Quality of earnings

Profit is computed under rules that require judgement. Most companies exercise that judgement honestly and some exercise it optimistically, and the accounts give you enough to tell which.

The core test

cash conversion = cash from operations ÷ net profit

Over five years. A business where this sits around 1 or above is converting what it reports into money. One where it sits at 0.5 is reporting twice the profit it collects.

One year proves nothing — a growing company legitimately absorbs cash into working capital (chapter 11). Five years is where the pattern shows, because working capital cannot absorb cash indefinitely without either reversing or being written off.

This is the single most useful quality test and it takes two numbers a year.

Where judgement enters

Five places, in rough order of how much they can move profit.

Revenue recognition. When a sale counts. For long contracts, percentage-of-completion requires estimating how complete the work is — and the estimate sets the revenue. The auditor's key audit matters (chapter 7) flag this when it is significant.

Provisions for bad debts. Management estimates how much of its receivables will not be collected. Under-providing raises profit now and produces a write-off later. Compare the provision with receivable days from chapter 11: rising days and a flat provision rate is a combination worth questioning.

Inventory valuation. Inventory is carried at the lower of cost and net realisable value, and someone decides what is realisable. Slow-moving stock carried at full cost is profit that has not yet been reversed.

Depreciation and useful lives. A longer assumed life means lower annual depreciation and higher profit. Within rules, and disclosed. A company that extended its asset lives has raised profit without selling anything, and the note says so.

Capitalisation. Costs recorded as assets rather than expenses do not hit this year's profit. Development costs, some software, interest during construction. All legitimate in the right circumstances, and all able to flatter if stretched. The tell is the same: profit improves, operating cash flow does not.

What good quality looks like

Cash conversion near or above 1, consistently.

Few exceptional items. Chapter 4: items that recur are not exceptional.

Stable accounting policies. Changes are disclosed, and a change that happens to raise profit in a difficult year is worth understanding.

Conservative estimates. Depreciation lives at the shorter end, provisions at the fuller end. Dull, and it means reported profit is more likely to be understated than overstated.

A clean audit opinion with no Statement of Impact of Audit Qualifications, from chapter 7.

What poorer quality looks like

Persistent gap between profit and operating cash flow, explained by growing receivables or inventory.

Profit growing faster than revenue, year after year, without an identifiable operational reason. Margins can expand for real reasons; sustained expansion with no mechanism is worth probing.

Frequent exceptional items, always costs and never gains.

Accounting changes that raise profit, particularly in a weak year.

Large other income relative to operating profit — chapter 4, and it means the profit is not from the business.

Related party revenue that is material, from chapter 15.

Earnings management is not fraud

An important distinction.

Within the rules there is a legitimate range for most estimates, and choosing the optimistic end of it is not illegal and not necessarily dishonest. Management has incentives — targets, options, covenants — and incentives shape judgement without anyone deciding to deceive.

What this means for you: you are not looking for a crime. You are forming a view on whether the reported number is likely to be the conservative or the optimistic version of a legitimate range, and adjusting your confidence accordingly.

The company that consistently reports slightly less than it could is worth more trust, and over a decade usually turns out to have been worth more money.

A simple adjustment

If you want one number instead of a judgement: use the five-year average of free cash flow rather than reported profit as the basis for chapter 17's multiples.

It is not perfect — chapter 6's caveat about growth capital expenditure applies — and it is immune to almost every estimate in this chapter, because cash either arrived or it did not.

The point

Cash conversion over five years is the core test: operating cash flow against net profit. Judgement enters through revenue recognition, provisions, inventory, asset lives and capitalisation, and in each case flattering profit leaves operating cash flow behind. You are not looking for fraud — you are judging whether the reported number is the conservative or the optimistic version of a legitimate range.