Alphanume

Insights

What Is a Stock Buyback? Authorizations vs Actual Purchases

Alphanume Team · October 27, 2026

A stock buyback is a company's purchase of its own shares. A buyback announcement is a statement about a plan to make those purchases. The difference matters: a board can authorize $500 million without the company immediately spending $500 million, and a later purchase report can show a much smaller amount.

To research buybacks, keep three quantities separate: the capacity authorized, the purchases executed, and the net change in shares outstanding. Alphanume's Stock Buybacks dataset tracks announcements and changes to programs. It helps establish what the company said it intended to do; purchase disclosures and financial statements are needed to establish what happened.

How a buyback changes ownership and earnings per share

Consider a hypothetical company with 100 million shares outstanding, a $10 share price, and $100 million of annual net income. Its market capitalization is $1 billion and its simple earnings-per-share calculation is $1. A shareholder who owns one million shares owns 1% of the company.

Suppose the company actually spends $100 million buying ten million shares at $10, and those shares cease to be outstanding. If nothing else changes, 90 million shares remain. The same shareholder now owns about 1.11%, and $100 million divided by 90 million shares gives about $1.11 per share.

Hypothetical quantityBefore purchasesAfter purchases
Shares outstanding100 million90 million
Shareholder's shares1 million1 million
Shareholder's ownership1.00%About 1.11%
Simple EPS at unchanged $100 million income$1.00About $1.11
Cash spent on repurchases$0$100 million

This example isolates the denominator. Reported basic EPS uses weighted-average shares over the reporting period, rather than simply the ending share count; diluted EPS can include additional potential shares. Cash has also left the company, and borrowing to finance a purchase can change interest expense. The arithmetic does not establish that the share price should rise by 11.1% or that shareholder wealth increased by the same amount.

Why a company might buy back stock

A repurchase can return cash to shareholders who sell, reduce the share count, offset shares issued through compensation, or express management's view that buying the company's equity is a good use of capital. Those are different propositions. A company can offset issuance while barely reducing shares outstanding, or shrink the share count while paying a price that later looks expensive.

Read management's stated rationale, then test it against the balance sheet and the purchase record. How much cash is available after operating needs? What financing supports the program? Is the authorization large relative to the company, and does it expire? The pct_of_market_cap field in the buyback dataset makes announced program sizes easier to compare, but it measures announced capacity rather than a guaranteed percentage reduction in shares.

Authorization, execution, and remaining capacity

Imagine a hypothetical board approves up to $500 million of repurchases. During the next quarter the company spends $120 million. Assuming no other changes to the program, $380 million remains available. The three numbers answer different questions:

  • $500 million: the size of the approved program.
  • $120 million: the amount actually spent during the period.
  • $380 million: the unused capacity under that program.

Do not add all three and call the result buybacks. Also distinguish a completed program from one terminated or suspended without using its full authorization. Alphanume groups those endings under completion_or_termination; the accompanying description is needed to understand which ending occurred.

The SEC's Item 703 disclosure table separates shares purchased, average purchase price, purchases under publicly announced programs, and remaining authorization. Its footnotes identify program approval dates, approved amounts, expirations, and terminations. That is useful evidence when reconciling a company announcement with later activity.

Why an increase is easy to double-count

Suppose a hypothetical company first announces a $200 million program and later announces a $300 million program. The second statement could mean an extra $300 million, an increase of the total to $300 million, or a replacement authorization that supersedes what remained. Those interpretations produce different capacity estimates.

The Buybacks API contract preserves this ambiguity. For an increase, authorized_amount_usd records the increment when stated, otherwise the new total. replaces_prior_program distinguishes replacement from addition when the filing says which applies; null means it does not say. Read event_description, remaining_prior_amount_usd, and the filing before building a running balance.

An accelerated share repurchase agreement or a company trading plan may also implement an existing authorization. Counting the authorization, the agreement, and the plan as three independent commitments would overstate the total. The feed retains one action per filing, so it should not be treated as a complete ledger of every implementation detail in a complex announcement.

Actual purchases need not equal the drop in share count

Consider another hypothetical quarter. A company begins with 100 million shares, repurchases five million, and issues three million through compensation or other transactions. It ends with 98 million. Net shares fell by two million even though gross repurchases were five million.

A share-count comparison alone cannot recover those flows. Historical Market Cap supplies dated company-size and share-count context, while Dilution helps identify relevant financing disclosures. Neither makes every unexplained share-count change a buyback or an offering. Reconcile the issuer's statement of equity, cash-flow disclosures, and repurchase table on compatible dates and share bases.

A useful buyback research screen

A concrete first screen is new authorizations worth at least 5% of company market value over a stated period. On GET /v1/capital/buybacks, use event_type=new_authorization, a date range, and min_pct_of_market_cap=5. Here, 5 means 5%, not 0.05. The ratio uses market cap on the last session before filing, which can differ from market cap before the announced event date.

For each result, retain the issuer's CIK, accession_no, announcement date, filed_at, amount, replacement flag, expiry, and description. A missing amount is unknown rather than zero; minimum-size filters exclude records without the required amount or ratio. The dataset covers qualifying actions in SEC 8-K filings from January 2024 and excludes routine quarterly purchase reports, so this screen finds announcement events rather than all companies that repurchased shares.

Keep the historical clock honest

event_date is the stated announcement date, with documented fallbacks. filed_at is when EDGAR received the source filing. Neither proves that an API user received the extracted row at that moment: historical records were backfilled, and corrections can arrive later.

For a market-reaction study, establish the earliest public disclosure using the underlying documents and obtain separate prices. For a study of a live data-driven process, save retrieval times and raw responses as the process runs. Then compare later purchase disclosures with the original authorization. That sequence answers a more useful question than whether the headline sounded bullish: how much capacity was genuinely new, how much was eventually used, and how the share count changed afterward.