Earnings Quality: Accruals, Cash Flow Gaps and Accounting Red Flags

Two numbers in the same filing, printed a dozen pages apart, are supposed to describe the same year. Net income sits at the bottom of the income statement. Cash from operations sits at the top of the cash flow statement. When they disagree by a lot, and keep disagreeing in the same direction, you are looking at the most useful signal available to anyone reading annual reports for a living.

Earnings are the output of hundreds of judgments about timing. When revenue is earned. How long a machine lasts. What share of receivables goes bad. Whether a cost belongs on the income statement or the balance sheet. Every one of those has a defensible range, and a company under pressure can sit at the aggressive end of all of them at once without breaking a single rule.

Cash is harder to shape. That is the whole idea. You should already be able to read an earnings report before going further.

Start with one subtraction

accruals ratio = (net income - cash flow from operations) / average total assets

A company reports net income of $480 million and cash from operations of $210 million, while total assets opened the year at $3,200 million and closed at $3,600 million, so average assets are $3,400 million.

(480 - 210) / 3,400 = 270 / 3,400 = 0.079

An accruals ratio of 7.9 percent. Over half the reported profit never arrived as cash, and the difference is parked somewhere on the balance sheet as a receivable, an inventory item or a capitalized cost.

The companion ratio takes ten seconds. CFO / net income = 210 / 480 = 0.44. A mature business with real depreciation normally converts more than a dollar of cash per dollar of reported profit, because depreciation is a non cash charge that gets added back. Forty four cents is the opposite of that, and three consecutive years below one is the version of this signal that matters.

Why the gap has predicted anything at all

Richard Sloan’s 1996 work on the accruals anomaly documented that firms with high accruals tended to underperform afterwards and firms with low accruals tended to outperform, and the explanation offered was that investors fixate on the earnings line and treat the cash portion and the accrual portion as interchangeable, when the accrual portion is far less persistent.

The mechanism is estimates reverting. A receivable nobody collects becomes a write off. An inventory build against demand that never shows up becomes a markdown. Revenue pulled forward out of next year leaves next year short, and next year arrives on schedule.

The evidence has weakened since publication, which is the usual fate of a well publicized anomaly, and the factor investing guide covers why that keeps happening. It has held up as a screen for what to avoid.

The balance sheet tells on the income statement

Second pass now, on the balance sheet. Compare growth rates.

Item	Prior year	Current year	Growth
Revenue	$1,800m	$2,050m	13.9 percent
Accounts receivable	$300m	$410m	36.7 percent
Inventory	$240m	$340m	41.7 percent

Turn receivables into days sales outstanding. A percentage does not tell you what changed operationally.

prior: 300 / 1,800 x 365 = 60.8 days current: 410 / 2,050 x 365 = 73.0 days

Twelve extra days of waiting to be paid. That has innocent explanations, including a shift in customer mix toward large enterprises with longer terms, and it has the other kind of explanation too: credit terms loosened to pull sales into the quarter, or product shipped to distributors who have not sold any of it on. The filing should say which. Silence in the filing is itself an answer.

Do the same arithmetic on inventory. Inventory growing at three times the rate of revenue is a demand forecast the company has already paid for in cash.

Then carry the example one more step. If capital expenditure was $160 million, free cash flow is 210 - 160 = $50 million against $480 million of reported profit. There are only three ways to explain a gap that size. The company is investing hard for growth, in which case the revenue should show up within a couple of years and you can hold them to it. Or maintenance capital spending genuinely runs above depreciation, in which case reported profit overstates the economics permanently. Or the profit is not there. Any discounted cash flow model built on reported earnings will be wrong in the same direction the accounting is wrong.

Adjusted earnings and the permanent one-off

Adjusted figures can be genuinely useful. A legal settlement, a factory fire, a real restructuring. Strip those out and you get closer to normal earning power.

The test is repetition. GAAP diluted EPS of $1.85 against adjusted EPS of $3.20 leaves $1.35 of difference, and suppose $0.90 of it is stock based compensation. That is a real transfer of value from existing shareholders to employees, it happens every single year, and excluding it produces a number that describes no version of the company that exists. An adjustment appearing in five consecutive annual reports is an operating cost wearing a euphemism.

There is a related check on the cash side. A company repurchasing more stock than its free cash flow covers is borrowing to do it, which is worth remembering when you read about buybacks. The debt lands on the balance sheet quietly and the EPS improvement lands in the press release loudly, so compare repurchases in the period to free cash flow before you read a shrinking share count as strength.

The things that live in the notes

Several of the better signals never appear in the headline statements. A change in depreciation schedules or estimated useful lives cuts the annual charge and lifts reported profit with nothing happening in the business, and it gets disclosed in the accounting policy note, usually in one sentence, usually without a number attached.

Capitalizing costs that used to be expensed moves spending off the income statement and onto the balance sheet, which improves profit now and creates amortization later, and software development and customer acquisition costs are where it usually happens.

Classification inside the cash flow statement is the sharpest of the three. Shift an outflow from operating to investing and operating cash flow improves while the total does not move at all, which flatters every ratio in this guide directly.

An auditor resignation, a change of auditor, a restatement or a disclosed material weakness in internal controls sits in a different category from all of the above. Those are rare, and they are the strongest single signals in the whole exercise.

Run it in this order, then move on

Accruals ratio and cash conversion for three years, read as a trend. Receivables and inventory growth against revenue growth, converted to days. Free cash flow against net income, with the gap explained. The GAAP to adjusted reconciliation, counting how many years each line has appeared. The accounting policy note for changes in estimates. The auditor and any control disclosures. Then the multiple you are being asked to pay, through the P/E calculator, set against the quality of the earnings sitting in the denominator.

One flag proves nothing, and most companies carrying one have a reasonable answer, but three or four in the same filing is a different animal, and the right response is usually to close the tab and go look at something else.

Where this fails is timing, and it fails badly. Accounting can deteriorate for years while the stock rises, which is exactly what makes shorting on these grounds so expensive, as the short selling guide sets out. The ratios also misfire on genuine fast growth, since a company doubling revenue builds receivables and inventory ahead of sales and will screen as aggressive for entirely honest reasons, so compare a company to its own history and to two or three close competitors, or the numbers mean very little.

And the last caution is the one that keeps people humble. Most of the large accounting failures of recent decades were visible in the filings before they were visible in the headlines, and so were the same patterns in dozens of companies that turned out to be fine. Screening produces candidates. Separating them is judgment, and the analysis checklist, the valuation basics guide and the stock screener are where that work continues.