An exchange-traded fund has two prices at once: the net asset value (NAV) — the per-share market value of the securities it actually holds — and the market price at which its shares change hands on the exchange. When the market price sits above NAV the fund trades at a premium; when it sits below, a discount. The premium/discount is the percentage gap between them, and it is the single cleanest read on whether an ETF's pooled-ownership wrapper is currently being priced fairly against its contents. The core tension is that NAV is a valuation (computed, often once a day) while the market price is a negotiation (continuous, demand-driven); the two are kept close by an arbitrage mechanism that is powerful but not infinitely strong.

How it's calculated / formed

The standard formula is identical for ETFs and closed-end funds:

> Premium/Discount % = (Market Price ÷ NAV) − 1

A fund whose shares trade at $19 against a $20 NAV shows a −5.0% discount; at $20.40 against $20, a +2.0% premium (Fidelity). Official NAV is struck once per day, after the close, using last-traded prices of the holdings. To bridge the gap during the session, ETFs also publish an intraday indicative value (iNAV / IIV), recalculated roughly every 15 seconds from the published creation basket. The SEC's Rule 6c-11 (the 2019 "ETF Rule") requires funds to post daily premium/discount data plus a rolling historical distribution on their websites; it also triggers a special website disclosure with an explanation of contributing factors whenever a fund's premium/discount exceeds 2% for more than seven consecutive trading days (SEC small-entity compliance guide). This is why the metric is so easy to monitor.

Premiums and discounts form whenever order flow pushes the market price away from fair value faster than supply can adjust — a wave of buyers lifts the price to a premium; concentrated selling opens a discount.

How it's used in practice

For most large, liquid ETFs (broad U.S. equity funds, major sector funds) the premium/discount is a transaction-quality check, not a signal: a buyer wants to avoid paying a premium or selling into a discount, especially around the open and close when spreads are widest. The practical guidance from issuers — Vanguard, Fidelity, State Street — is to use limit orders, trade during liquid hours, and check the live premium before sending a marketable order.

The keeper of the peg is the creation/redemption mechanism. Authorized Participants (APs) — large broker-dealers — can exchange ETF shares for the underlying basket and vice versa. When an ETF trades at a premium, an AP buys the underlying securities, delivers them to the fund for newly created shares, and sells those shares into the rich market, pocketing the spread; this adds supply and pushes price down toward NAV. At a discount the trade runs in reverse: buy cheap ETF shares, redeem them for the more-valuable basket. This arbitrage is what makes ETFs' deviations normally small and short-lived — the structural feature that distinguishes them from closed-end funds.

A persistent, structural premium or discount is more interpretable than a fleeting one. It usually signals that arbitrage is impaired — and the reasons (stale NAV, closed underlying market, illiquid holdings) tell you something real about the wrapper and its contents.

Adoption, debate & evidence

That premium/discount is real and measurable is not contested; the debate is over which deviations are noise and which are information.

• Domestic, liquid ETFs: deviations are small and transitory. Empirical work (e.g. Petajisto and others) confirms that funds holding liquid, U.S.-traded securities track NAV tightly because arbitrage is cheap and fast.
• International ETFs: the arbitrage barrier is structural. When the underlying market (e.g. Tokyo, Mumbai) is closed while the ETF trades in New York, the ETF price reflects newer information than the stale NAV, producing apparent premiums/discounts that are partly an artifact of non-overlapping hours. Research on international and emerging-market ETFs finds higher and more persistent deviations than domestic funds, attributed to these arbitrage frictions (ScienceDirect, Premiums and discounts in ETFs; ScienceDirect, feedback trading in emerging-market ETFs).
• Bond ETFs: the most cited stress case is March 2020, when several large investment-grade and high-yield bond ETFs traded at sharp discounts. Magnitudes vary by measure: average absolute deviations widened only into the ~1% range for LQD/HYG over March–April 2020, but the largest closing-basis discount on the IG fund LQD was reported near −5% at the mid-March peak (SEC Fixed Income Advisory Committee materials). The contested interpretation: many argued the ETF price was the truer mark and the published NAV was stale — built from dealer quotes for bonds that hadn't traded — so the "discount" was really NAV lagging a falling market. This episode is the textbook illustration that NAV is not gospel for illiquid underlyings.
• Closed-end funds (the contrast case): CEFs have a fixed share count and no creation/redemption arbitrage, so discounts can persist for years. The academic "closed-end fund puzzle" — discounts that exist at all, vary over time, and don't reliably close — remains unresolved, with leading explanations spanning investor sentiment (Lee–Shleifer–Thaler), distribution policy, embedded tax liabilities, and ownership concentration. Studies have found average discounts materially deeper for funds with large blockholders (one cited figure ~14% vs ~4% without) — but the precise figures are sample-specific and should be treated as illustrative, not universal.

Strengths & limitations

Where it works: Premium/discount is an honest, real-time, formula-clean indicator of execution quality and of arbitrage health. A sudden, large deviation in a normally tight ETF is a genuine red flag — it often means trading is dislocated or the AP mechanism has seized.

Where it fails / the #1 misuse: Treating it as a cheapness signal. For ETFs it is not a value play — the arbitrage mechanism is designed to erase it, so a discount is usually either noise or a warning about stale/illiquid pricing, not an opportunity. The deeper trap is comparing the live ETF price to a stale NAV (international funds during closed hours, bond funds in a fast market) and concluding the fund is "cheap" when in fact the NAV is simply old. For CEFs, the related error is assuming a discount must revert — nothing forces convergence to NAV, and discounts can persist or deepen indefinitely (Fidelity).

Sources

• Fidelity — CEF Discounts and Premiums (formula; persistence; no mandated convergence)
• Fidelity — ETF Premiums and Discounts
• Vanguard — Navigating ETF premiums and discounts (execution guidance)
• State Street — Mechanics of ETF trading (AP arbitrage)
• Invesco — ETF arbitrage, premiums and discounts
• SEC — Exchange-Traded Funds: Small Entity Compliance Guide (Rule 6c-11) (2%/7-day triggered disclosure)
• SEC Fixed Income Advisory Committee — Pricing and Liquidity of Fixed Income ETFs in the COVID-19 Crisis of 2020 (March 2020 deviation magnitudes)
• ScienceDirect — Premiums and discounts in ETFs: arbitrage in domestic and international funds (domestic vs international persistence)
• ScienceDirect — Premiums, discounts and feedback trading: emerging-market ETFs
• Lee, Shleifer & Thaler / UT-Dallas — Persistence and Predictability of CEF Discounts (closed-end fund puzzle, sentiment)

Disputes flagged: (1) The interpretation of March 2020 bond-ETF discounts — "ETF mispriced" vs "NAV stale" — is genuinely contested; cited magnitudes are approximate and report-dependent. (2) The closed-end fund discount puzzle has no consensus cause; the blockholder discount figures (~14% vs ~4%) are sample-specific and illustrative only.