Stock-Based Compensation and Dilution

Canonical data and boundary. Consolidated figures tie to Wexford Instruments (USD millions). The stock-based-compensation charge of 15.0 used below is this article's illustration, layered onto Wexford at its scale — Wexford's canonical cash-flow reconciliation (net income 62.3, plus depreciation and amortisation 50.0, less working capital 14.0, equals 98.3) carries no separate stock-based-compensation add-back, so the 15.0 does not alter the canonical statements and other pillars cite it as this article's illustration. This article covers how equity pay is recognised and disclosed. The arithmetic of a changing share count — how dilution and buybacks move ownership percentages — is covered in Pillar 11 and is linked rather than repeated.

The question sounds philosophical and is worth several percentage points of reported profit at a great many companies. It is also the most actively argued item in this pillar — sophisticated investors disagree about it in public, and the disagreement is not merely a matter of preference. This article sets out the accounting, the strongest case on each side, and where the argument actually resolves.

The accounting, briefly

Equity compensation comes in a few forms. Share options give the right to buy shares at a set price, valuable only if the price rises above it. Restricted stock units are promises of shares on a future date, valuable at whatever the price then is. Performance shares vest only if targets are met. All of them are compensation, and all are recognised as an expense.

Measurement uses grant-date fair value, spread over the vesting period. For restricted units that is close to the share price at grant. For options it requires a valuation model, whose inputs include expected volatility, expected life, and the risk-free rate — so the expense is an estimate built on estimates, which is the same trade accrual accounting makes everywhere. Once set at grant, the charge for equity-classified awards is generally not remeasured for later price moves, which means the expense recorded and the value eventually transferred are two different numbers — a point both sides of the debate below rely on.

The argument, at full strength on both sides

The case that it is not a real expense. No cash leaves the business — the company writes no cheque, and its bank balance is identical whether or not options are granted. The cost, on this view, falls on existing shareholders through dilution rather than on the company, and it is already captured by the increase in share count. Which leads to the strongest version of the argument: charging the income statement and counting the extra shares in diluted EPS counts the same thing twice. That is a serious point and it deserves to be stated as seriously as its proponents state it.

The case that it is a real expense. The company received labour and gave up something of value for it. The decisive test is the equivalence: if the company had issued those shares to investors for cash and used the cash to pay salaries, nobody would question that the salaries were an expense — and the economics are identical. Treating equity pay as free therefore makes labour appear free whenever it is paid in a particular currency, which would let two companies with the same workforce and the same output report different profitability purely because of how they chose to settle payroll. On the double-counting objection, the answer is that the two figures measure different things: the expense measures the value granted, while the share count measures the ownership eventually transferred, and neither is a substitute for the other.

Where this resolves. The accounting standard-setters concluded — after prolonged and contested debate — that it is an expense, and that is the reporting requirement under both frameworks. The remaining live disagreement is best understood as being about measurement rather than existence: whether grant-date fair value is the right number, whether option models are reliable, and how the expense should be presented. A reader does not need to settle it. What a reader needs is to notice when a figure has been presented with the expense removed — which is extremely common.

Adjusted earnings, and the cash that hides in financing

Stock-based compensation is among the most commonly excluded items in "adjusted" or non-GAAP earnings, on the reasoning that it is non-cash. The effect is not marginal: removing an illustrative 15.0 charge lifts operating income from 100.0 to 115.0 — a 15% improvement — and, tax-effected at the canonical 24.0% rate, lifts net income from 62.3 to 73.7, up 18.3%, without anything changing in the business.

And here is the part most often missed. Companies that grant equity frequently repurchase shares to limit the resulting dilution — and a buyback is entirely cash. On the illustration, the company granted 15.0 of stock-based compensation and repurchased 12.0 of stock — Wexford's canonical buyback — so a cash outflow equal to 80% of the supposedly non-cash charge left the business in the same year. It appears in financing activities rather than operating expenses, so a reader looking only at adjusted operating profit never encounters it. The expense is non-cash; the company's response to it very often is not. That is not an accusation — offsetting dilution is a reasonable use of capital — but it means "non-cash, therefore ignorable" does not survive contact with the cash-flow statement.