Stock based compensation

4.2.1 The economics first

A company can pay an engineer three ways: cash now (salary), cash later (bonus, pension), or a slice of the company (shares or options). That third option feels free because no cash leaves. It isn't. The company is settling a real compensation obligation by printing new claims on itself, and existing shareholders pay for it through dilution: the same future profits divided across more shares. None of it shows up in the bank account. Accounting's job is to force this cost into the income statement anyway. Your job, covered ahead, is to make sure the cost doesn't slip back out through the cash flow statement. What follows is the equity-settled case you'll meet every day. Share-based compensation gets a deeper treatment on the cash-settled share-based payment page, which adds cash-settled awards. Their liability is remeasured to fair value at every reporting date instead of being fixed at grant.

What is the difference between RSUs and stock options?

Two instruments dominate:

• Restricted stock units (RSUs): a promise of one share per unit if you stay (and sometimes hit targets). Value at grant ≈ the share price itself. An RSU is worth something even if the stock falls. That's why the industry shifted heavily from options to RSUs after the dot-com bust and the 2006 expensing rules.
• Employee stock options (ESOPs in Indian usage): the right to buy a share at a fixed exercise (strike) price, typically the market price at grant. Options vest over ~3–4 years and can be exercised for up to ~10. Each one is worth much less than an RSU, because you have to pay the strike. It's worthless if the stock stays below the strike ("underwater"), and far more leveraged to the upside.

4.2.2 Measurement: grant-date fair value, frozen

ASC 718 (US GAAP), IFRS 2 and Ind AS 102 all use the same core rule for equity-settled awards:

Measure the award's fair value once, at grant date, and recognize it as compensation expense over the vesting (service) period. Do not remeasure for later stock-price moves.

• RSU fair value = grant-date share price, less the present value of any dividends the holder won't receive while unvested.
• Option fair value = an option-pricing model (Black–Scholes or a lattice/binomial model) driven by the share price, strike, expected life, volatility, dividend yield and the risk-free rate. A typical at-the-money 4-year option is worth very roughly 25–40% of the share price, and the SBC note discloses the inputs. Awards with market conditions (e.g., share-price or TSR hurdles) get a Monte-Carlo value. That expense is never reversed, even if the hurdle is missed, because the chance of missing was already priced in.
• Service and non-market performance conditions (stay 4 years, hit an EPS target) work the other way. They don't enter the per-unit value. Instead, the number of awards expected to vest is adjusted, and expense is finally trued up to what actually vests.

Because the value is frozen at grant, a stock that triples after grant creates no extra expense. The P&L records the old, small number while the economic transfer to employees balloons. Hold that thought for the half-truth section ahead.

How do cliff and graded vesting differ?

Vesting schedules. A cliff vest (100% at end of year 3) is expensed straight-line. A graded vest (25% per year over 4 years) exposes a real standards difference:

• IFRS 2 / Ind AS 102: each tranche must be treated as a separate award, expensed over its own vesting period. That front-loads the expense (the "accelerated" or graded-attribution method).
• US GAAP: for awards with only service conditions, companies may elect straight-line over the total period, subject to a floor (cumulative expense ≥ the vested portion). Most US companies elect straight-line.

Here's a worked contrast with synthetic, clean numbers. The grant is 120,000 options at a grant-date fair value of ₹100 each (total ₹1.2 crore), vesting 25% at the end of each of years 1–4.

Here's the graded logic. Tranche 1 is 30,000 options × ₹100 = ₹30 lakh, expensed over year 1 alone. Tranche 2 is spread over years 1–2 at ₹15 lakh/yr, tranche 3 over three years and tranche 4 over four. Strictly, each tranche would also carry a slightly different fair value because its expected life differs. We hold it constant to isolate the attribution effect. Same total, very different timing. An Ind AS company's SBC expense spikes in the year of a big grant and then fades, while a US straight-liner's is smooth. Year-over-year margin comparisons across the two regimes have to respect this.

Forfeitures. IFRS 2/Ind AS 102 estimate expected forfeitures and true up. US GAAP (post-2016) lets companies either estimate them or simply account for forfeitures as they occur, so read the policy note.

4.2.3 The triple entry: one decision, three statements

SBC makes a great drill for the closed loop you built in the statement-linkage capstone, because one annual entry touches all three statements:

1. Income statement: Dr Employee stock compensation expense. It's booked inside COGS/SG&A/R&D by function (US) or employee benefits expense (India). Profit falls.
2. Cash flow statement: in the indirect CFO build, SBC is added back to net income, since no cash left the building. CFO is unaffected by the expense.
3. Balance sheet: Cr Share options outstanding account / APIC. Equity is built up as the credit side of the expense. Under Schedule III (India) it appears as its own line inside Other Equity. Under US GAAP it accumulates in additional paid-in capital.

So the loop closes neatly. Net income is lower by the expense, but equity is higher by the same credit, so the total equity change nets to zero cash effect. That matches what happened economically. No cash moved, and a claim was granted.

