A buyback (share repurchase) is when a company uses cash to buy its own shares from the market and retire them (or hold them as treasury stock), returning capital to shareholders by shrinking the share count rather than by paying a dividend. It is one of the two principal ways a firm distributes excess cash, and a core capital-allocation decision sitting alongside dividends, debt paydown, M&A, and reinvestment. Its central tension: a buyback mechanically lifts per-share metrics (EPS, ownership %) and can create real value if shares are bought below intrinsic value — but it creates no fundamental value on its own, and is routinely criticised for boosting optics (and executive pay) while companies tend to repurchase most heavily at high prices, not low ones.

How it's done — the methods

Open-market repurchase (OMR) — the company buys shares on the exchange at prevailing prices over time under a board-authorised program. This is the dominant form: it accounts for more than 95% of buyback programs worldwide (Corporate Finance Institute; CFA curriculum). A program authorisation (e.g. "$10B buyback") is a ceiling, not a commitment — management can buy little or none of it, which matters greatly for interpreting announcements.

Fixed-price / Dutch-auction tender offer — the company offers to buy a set number of shares directly from holders, usually at a premium to market, over a short window. Faster and a stronger signal, but more expensive and administratively heavier; often used in restructurings or takeover defence.

Accelerated share repurchase (ASR) — the company pays an investment bank upfront for a large block of shares (borrowed by the bank), retiring them immediately; the final price settles later via a forward contract tied to the average price over the period. Used when management wants to shrink the count now.

Mechanics of EPS accretion. Buying shares reduces shares outstanding, raising EPS — but the cash spent (or interest on debt raised to fund it) reduces net income or forgone interest income. A repurchase is accretive only when the earnings yield (E/P) on the stock exceeds the after-tax return the cash would otherwise earn (McKinsey). It is value-neutral in EPS terms if the stock is bought at a P/E equal to that hurdle.

How it's used in practice

• Returning excess cash flexibly. Unlike a dividend, a buyback carries no implicit ongoing commitment — it can be dialled up or down without the market penalty that a dividend cut triggers. This makes it the preferred vehicle for distributing variable or one-off cash (e.g. windfall years, repatriated cash).
• Tax efficiency (historically). In the US, returning cash via buyback let shareholders defer/convert ordinary-income dividend tax into deferred capital gains. The Inflation Reduction Act introduced a 1% excise tax on net repurchases effective Jan 1, 2023, narrowing but not erasing that edge (Congress.gov CRS R47397).
• Anti-dilution. Mature firms (especially tech) use buybacks to mop up shares issued through stock options and RSUs, keeping the count roughly flat. Here the "buyback" is really an offset to stock-based compensation, not net capital return — a distinction analysts must net out.
• Signalling undervaluation. Management buying its own stock is read as "we think it's cheap." A gross buyback that genuinely shrinks the count carries more signal than one that merely absorbs SBC.
• What analysts watch: net share count change (diluted shares, year over year), buyback yield (repurchases ÷ market cap), price paid vs. estimated intrinsic value, and whether buybacks are funded by free cash flow or by debt.

Adoption, debate & evidence

Buybacks are now the largest channel of US corporate cash return — per S&P Dow Jones Indices, S&P 500 repurchases set a calendar-year record of ~$923B in 2022, then fell ~13.8% to ~$795B in 2023 (on a rolling-12-month basis the figure briefly topped $1T mid-2022; the 2023 decline is partly attributed to the IRA excise tax). They are heavily contested.

The bull case has real evidence — but it is conditional. Comment & Jarrell (1991) document positive announcement-period abnormal returns rising with the percent of shares sought and falling with recent stock returns — consistent with an undervaluation signal. Ikenberry, Lakonishok & Vermaelen (1995) found a positive four-year buy-and-hold abnormal return averaging ~12.1% after OMR announcements — but concentrated entirely in "value" stocks (~45.3%) with no drift for "glamour" stocks. The edge is in cheap firms buying cheap stock, not buybacks generically.

The critique is also evidence-backed. McKinsey's analysis shows buyback-driven EPS growth correlates poorly with total shareholder return; TSR is driven by revenue, earnings, and especially ROIC — buybacks "boost earnings without improving returns." Empirically, companies tend to repurchase most aggressively when cash is plentiful and stocks are high (pro-cyclical), and cut buybacks in downturns when shares are cheap — the opposite of value-creating timing. Mashruwala (2025, JBFA) finds firms are far more likely to do EPS-accretive repurchases when they would otherwise just miss analyst EPS forecasts, and that such repurchases transfer wealth from continuing to selling shareholders. Lazonick's influential 2014 HBR critique frames buybacks as diverting cash from R&D and labour to inflate EPS-linked executive pay.

Honest synthesis: a buyback is value-creating if and only if shares are bought below intrinsic value with cash that has no higher-returning alternative. Absent that, it is value-neutral at best, and the "accretive EPS" is an accounting artefact, not value creation.

Strengths & limitations

Strengths: flexible, tax-deferred-historically capital return; legitimate undervaluation signal from value firms; offsets dilution; concentrates ownership for staying shareholders.

Limitations / failure modes: creates no fundamental value by itself; firms systematically buy high and stop low (pro-cyclical mistiming); debt-funded buybacks raise leverage and fragility; can mask SBC dilution if only "net" is reported; can starve reinvestment. The single most common misuse — by management and analysts alike — is treating mechanical EPS accretion as value creation, ignoring the price paid relative to intrinsic value. The second is conflating a gross buyback with a net reduction in share count when much of it merely absorbs option/RSU issuance.

Sources

• McKinsey — How share repurchases boost earnings without improving returns (EPS vs. TSR/ROIC; pro-cyclical timing).
• Ikenberry, Lakonishok & Vermaelen (1995), Market Underreaction to Open Market Share Repurchases (NBER w4965 / JFE) — 4-year drift ~12.1%, value vs. glamour split (45.3% / ~0%).
• Comment & Jarrell (1991) — announcement abnormal returns and the undervaluation signal.
• Mashruwala (2025), EPS-motivated share repurchases and wealth transfer, Journal of Business Finance & Accounting — accretion concentrated around EPS-forecast misses; wealth transfer to selling shareholders.
• Corporate Finance Institute; CFA Level 1 (AnalystPrep) — repurchase methods; OMR >95% share.
• Congress.gov CRS R47397 / EY / Deloitte — 1% IRA excise tax (IRC §4501) on net repurchases, effective Jan 1, 2023.
• S&P Dow Jones Indices, S&P 500 buyback press releases (Q4 2022, FY 2023) — record ~$923B in 2022, ~$795B in 2023 (-13.8%).
• Lazonick, Profits Without Prosperity, Harvard Business Review (2014) — buybacks-vs-investment / executive-pay critique (contested).

Flag: the academic positive-drift figures (Ikenberry, Comment & Jarrell) are robust but vintage and concentrated in value names; do not generalise to glamour/expensive stocks. The Lazonick critique is influential but politically contested — treat as a hypothesis, not settled fact.