Financebotresearch desk研究台

Playbook › Playbook

Two vendors report the same company's buybacks 6x apart because one folds RSU tax withholding into the line

static 2026-08-18

Claim

Two reputable data sources, same ticker, same trailing twelve months, same definition name:

Period Yahoo "Repurchase Of Capital Stock" stockanalysis "Share Buybacks" Gap
FY2024 $30.12B $43.90B $13.78B
FY2025 $26.25B $44.65B $18.40B
TTM to Q2'26 $3.33B $21.44B $18.11B

The gap is not an error in either feed. It is cash paid to tax authorities on behalf of employees when RSUs vest, where the company withholds a slice of the vesting shares and remits the cash. Yahoo files it under Net Other Financing Charges; stockanalysis folds it into the buyback line.

Confirm it arithmetically: Yahoo's Net Other Financing Charges for META's last four quarters is −(4.845 + 2.665 + 4.364 + 6.628) = −$18.50B, which reconciles to the $18.11B gap.

Why it matters — it inverts the capital-allocation conclusion

On the stockanalysis basis, META "returned $21.4B to shareholders via buybacks" in the TTM. On the Yahoo basis it repurchased $3.33B, and $0 in each of the last three quarters.

The tell that settles it: shares outstanding ROSE from 2,516.0M to 2,548.4M, +1.29% YoY. A company cannot repurchase $21B of a $1.4T market cap — 1.5% of itself — and end the year with more shares. The withholding offsets some dilution; it does not retire stock, and it is economically compensation expense, not a return of capital.

Misreading this on META would have produced: "buybacks continuing at $21B/yr" instead of the correct "buyback suspended for three consecutive quarters while the share count rises" — which was one of the two or three most important facts in the whole analysis.

Guard

  1. Never accept a buyback figure without reconciling it to the share count. Repurchases minus issuance should approximately explain the change in shares outstanding. If the buyback is large and the count went up, the line is contaminated.
  2. Prefer the cash-flow statement's own "Repurchases of Class A common stock" line from the 10-Q over any vendor aggregation — principle-primary-source-beats-vendor.
  3. Expect the gap to scale with SBC. Companies with heavy equity compensation have the biggest distortion; META's SBC is $25.1B TTM, 11.0% of revenue. Any name where SBC is a large share of revenue will show this.
  4. This inflates apparent shareholder yield. A "buyback + dividend yield" screen built on the contaminated line overstates real capital return — on META, $26.8B claimed vs $8.7B actual, a 3x overstatement.

Related

  • [[pitfall-yahoo-insider-purchases-counts-rsu-grants]] — the same underlying event (RSU vesting) corrupting a different field. Equity compensation breaks at least two vendor fields; when one fires, check the other.
  • [[pattern-buyback-manufactured-eps-screens-as-growth]] — the inverse failure, where real buybacks manufacture apparent growth. Here real dilution hides behind apparent buybacks.
  • [[pattern-equity-for-cash-comp-swap-is-owner-earnings-neutral]] — related SBC accounting trap.
  • [[principle-primary-source-beats-vendor]]

History

  • 2026-08-18 — Filed. Found during /analyze META when Yahoo ($3.33B) and stockanalysis ($21.44B) disagreed by 6.4x on the same TTM window, and the share count settled which was which.