Market Pulse
Mixed observation times, Oct 20–Oct 25, 2026 (5 days apart)
Learn · Reading the accounts· Intermediate · 6 min

Stock-based compensation

Paying staff in shares is a real cost that never leaves the bank account. That combination is why it is the most argued-about line in the accounts.

A company can pay an engineer in two ways: cash, or shares in itself. The second is genuinely cheaper for the company's bank balance, and it is not free. Working out who actually pays is the whole lesson.

Where it appears

Stock-based compensation — usually abbreviated SBC — is recorded as an operating expense on the income statement, mixed into cost of revenue, research and development, and sales and administration depending on who received it. It reduces reported profit exactly as a salary would.

Then, on the cash flow statement, it is added straight back. That is not a trick. The cash flow statement starts from net income and reverses every expense that did not consume cash, and SBC did not: no money left the company.

Fact
The add-back is correct accounting and an incomplete picture
It is true that SBC consumed no cash. It is also true that the company handed over a slice of itself. Cash flow statements measure cash, so they do the first thing and cannot do the second. The cost shows up elsewhere — in the share count.

Who actually pays

Existing shareholders. If a company has 100 million shares and issues 2 million to staff, every prior owner now holds a slightly smaller fraction of the same business. Their claim on future earnings fell by roughly 2%.

Wealth Signal calculation
An illustrative year
A company earns $500m and has 100m shares, so earnings per share is $5.00. It issues 3m shares to employees. Next year, with identical earnings of $500m, EPS is $500m ÷ 103m = $4.85 — a 3% fall in per-share earnings with no change whatsoever in the business. That is the cost, and it is paid by the shareholder rather than by the company.

The buyback offset

Many companies spend cash buying shares back to keep the count flat. Watch what that does: the company is issuing shares to staff and then using real cash to buy them off the market. The net effect is that the company paid cash for labour, via a detour.

Interpretation
Buybacks that only offset dilution are compensation, not capital return
When a company reports a large buyback programme and a share count that barely moves, the buyback is not returning capital to shareholders. It is paying the payroll through the market. The two are frequently reported as though they were the same thing, and they are not.

The adjusted-earnings argument

Many companies present an adjusted profit figure that excludes SBC, on the reasoning that it is non-cash and volatile. The counter-argument is short: the staff would not have worked for nothing, so excluding what they were paid overstates what the business earns.

Both sides have a point about measurement and only one side has a point about economics. See adjusted earnings for the general pattern.

What to actually check

  • SBC as a share of revenue. A single percentage tells you how much of the business is being paid out in ownership. Compare it across years for the same company, not across industries.
  • Diluted share count, over five years. This is the number that matters, and it is the one that cannot be adjusted away. If it rises steadily, dilution is real regardless of what the earnings release says.
  • Buyback spending versus the change in share count. If billions were spent and the count is flat, you now know what the buyback was for.
  • Free cash flow with SBC subtracted. Not a standard measure, but a useful sanity check on how much the business would generate if it paid its people in money.
None of this makes stock compensation bad. Aligning staff with owners is a real benefit, and start-ups could not hire without it. The point is only that it is a cost with an identifiable payer, and that the payer is you.
Education, not advice. These lessons explain how things work. They do not tell you what to buy, and Wealth Signal does not publish recommendations. Full disclaimer.