Goodwill Impairment: The Deal's Verdict, Delivered Years Later

The Acquisition Is Re-Priced Every Year, Whether Anyone Wants It Re-Priced or Not

A transaction closes and the valuation work appears to end. The price was agreed, the bridge settled the consideration, the funds moved. What actually happens is that the valuation moves from a negotiation into a reporting obligation. The premium paid over the fair value of what was acquired sits on the balance sheet as goodwill, and from the first reporting date onward it must be tested — annually, and whenever there is an indication of impairment — against whether the acquired business can still recover it. The assumptions that justified the price become assumptions that have to keep holding, in public, in front of auditors, for as long as the goodwill remains.

This is the most under-appreciated consequence of an acquisition, and it is the reason valuation discipline at the point of purchase is not merely a negotiating posture. A price that required aggressive growth to justify has not been made safe by the counterparty accepting it. It has been converted into a carrying value that requires the same aggressive growth to survive an annual test administered by people who were not in the room. The verdict is delivered years later, by a mechanism the acquirer does not control, in a document everybody reads.

This article sets out how that verdict is arrived at: where goodwill comes from and why it cannot be tested on its own, how the cash-generating unit is defined and why the level chosen frequently decides the outcome before any cash flow is forecast, the two permitted measures of recoverable amount, the three assumptions that carry the weight of the calculation, and why headroom that looks comfortable can be eliminated by a movement in the discount rate that nobody would describe as dramatic. It completes a series: the enterprise-to-equity bridge decides what a seller receives, the quality-of-earnings review decides what number the multiple was applied to, and the impairment test decides, some years afterwards, whether the resulting price was supportable.

Goodwill Is What Is Left When Everything Identifiable Has Been Recognised

Goodwill arises on a business combination as consideration transferred less the fair value of the identifiable net assets acquired. Every word of that is doing work. Consideration includes contingent elements measured at fair value, not only cash paid at completion. Identifiable net assets are measured at fair value on acquisition rather than carried across at the seller's book values, and the exercise routinely recognises assets the seller never had on its own balance sheet — customer relationships, brands, technology, order backlogs — because internally generated intangibles are generally not recognised by the entity that generates them but are recognised by an acquirer who has paid for them.

The purchase price allocation is therefore consequential rather than administrative, and it is a valuation exercise in its own right. The more of the price that is attributed to identifiable intangible assets, the less remains as goodwill — but those intangibles are typically amortised over finite useful lives, so the allocation trades a residual balance that is tested annually for a charge that runs through profit every year. An allocation that pushes value into goodwill defers cost and concentrates risk; an allocation that recognises more identifiable intangibles accepts a predictable drag in exchange for a smaller residual. Where our valuation team performs a purchase price allocation, both consequences are set out for the board before the allocation is finalised, because the decision is frequently presented as technical when its effect on reported earnings for the following decade is not.

What makes goodwill distinctive is that it cannot be tested on its own. It generates no cash flows independently of other assets; it is, by construction, the part of the price that could not be attributed to anything identifiable. So the standards require it to be allocated to the cash-generating units, or groups of units, expected to benefit from the synergies of the combination — and it is the recoverable amount of those units, not of the goodwill, that is tested. That allocation decision, made once at acquisition and revisited when the business is reorganised, has more influence over whether an impairment is ever recognised than most of the forecasting that follows it.

The Level the Unit Is Drawn At Frequently Decides the Answer

A cash-generating unit is the smallest identifiable group of assets that generates cash inflows largely independent of the cash inflows from other assets. The definition is deliberately restrictive at the level of ordinary assets, but goodwill is allocated to units — or to groups of units — that reflect the level at which it is monitored internally, subject to a ceiling at the operating segment level. Between the smallest independent unit and that ceiling there is judgement, and the judgement is decisive.

The mechanism is simple to state. Test a small unit, and the acquired business defends its own carrying value on its own cash flows, with no help. Test a large group of units, and the acquired business is pooled with everything else in the group, so headroom generated by unrelated successful operations can absorb the shortfall of the acquisition entirely. The same acquisition, the same performance, the same forecasts, can produce an impairment at one level and none at the other. This is not a loophole; it is a consequence of testing goodwill in the unit that benefits from the synergies. But it means the boundary is the first thing to examine on any impairment question, and it is examined too rarely. In the mandates our advisers support, the CGU boundary is reconciled back to the allocation made at acquisition before any cash flow forecast is reviewed, because a boundary that has moved since acquisition can mask an outcome the original allocation would have exposed.