What happens on exercise or settlement?

On exercise (options), the employee pays the strike and the accumulated grant-date value moves within equity.

Take a synthetic example: 30,000 vested options, strike ₹500, accumulated fair value ₹100 each, face value ₹10, and a market price on exercise day of ₹900.

Dr Cash                                 ₹1.50 cr   (30,000 × 500)
Dr Share options outstanding account    ₹0.30 cr   (30,000 × 100)
   Cr Equity share capital                  ₹0.03 cr   (30,000 × 10)
   Cr Securities premium (APIC)             ₹1.77 cr   (balance)

Notice what doesn't appear. The option is in the money by ₹400/share beyond its grant-date value (₹900 − 500 = 400 intrinsic vs ₹100 booked), and that extra transfer to the employee never touches the P&L, because grant-date measurement is final. Shareholders pay it through dilution, out of sight of the income statement. Taxes see it differently. The employer's tax deduction in both the US and India is generally based on the exercise-date intrinsic value. SBC-heavy companies show odd effective-tax-rate lines as a result, the same pattern you see in deferred tax.

On RSU settlement, shares are issued and amounts reclassify within equity. There's one cash reality here, though. Companies usually withhold shares to pay the employee's taxes and remit cash to the tax authority. That's a real cash outflow, classified as a financing outflow under US GAAP. For big US tech it runs into billions, which makes for "non-cash" compensation with a very cash tax bill attached.

4.2.4 Dilution: the treasury-stock method

You met diluted EPS on the basic vs diluted EPS page. It asks a question: if all in-the-money options were exercised, how many extra shares would exist? The treasury-stock method (TSM) answers with a thought experiment. Assume exercise. Then assume the company uses the exercise proceeds to buy back shares at the average market price of the period. Only the net new shares count.

Incremental shares = N options × (1 − strike ÷ average market price)

Try it with 10 crore options, a strike of ₹250 and an average price of ₹400. Assumed proceeds are ₹2,500 cr, which buy back 2,500 ÷ 400 = 6.25 cr shares, so incremental = 3.75 cr shares. Done directly, that's 10 × (1 − 250/400) = 3.75 cr. Options with a strike above the average price are antidilutive and excluded. Unvested RSUs go through the same machinery with a (near-)zero strike, so they dilute almost one-for-one. Technically, unrecognized compensation cost is also treated as assumed proceeds under ASC 260 and IAS 33. The filings' diluted count already builds in this refinement, so use the filed number instead of rebuilding it.

Diluted EPS = income attributable to parent shareholders ÷ (basic weighted shares + incremental shares). Always use the diluted count for per-share value work. For an SBC-heavy company, also check the option/RSU overhang in the SBC note, meaning total outstanding awards as a % of shares out. At 2% it's background noise. At 8–10%+ it's a real claim on your ownership.

4.2.5 Why "SBC is non-cash" is a half-truth you must reject

The add-back in CFO is mechanically correct and analytically dangerous. Hold four truths at once.

1. It's real compensation. If the company paid the same value as cash salary, CFO would fall rupee-for-rupee. Paying in shares changes the route of the cost (dilution instead of cash), and the cost is still there. A company that "generates" ₹1,000 cr of CFO including a ₹300 cr SBC add-back is buying a third of its apparent cash generation with its shareholders' ownership.
2. Grant-date expense can badly understate the current transfer. In a rising stock, this year's P&L carries the stale grant-date value of old awards while employees receive far more value at vest/exercise. A falling stock does the reverse, and expense keeps accruing on underwater options no one values.
3. The buyback tell. Many mature companies run buybacks roughly sized to mop up SBC issuance. The cash for those buybacks shows up in CFF, below the FCF line most analysts quote. Microsoft's SBC is on the order of $10–11bn a year against ~$245bn revenue (~4%; illustrative, modeled on ~FY24 scale, so pull the 10-K). Part of its buyback simply neutralizes that issuance. At some younger software firms SBC has run at 20–40% of revenue, and Snowflake has been near the top of that range (check in its latest 10-K). At those firms, an "FCF margin" quoted with SBC added back is close to fiction.
4. India runs cooler, but not cold. Infosys' RSU expense is well under 1% of revenue (illustrative, ~FY24/FY25 scale, so pull the AR's ESOP note), and Indian promoters dislike dilution. But new-age listings (Zomato/Eternal, Paytm) carried heavy ESOP charges at the very moment they claimed "adjusted EBITDA" profitability. The adjustment was, largely, adding SBC back. SEBI-registered analysts fight this fight in India too.

How should SBC be treated in free cash flow?

There are two consistent ways to treat it. Pick one, because using both at once double-counts:

• Treat SBC as a cash cost: FCF = CFO − capex − SBC. It's simple and conservative, and it's Damodaran's default. Then value using the current diluted share count.
• Leave FCF alone but model dilution: keep the add-back, but grow the share count for expected future issuance (and value existing options separately). It's harder to do well.

You can never defend adding back SBC while also using a frozen share count. That values the company as if employees work for free.