Full explanation
Overview
Some of the most valuable things a company owns, you can't touch. And some of them never appear on its balance sheet at all.
Coca-Cola's brand is worth tens of billions — and it is nowhere on Coca-Cola's balance sheet. That's not an oversight. That's IAS 38, working exactly as designed.
Three tests to even be one
An intangible asset is three things at once: identifiable, non-monetary, and without physical substance. Identifiable matters most — it means the asset is either separable, so you could sell it on its own, or it arises from legal or contractual rights. Then the usual two recognition tests: probable future economic benefits, and cost that can be measured reliably. Software, patents, licences, trademarks, customer contracts. Not the building they sit in — that's IAS 16.
Research vs Development
Now the divide that runs through the whole standard: research versus development. Research is the phase where you're still finding out whether something is even possible. IAS 38 is blunt about it — research costs are ALWAYS expensed. Never capitalised. Development is different: this is where you're building something you already know works. And development costs CAN be capitalised — but only if you clear six hurdles, all of them, at the same time.
Six conditions, and you need every single one. Miss one, and the spending goes straight to profit or loss.
The six criteria for capitalising development
Here they are. One: technical feasibility — you can actually finish it. Two: intention to complete it. Three: ability to use or sell it. Four: it will generate probable future economic benefits — and for an intangible you intend to use internally, you must show its usefulness. Five: adequate technical, financial and other resources to finish the job. And six: you can measure the expenditure reliably. All six. Simultaneously. From the date the last one is met — and not a day earlier.
One project, three phases
Let's put real numbers on it. Meridian Limited spends seven hundred thousand euros developing a new product. January to March is pure research — one hundred and twenty thousand euros, exploring whether it's even viable. April and May, they've moved into development, but the technical feasibility isn't yet proven — another one hundred thousand. Then, on the first of June, the last condition falls into place. From June to December they spend four hundred and eighty thousand.
€220,000 expensed. €480,000 capitalised.
So what goes where? The research phase — one hundred and twenty thousand — expensed. The development spend BEFORE the criteria were met — one hundred thousand — also expensed. Two hundred and twenty thousand euros straight through profit or loss. Only the four hundred and eighty thousand spent after all six conditions were satisfied gets capitalised. And here's the trap: you can never go back and capitalise the earlier spend retrospectively. Once it's expensed, it stays expensed.
That's the spending you CAN capitalise. Now the category you never can — no matter how valuable it is.
Never recognised. No matter the value.
IAS 38 draws an absolute line. Internally generated goodwill: never recognised. Internally generated brands, mastheads, publishing titles, customer lists, and items similar in substance: never recognised. It doesn't matter that everyone knows the brand is worth a fortune. The standard's reasoning is that you cannot separate the cost of building that brand from the cost of running the business as a whole — so any number you put on it would be invented. That's why Coca-Cola's brand isn't there.
Buy the same brand — everything changes
But buy the same brand from someone else, and everything changes. A purchased intangible has a price — an actual transaction, negotiated at arm's length. So it goes on the balance sheet at cost. Meridian buys a ten-year software licence for three hundred thousand euros: recognised, and amortised at thirty thousand a year. Same type of asset. Completely different answer, depending on whether you built it or bought it. And intangibles acquired in a business combination are recognised at fair value — even ones the seller could never have recognised themselves.
Once it's on the balance sheet, one question decides how it's measured from then on: does it have a finite life, or not?
Finite life, or indefinite ?
A finite-life intangible is amortised over its useful life — Meridian's capitalised development costs, four hundred and eighty thousand over five years, is ninety-six thousand a year, starting when the asset is available for use, not when you finish paying for it. An indefinite-life intangible is different: indefinite does NOT mean infinite. It means there's no foreseeable limit on how long it will generate cash. You don't amortise it at all — instead you test it for impairment every single year, whether or not there's any sign of a problem. And you reassess that indefinite judgement annually.
Revaluation exists — but barely
Two more measurement points. There is a revaluation model for intangibles — but it's almost theoretical, because it requires an active market for that exact asset, and active markets rarely exist for something as unique as a brand or a patent. Most intangibles sit at cost less amortisation, forever. And amortisation must reflect the pattern of consumption — a 2014 amendment specifically restricted revenue-based amortisation, because how fast you earn money from an asset is not the same as how fast you consume it.
Cloud software — usually not an asset
Now the question this standard gets asked more than any other — cloud software. When you pay for software as a service, you are not buying an intangible asset. The Interpretations Committee settled that in 2019: a contract that only gives you the right to access the supplier's software is a service contract, so you expense it across the term. Then in 2021 they went further, on the money companies spend configuring and customising that cloud software — often millions of euros. Those activities don't create a resource you control separately from the software itself. So unless what you build genuinely meets the intangible definition on its own, it's an expense. And websites have their own interpretation, SIC 32: the planning phase is expensed, the development phase can be capitalised — on exactly the six criteria you just saw.
One more thing worth knowing — because if you also read US accounts, this is where the two worlds split.
The same project, two different profits
Under US GAAP, research AND development are both expensed as incurred — with a narrow exception for certain software costs. Under IFRS, development is capitalised once those six criteria are met. So take two identical companies, running identical projects — the IFRS one shows an asset and higher profit; the US one shows neither. Nothing about the underlying business differs. Only the rulebook does. It's one of the largest remaining differences between the two frameworks.
Buy it · value it · test it · sell it
And IAS 38 sits inside a family. Something physical you use? IAS 16 — Property, Plant and Equipment. No physical substance? IAS 38, here. Held to earn rent or for capital appreciation? IAS 40, Investment Property — that's next. Held to sell in the ordinary course of business? IAS 2, Inventories. And whatever category it lands in, if its value falls too far, IAS 36 Impairment tests it. Buy it, value it, test it, sell it — one asset lifecycle, five standards.
IFRS 18 — where amortisation lands
And one change that's already in motion. IFRS 18 replaces IAS 1 for periods beginning January 2027, and it reorganises the income statement into three categories — operating, investing and financing. Amortisation of intangibles sits in operating. But here's the part that will bite: IFRS 18 pulls management-defined performance measures into the audited notes. So a company reporting adjusted earnings that exclude amortisation of acquired intangibles — and a great many do — must now reconcile that figure to an IFRS subtotal, where the auditor and the reader can both see it. And because it applies retrospectively, the 2026 comparatives are being restated right now.
Where the balance sheet is knowingly incomplete
Why does an investor care? Because this is one of the few places where the balance sheet is knowingly incomplete. A pharmaceutical company with a huge research pipeline, or a consumer business built on a famous brand, can be worth vastly more than its net assets suggest — and IAS 38 is the reason. So read the intangibles note carefully: how much was capitalised this year, how much went straight to expense, and whether any indefinite-life assets are quietly sitting there, never being amortised.
Every euro traceable to a rule
On Meridian's own accounts: intangible assets, four hundred and eighty thousand of capitalised development, less ninety-six thousand of amortisation, carried at three hundred and eighty-four thousand. Plus the software licence, three hundred thousand less thirty thousand, at two hundred and seventy thousand. And in profit or loss, two hundred and twenty thousand of research and pre-criteria development, expensed. Every euro traceable to a rule.