Boundaries do legitimately move. Businesses are reorganised, reporting structures change, and a unit that was independent at acquisition may genuinely be integrated three years later. Where that happens, goodwill is reallocated between units, usually on a relative value basis. What deserves attention is the direction of travel: a reorganisation that pools an underperforming acquisition into a larger, healthier group has changed the test in the acquisition's favour, and whether the reorganisation reflects how the business is now actually managed and monitored is a question of fact that can be evidenced from internal reporting rather than debated in the abstract. Where we act on an impairment question, our financial due diligence team asks for the internal management pack before the impairment memorandum, because the pack shows how the business is monitored in practice and the memorandum shows how it has been described for the purpose of the test.

A second consequence follows from the pooling. Where an acquisition's goodwill sits inside a large group of units, the internally generated value of the rest of that group is protecting it. The standards do not permit internally generated goodwill to be recognised as an asset, but nothing prevents it from providing the headroom that keeps acquired goodwill on the balance sheet. That is a structural feature of the model and it is worth understanding for what it is: an absence of impairment is evidence that a unit as a whole recovers its carrying amount, and it is not, on its own, evidence that a particular acquisition performed as underwritten.

Value in Use and Fair Value Less Costs of Disposal Answer Different Questions

Recoverable amount is the higher of a unit's value in use and its fair value less costs of disposal. An impairment arises where the carrying amount of the unit, including allocated goodwill, exceeds that recoverable amount. Because the test takes the higher of the two, an entity needs only one of them to clear the carrying value — which means the choice of which to prepare, and how much effort to put into the second, is itself a decision with consequences.

Value in use is an entity-specific measure: the present value of the future cash flows expected to be derived from the unit in its current condition, discounted at a rate reflecting current market assessments of the time value of money and the risks specific to the asset. Its defining constraint is that it is measured on the asset as it stands. Cash flows from a future restructuring the entity is not yet committed to, or from enhancing the unit's performance beyond its current condition, are excluded. That constraint is where most of the argument sits in practice, because management forecasts are naturally built on plans, and plans naturally include improvement. Separating the cash flows of the asset in its current condition from the cash flows of the asset as management intends it to become is the substantive work in a value-in-use calculation, and it is where our valuation team spends its review time rather than on the arithmetic of the discounting.

Fair value less costs of disposal is a market measure: what a market participant would pay for the unit, less the costs of selling it. It is not entity-specific, so it can include benefits a market participant would obtain that the current owner cannot — but it must be supportable by market evidence rather than by internal expectation. Where recent transactions in comparable businesses exist, or where the unit is closely comparable to listed peers, this measure can be more robust than a discounted forecast. Where such evidence is thin, a calculation labelled fair value that is in substance a discounted cash flow with a different discount rate is unlikely to withstand scrutiny, and it will be identified.

The practical point is that the two measures fail in different conditions. Value in use falls when the entity's own forecasts deteriorate. Fair value less costs of disposal falls when market pricing for that kind of business deteriorates, which can happen while internal forecasts are entirely unchanged. A business whose sector has de-rated may find its market-based measure has moved sharply below a value-in-use calculation that nobody has touched — and the disclosure of which measure was used, and why, is one of the more informative lines in the notes.

Three Assumptions Carry the Weight, and One of Them Moves on Its Own

A value-in-use calculation has many inputs and three that matter. The forecast cash flows over the explicit period, the terminal growth rate applied beyond it, and the discount rate. Everything else is arithmetic. The distinctive feature of the three is that they behave very differently: two are prepared by management and defended by management, and the third is substantially set by markets and can move materially between one test and the next without anybody inside the business doing anything.

The forecast cash flows attract the most attention and are the most obviously contestable, but they also carry an accountability trail that few other estimates have. Prior-year forecasts exist. Outturn exists. A calculation that assumes recovery to a level the unit has forecast in each of the preceding three years and not reached is asserting something that the entity's own records contradict, and the comparison is straightforward to run. In the mandates our advisers support, the first exhibit prepared is not a new forecast but a comparison of previous forecasts against outturn, because credibility established that way is worth more in the audit discussion than any amount of additional modelling.

The terminal growth rate is the quietest and frequently the most powerful. Because the terminal value typically dominates the present value in any long-lived unit, a rate that looks conservative in isolation can carry an implausible implication — that the unit grows in perpetuity at a rate approaching or exceeding the long-run growth of the economy it operates in, which implies it eventually becomes an ever-larger share of that economy. The discipline is to state the rate alongside the long-run inflation and growth expectations for the relevant market and to justify any excess explicitly, rather than to select a figure that produces the required answer and disclose it without reference.

The discount rate is the one that moves on its own, and it is the reason a test that passed comfortably can fail without any deterioration in the business. It reflects current market assessments of the time value of money and the risks specific to the unit — which means changes in risk-free rates, in equity risk premia and in the unit's risk profile all feed into it. Because the terminal value is the largest component of most calculations and the terminal value is inversely sensitive to the gap between the discount rate and the terminal growth rate, a modest increase in the rate compresses that gap and reduces the recoverable amount disproportionately. Where the two are close together, the calculation is unstable by construction: small movements in either produce large movements in value. Our valuation team runs the discount rate and terminal growth sensitivity as a grid rather than as single-point movements, because it is the combination that produces the surprise, and the combination is what the disclosure requirements are directed at.

