and a company that has depreciated its assets fully has not thereby saved for their replacement.
The first decision: capitalise or expense
Before any depreciation policy applies, someone must decide whether a payment is an asset at all. Buying a machine is capital expenditure — it becomes an asset and is depreciated. Repairing that machine is an expense — it hits profit immediately. The boundary is defined by whether the spending extends the asset's life or capability beyond its original condition, and in practice the boundary requires judgement: major overhauls, software development, and improvements to leased property all sit near the line, and companies must set and disclose policies about where they draw it.
The consequence is worth understanding precisely, because it is the largest single lever in this article. Capitalising a cost rather than expensing it raises reported profit this year, raises reported assets, and changes nothing about cash. On the illustration: spending 30 with a five-year life reduces profit by 30 if expensed, or by 6 if capitalised — a 24 difference in year-one profit, with assets 24 higher in the capitalised version and the bank balance identical either way. Neither treatment is dishonest and the correct one depends on the facts, but a reader comparing two companies with different capitalisation policies is comparing two conventions.
Method, life, and residual value
Three policy choices determine the annual charge, and all three are estimates. Method. Straight-line spreads cost evenly and is by far the most common. Accelerated methods such as declining balance charge more in early years, on the reasoning that many assets deliver more value when new. Units of production ties the charge to actual usage, which suits assets whose life is measured in output rather than time. Useful life. How long the asset will serve — an estimate, and one that varies legitimately between companies operating identical equipment in different ways. Residual value. What it will be worth at the end, which reduces the amount to be charged.
The total charged over an asset's life is the same under every method — only the timing differs, and the timing differences are large. On a 100 asset: straight-line over eight years charges 12.5 in year one; straight-line over five years charges 20.0; declining balance over five years charges 40.0. That is a spread of 27.5 in the first year — 28% of the asset's entire cost — from three defensible policies applied to one identical machine. Later years reverse the pattern, which is the point: accelerated methods do not reduce total profit, they move it. (A declining-balance schedule reaches full cost only by switching to straight-line or adjusting the final year, which is standard practice and is stated here as a principle rather than tabulated.)
Amortisation is the same mechanism applied to intangible assets with finite lives — purchased software, patents, customer relationships acquired in a transaction, licences. Straight-line dominates because a usage pattern is rarely observable. Two boundaries matter. Intangibles with indefinite lives, and goodwill, are not amortised at all under current standards; they are tested for impairment instead, which the next article covers. And internally generated intangibles are largely not capitalised: a company that builds a brand through decades of advertising carries nothing on its balance sheet for it, while a company that buys that same brand records it as an asset and amortises it. Two identical economic positions, two entirely different balance sheets — one of the widest gaps between accounting and economic reality that a reader will encounter.
Worked example
Worked example: Wexford Instruments (canonical figures, USD millions). Wexford's income statement shows D&A of 50.0, which is two different things combined: depreciation of 42.0 on physical assets and amortisation of 8.0 on finite-lived intangibles. Trace each to the balance sheet. PP&E opened at 400.0, capital expenditure added 78.0, depreciation removed 42.0, and the closing balance is 436.0 — exactly as reported. Intangibles opened at 68.0, amortisation removed 8.0, closing at 60.0. Goodwill sat unchanged at 180.0, because goodwill is not amortised. Now the ratio. Capital expenditure of 78.0 against D&A of 50.0 is 1.56×; against depreciation alone it is 1.86×. A company spending well above its depreciation charge is adding capacity rather than merely replacing it — or is replacing assets that cost more than the ones they succeed. The statements cannot distinguish those two, and this article does not pretend otherwise; what the ratio establishes is that the question is worth asking. And the caution. Wexford's 42.0 depreciation charge reflects its chosen methods, lives, and residual values. A competitor with identical machinery and longer assumed lives would report a smaller charge, higher profit, and higher carrying values — with no difference whatsoever in the physical assets or the cash. (Canonical figures; ties verified against the Wexford page.)
Frequently asked
8 questions
What is depreciation?
Spreading the cost of a long-lived asset across the periods it helps earn revenue — the matching principle applied to things that last. It's typically the largest non-cash charge in a set of accounts.
Is depreciation a cash cost?
No. The cash left when the asset was bought, in an earlier year, as an investing activity. The annual charge moves no money, which is why it's added back at the top of the cash-flow statement.
Does book value tell me what an asset is worth?
No — book value is original cost less accumulated charges, an arithmetic result of a policy choice rather than an appraisal. A fully depreciated machine carried at zero may be running productively and would fetch a real price.
Is accumulated depreciation money set aside for replacement?
No. It's a contra-account recording how much of the original cost has been charged. No cash has been reserved, and a company that has fully depreciated its assets has not thereby saved to replace them.
What's the difference between capitalising and expensing?
Expensing hits profit immediately; capitalising creates an asset charged over future years. On the illustration here, spending 30 with a five-year life costs 30 of year-one profit if expensed and 6 if capitalised — a 24 difference, with assets 24 higher and cash identical. The correct treatment depends on whether the spending extends the asset's life or capability.
Which depreciation method is right?
None universally. Straight-line spreads cost evenly, accelerated methods charge more early, units-of-production ties the charge to usage. Total charged over the asset's life is identical under all of them — only timing differs, and on a 100 asset the year-one charge ranges from 12.5 to 40.0 across three defensible policies.
Why isn't goodwill depreciated?
Because it has no determinable finite life. Under current standards, goodwill and indefinite-lived intangibles are tested for impairment rather than amortised — which the next article covers.
Why does a bought brand appear on the balance sheet when a built one doesn't?
Because internally generated intangibles are largely not capitalised while acquired ones are. A company that builds a brand over decades carries nothing for it; a company that buys the same brand records an asset and amortises it. Two identical economic positions, two entirely different balance sheets.
References
- IFRS Foundation — IAS 16 Property, Plant and Equipment (depreciation methods, useful life, residual value) —
- IFRS Foundation — IAS 38 Intangible Assets (amortisation of finite-lived intangibles; internally generated intangibles largely not recognised) —
- SEC — Beginners' Guide to Financial Statements —
Educational and informational only — not investment advice, a recommendation, or an offer to buy or sell any security. Investing involves risk, including the possible loss of principal. Worked examples use fictional companies and figures.