The Income Statement

From revenue to net income, one line at a time

Revenue is the top line, and it is the line with the most judgement in it. It is recognised when a performance obligation is satisfied — when control of a good or service passes to the customer — which is why a software company that signs a three-year contract books a third of it a year, why a retailer books it at the till, and why the accounting rules on this single question fill a standard. Revenue is not cash received; it is the amount the company believes it has earned. Cost of goods sold is the direct cost of what was sold: materials, manufacturing, freight, and the direct labour that made the product. Subtracting it gives gross profit, and gross profit divided by revenue is the gross margin — the single number that best describes the product itself. A business with a 60% gross margin can afford to spend on sales and research; one with 15% has to sell enormous volumes to pay for anything. Below the gross-profit line come operating expenses: selling, general and administrative costs and research and development. These are period costs rather than product costs, and they are largely fixed within a year — a hiring plan is a decision, not a variable that tracks revenue. Subtracting them, and depreciation, gives operating profit or EBIT: earnings before interest and taxes, which measures the business rather than its financing. EBIT over revenue is the operating margin. Then the capital structure arrives. Interest expense is the cost of debt, subtracting it gives pre-tax income, and applying the tax rate gives net income. Divide by the share count and you have earnings per share — and you should use the diluted figure, which counts every share that could exist through options, warrants and convertibles, because that is what the claim on the residual might actually have to be shared with. One year, three margins — Revenue: $4,800m · Gross profit at a 40% margin: $1,920m · Operating profit after $1,150m SG&A and $270m D&A: $500m — a 10.4% margin · Net income after $80m interest and 21% tax: $331.8m — a 6.9% margin · Diluted shares: 190m · Earnings per share: $1.75 Adjusted earnings are the company’s own opinion of its year. Every excluded item — restructuring, impairments, stock compensation — was a real cost that somebody paid. Read the gap between reported and adjusted EPS, because that gap is where the argument is.

Why the margins move so much more than revenue

A gross margin is the product’s economics; an operating margin is the business’s; a net margin is the owners’. Because costs below the gross-profit line are largely fixed, a small change in gross margin passes through to operating profit nearly undiluted. In the year above, a two-point fall in gross margin — say discounting to hold volume — cuts gross profit to $1,824m and operating profit to $404m: a 19% fall in operating profit from a 5% fall in gross profit. The same arithmetic runs in the other direction, which is why operating leverage is attractive as well as dangerous. A company with high gross margins and a fixed cost base can raise revenue 10% and raise operating profit 25%. It also means that a single quarter of soft revenue can turn a modest miss into a large one, and it is why the market reacts to a margin line far more violently than to a revenue line. Two other traps sit in the same statement. The first is that operating profit excludes stock-based compensation in the adjusted figures while dilute share counts include the shares it creates — one number ignores the cost, the other records the dilution. The second is that a company can hold margins up for a period by capitalising costs onto the balance sheet, which is F8’s subject: profit up, cash flat, and nothing on the income statement to show you why. The statement is a flow over a period, and it records accrual, not cash. A sale on credit is revenue today and cash later, which is why the cash flow statement (F4) is the check on everything here.

The note that explains the revenue line

A revenue figure on its own is a single number standing in for a business. The **segment note** is where it comes apart, and it is the first note a professional reads — because a company growing 12% overall can be a company whose largest division is shrinking and whose growth came from a small one that cannot carry it. Three things in that note change the analysis. **Segment mix** decides what the blended margin means: a rising share of a 20%-margin product inside a 60%-margin company dilutes earnings per dollar of revenue even while the headline grows. **Geographic mix** decides how much of the number is exposed to a currency you are not analysing. And **customer concentration**, when disclosed, can put a whole thesis on one counterparty’s purchasing decision. The practical habit is to write the revenue line as a sum before forecasting anything, because one forecast for the whole company hides three disagreements inside it. If divisions A, B and C have different drivers — a subscription base, a cyclical price, a new product — then the honest model has three lines and one total. The rule of thumb: any division above 15% of revenue earns its own line. A 12% year, taken apart — Total revenue: up 12% · Largest segment, 60% of revenue: down 3% · Second segment, 25%: up 40% · Small segment, 15%: up 9% · What the headline hides: the engine is shrinking and the story is a small, fast-growing division ← Read the segment note before you build a model, not after. Discovering that the growth came from 15% of the business turns a forecast into a different question rather than a tweak.