The consequence for a board is worth stating plainly. Headroom is not a buffer of a fixed size. It is the output of a calculation whose most volatile input is set outside the business, and the disclosure of key assumptions and their sensitivities exists precisely so that a reader can see how much of the headroom is a judgement about the business and how much is a judgement about the cost of capital. A unit reporting substantial headroom on a discount rate materially below what a market participant would apply is reporting the assumption, not the buffer.

Move the Assumptions and Watch the Headroom Disappear

The instrument below computes a recoverable amount from the three load-bearing assumptions and sets it against a carrying value, exactly as a value-in-use test does. It is seeded with an illustrative unit that passes its test with room to spare. The method is stated in full above the instrument, and the arithmetic is ordinary discounting — there is nothing proprietary in it.

The point of the exercise is not the headline figure but the fragility. Move the discount rate by a quarter of a point at a time and watch what happens to a buffer that looked comfortable. Then narrow the gap between the discount rate and the terminal growth rate and observe that the same movement produces a much larger effect. That instability is a property of the model, it is present in every long-lived unit tested this way, and it is why the sensitivity disclosures are read before the headroom figure by anyone who has done this work.

Non-Cash Does Not Mean Without Consequence

An impairment charge moves no cash. It is frequently introduced on that basis, and the description is accurate as far as it goes. What it omits is that the charge is a disclosure, and disclosures are read by parties whose decisions do move cash.

Lenders read it first. Covenant packages are typically constructed to look through non-cash items, so a well-drafted facility will not be tripped by the charge itself. But the charge is a statement by the borrower's own directors, supported by its auditors, that a unit is not expected to recover its carrying amount — and that statement informs how the next refinancing is priced, how covenant headroom is modelled going forward, and how readily an amendment is granted when one is needed. Analysts read it as information about the forecasts, not about the accounting. A charge tells them the assumptions underpinning a previously supported carrying value have been revised downward, and it invites the question of what else in the guidance rests on the same assumptions. In the mandates our transaction advisory practice runs, the disclosure is drafted alongside the calculation rather than after it, because a number that is defensible and a sentence that explains it are two separate pieces of work and only one of them is read by the market.

Boards read it, or should, as the closing of a loop. The impairment test is the only mechanism in financial reporting that systematically compares an acquisition's original thesis to its outcome, and it does so with the same rigour whether the board wishes to revisit the decision or not. That is uncomfortable, and the discomfort is the point: a governance process that reviews acquisitions only when they succeed learns nothing. Where our transaction advisory practice is asked to support a post-completion review, the exercise is deliberately structured around the assumptions relied on at acquisition rather than around current performance, because the useful question is not whether the business is doing well but whether it is doing what was underwritten.

The most substantial cost of an impairment is generally the one that arrived before the charge did. By the time a test fails, the operating shortfall it reflects has usually been evident internally for some time, and the value was lost in the period the business underperformed rather than on the day the accounting caught up. This is why an impairment is better understood as a verdict than as an event. It records a conclusion that the facts reached earlier.

A Price Is Agreed Once and Defended Every Year After

The discipline this argues for is not caution at the point of acquisition for its own sake. Businesses are worth paying for, premiums are frequently justified, and a board that never pays above the fair value of identifiable net assets will acquire very little. What the impairment mechanism does is impose a consistency requirement across time: the assumptions used to justify the price are the assumptions that will be tested, by a process the acquirer does not control, in a document its lenders and analysts will read.

The practical consequence is that the assumptions relied on at acquisition should be recorded at acquisition, in a form specific enough to be tested afterwards. A growth rate, a margin trajectory, a synergy delivery schedule with dates against it — written down while they are still a thesis rather than reconstructed later from a model that has been updated a dozen times. Boards that do this find the annual test straightforward, because the comparison they are being asked to make is one they have already been making. Boards that do not find themselves defending a carrying value against a forecast whose relationship to the original case nobody can now establish. Our valuation team asks for the acquisition case as the first document on any impairment engagement, and where it cannot be produced in its original form, that is itself the most useful finding of the review. These are observations drawn from work delivered by a network of 80+ specialist consultants across 22 countries, which is why they are offered as a pattern rather than as one adviser's recollection.

This completes the series. The quality-of-earnings review establishes what number the multiple should be applied to, the enterprise-to-equity bridge determines what happens to the value that multiple produces, and the impairment test re-marks the whole of it against what actually occurred. The three are the same valuation question asked at three points in time, and the answers are supposed to agree. Where they do not, the disagreement is information — and it is available to any board willing to run the test on itself before the auditors run it.