Why US ETFs Rarely Distribute Capital Gains

The mechanism, step by step

Step 1: the mutual fund's problem. Investors redeem; the manager sells appreciated holdings to raise cash; those sales realize gains; US tax law (the RIC rules) requires the fund to distribute realized net gains annually; the distribution is taxable to whoever still owns the fund on the record date. Long-term holders thus pay tax triggered by strangers' decisions, in a year they may have been down.

Step 2: the ETF's structural alternative. ETF redemptions happen through Chapter 31's window: an authorized participant delivers ETF shares and receives a basket of securities. No securities are sold, so no gains are realized inside the fund.

Step 3: the tax provision that makes it powerful. Under the relevant Internal Revenue Code provision governing regulated investment companies, a fund's distribution of appreciated property in redemption of its own shares does not trigger gain recognition at fund level. Combine that with the manager's ability to select which lots go into the redemption basket, and the fund can systematically send out its lowest-basis shares: the ones carrying the largest embedded gains.

Step 4: the result compounding over years. The fund's unrealized gain position is continuously scrubbed, its holdings' average basis stays high, and the shareholder's tax deferral runs uninterrupted until they choose to sell: at which point it's their own gain, at their own timing, at long-term rates, potentially reaching the step-up at death (Chapter 52).

What it's worth

The measurable evidence: in a typical year, the overwhelming majority of US equity ETFs distribute zero capital gains, while a large share of US equity mutual funds distribute something, sometimes several percent of NAV; and painfully, distributions have occurred in down years, when redemption pressure forces selling into losses for the fund's price while realizing gains from old, low-basis lots.

Estimating the value: a taxable investor avoiding, say, 1–2% of NAV in annual gain distributions at long-term rates saves roughly 0.2–0.5% per year in tax drag, compounding across decades. Academic and practitioner studies of ETF "tax alpha" land in a similar range for equity strategies. Set beside the fee differences investors obsess over, this is one of the largest single advantages available in a US taxable account; and it is entirely a wrapper effect, identical index inside.

Where the mechanism doesn't work

Honest boundaries, because the advantage is often oversold:

• Bond ETFs. Cash creations and redemptions are common (Chapter 31), so gains do get realized; bond ETFs distribute capital gains more often than equity ETFs, and their income is ordinary anyway.
• Commodity and currency products. Different structures (grantor trusts, partnerships) with their own regimes (Chapter 52's collectibles and K-1 notes).
• Active ETFs with high turnover. In-kind redemptions scrub gains that arise from flows, not gains from the manager's own trading. A strategy turning over 100% a year generates realized gains no basket can fully absorb.
• Inside tax-advantaged accounts. Worth nothing at all: no tax to defer (Chapter 28's location logic).
• Outside the United States. This is the crucial one for most of this book's readers: the mechanism is a US tax-code artifact. European and most Asian investors are taxed under their own national systems (Chapter 53), where a fund's internal realizations usually aren't the taxable trigger in the first place. UCITS ETFs do use in-kind creation/redemption for operational efficiency, and it helps tracking, but the tax miracle does not travel. Any article promising Europeans "ETFs are more tax-efficient" on these grounds is importing a foreign statute.

The heartbeat criticism, addressed

Some funds have used "heartbeat trades": a short-lived, pre-arranged inflow of cash or securities from a partner, followed by an outsized in-kind redemption that carries out low-basis positions: to scrub gains beyond what organic flows would allow, particularly around index rebalances or in active funds. Journalists have documented the practice, and it has drawn Congressional and academic attention; proposals to curb in-kind treatment have surfaced periodically without becoming law.

The balanced reading: the core mechanism is a decades-old, plainly-legislated feature of how redemptions in kind work, and it benefits ordinary index-fund holders enormously. Heartbeat trades are an aggressive optimization at the edge of that feature, concentrated in a minority of funds, and they are the part of this story with genuine policy risk attached. An investor's practical exposure to that risk is modest, since a rule change would apply prospectively to future gains, but pretending the criticism is baseless would violate this book's standards. It isn't baseless; it's specific, and it's about the edge, not the engine.

A worked example

Two American investors, each $200,000 in an S&P 500 strategy, 20 years, taxable accounts.

• Ellen holds the index mutual fund. Over the period her fund distributes capital gains in eleven of twenty years, averaging ~1.3% of NAV. She pays tax each of those years at long-term rates, from money that would otherwise have compounded.
• Frank holds the index ETF. Capital-gains distributions: zero. He pays tax on dividends, like Ellen, and nothing else until he sells; and he sells nothing, donating appreciated lots to charity in retirement and leaving the rest to a step-up.

Same index, same fees to within a basis point, same market. The difference: roughly 0.3%/yr of tax drag compounding, worth a meaningful five-figure sum on Ellen's original stake: created entirely by which wrapper each of them chose in year one.