When the revenue is allowed to exist

The revenue line looks like the least arguable number in the statement, and it is one of the most rule-dependent. The standard that governs it requires a company to identify a contract with a customer, identify the distinct promises in it, determine the price, allocate the price across those promises, and recognise revenue as each one is satisfied. Applied to a simple sale of a physical product that is almost automatic. Applied to a multi-year software subscription sold with implementation services and a free year of support, the same rule can put eighty percent of the cash in a different year from the rest. The consequence is that two businesses can report identical revenue in the same period and be in completely different places. A company that bills annually in advance collects the cash on day one and recognises revenue across twelve months, so its deferred revenue balance — a liability — grows faster than its income statement shows. A company that recognises revenue on delivery recognises all twelve months at once. The first is the healthier cash profile and the more conservative revenue profile; a screen that reads only revenue will rank them identically, and a screen that reads deferred revenue and billings alongside it will not. For an analyst the practical questions are concrete. What is the **liability for deferred revenue**, and is it growing faster or slower than recognised revenue? Is any revenue recognised from a contract whose customer has not paid, and how large are the receivables against it? Has there been any change of estimate, any restatement, or any disclosure of a contract modified after signing? Those three checks catch most of the cases where revenue timing is doing work that the business is not, and they can all be run from the filings without a model. One contract, two places for the cash — Annual subscription billed up front: Cash now, revenue spread over twelve months, deferred revenue grows · Same product billed monthly: Revenue recognised as billed; little deferred balance · Perpetual licence with support: Price allocated across both promises; support earned over the term ← A jump in receivables relative to revenue is not proof of anything on its own — some businesses legitimately extend terms. It is, however, the point at which the revenue line stops being sufficient evidence on its own, so read the note that explains it.

The tax line, and the three rates inside it

The tax line is treated as the mechanical item at the bottom of the income statement, and treating it that way is how a forecast goes wrong by more than any operating assumption is likely to. Three different rates live in that line, and they answer three different questions. The **statutory rate** is the law: the headline corporate rate applied to domestic taxable income. It is a look-up, it changes only when the law does, and it is the rate the other two are described against. The **effective rate** is the provision shown in the income statement divided by pre-tax book income, and it departs from the statutory rate for a list of ordinary reasons — profits earned in jurisdictions taxed differently, state taxes, credits, non-deductible expenses, valuation allowances moving, prior-year audit settlements, and the tax effects of share-based compensation. Several of those are one-off by nature, which is why this year’s effective rate is a poor forecast of next year’s. The third is the one that belongs in a cash flow model: the **cash tax rate**, which is what the company actually paid, and which appears in the cash-flow statement rather than in the income statement. It differs from the effective rate by the movement in deferred taxes, and the difference can persist for years. A company can report a strikingly low effective rate while paying close to the full statutory amount in cash, and a company can do the reverse, and the two statements describe the same year without contradicting each other. The page that makes all of this readable is the effective-tax-rate reconciliation, which lists each item’s contribution in percentage points. It is short, it is required, and it is the fastest way to find out whether a low tax rate is a durable feature of where the company earns its money or a temporary artefact of a credit, a settlement or a jurisdiction that could be changed by a single legislative session. When such a rate normalises, net income falls and nothing in the operating business has moved — which is a share-price event that appears in no operating forecast. The practical rule is short. Forecast the cash tax rate from the reconciliation rather than extrapolating the last reported effective rate, treat any line in the reconciliation that has moved by more than a point or two as a question that needs an answer, and keep in mind that tax changes arrive as re-ratings of whole businesses rather than as operating events. The line looks like arithmetic. It is a forecast. • Statutory is the law, effective is this year’s provision, cash is what was actually paid. • One-off items make a single year’s effective rate a weak forecast. • Deferred taxes are the reason the effective and cash rates can differ for years. • The tax-rate reconciliation, in percentage points, is the fastest read on whether a low rate lasts. In a valuation model the cash tax rate is the one that touches free cash flow, while the effective rate is the one that touches earnings per share. Using one where the other belongs is a small error that compounds across a forecast.

Sources

• The form and content of the income statement, and what each line excludesSEC Regulation S-X, Rule 5-03; FASB ASC 220
• Revenue recognition: when a sale is allowed to be called revenueFASB ASC 606, Revenue from Contracts with Customers
• Diluted earnings per share, and why the diluted figure is the honest oneFASB ASC 260, Earnings Per Share
• Non-GAAP measures, and the SEC’s conditions on using themSEC Regulation G; SEC Compliance and Disclosure Interpretations