ETF Liquidity Metrics: Amihud Ratio, Spreads and Creation Flow

Measure ETF liquidity in three steps: quote the bid-ask spread, compute the Amihud illiquidity ratio, then read creation and redemption flow to forecast your trading cost.

2026-11-18 · 10 min read

Key takeaways

  • Amihud's ILLIQ turns daily returns and dollar volume into an estimate of price impact, but it says nothing about the spread you actually pay.
  • In the US, Rule 6c-11 requires every ETF website to publish its median 30-day bid-ask spread, which makes spread data free and comparable.
  • Creation and redemption data show whether the primary market can absorb your order; heavy concentration in a few authorized participants is a warning sign.
  • Read the three metrics together: a wide spread plus one-sided creation flow is a costlier trade than any single metric suggests.

Why one liquidity number is never enough

ETF liquidity is not a single quantity. It is three separate things: the price you pay to enter the market, the price impact of your size, and the reliability of the primary market that restocks or absorbs shares behind the ticker.

Those three can disagree. A fund can show a tight quoted spread and still deliver a poor fill on a 5,000-share block. Another can trade thousands of times a day while every trade is market-maker round-tripping rather than genuine investor demand.

This framework measures each layer separately, then combines them into an expected trading cost. The building blocks — spread data and creation/redemption files — are partly public, which matters because a portfolio manager who can only see tape volume is working with the weakest of the available signals.

The three layers of ETF liquidity

  1. Secondary-market cost: bid-ask spread and market depth on the listing venue.
  2. Price impact: how much the price moves per unit of volume traded — this is where the Amihud illiquidity ratio comes in.
  3. Primary-market capacity: the ability of authorized participants (APs) to create or redeem baskets, which ultimately decides how much size the fund can absorb in a day.

For background on how these layers interact, see our pieces on ETF Liquidity Provision: Authorized Participants and Market Makers and Bid-Ask Spread and Market Depth: A Practical Guide to Execution Costs.

Step 1: Anchor on the bid-ask spread

The quoted spread is the immediate, observable cost of a round trip. Express it as a percentage of the mid price, not in dollars: a one-cent spread on a $25 ETF is eight times more expensive in relative terms than a one-cent spread on a $200 ETF.

Quoted spread versus effective spread

The quoted spread is what you see. The effective spread is where your order actually executed relative to the mid at the time of arrival — usually wider than quoted for market orders, tighter for patient limit orders that get filled inside the quote. Track both if your order management system records arrival midpoints; otherwise the quoted spread is a usable upper bound.

What "normal" looks like

Spreads vary enormously by asset class and by whether the underlying market is open. Morningstar's analysis of US-listed actively managed equity ETFs put the median 30-day spread at 0.12% for large-cap funds and 0.20% for small-cap funds as of 19 September 2025, with the typical emerging-market equity ETF roughly double the large-cap figure (Morningstar).

For European listings, Dimensional's study of ETF listings found time-weighted average intraday spreads of roughly 15–17 basis points for US and European equity ETFs, against about 25–26 bps for sector and global equity ETFs and roughly 28 bps for global and emerging-market fixed income (Dimensional). The pattern is consistent: the harder the underlying basket is to assemble and hedge, the wider the fund trades.

The disclosure you can rely on — in the US

In the US, Rule 6c-11 requires every ETF relying on it to publish on its website, daily, its median bid-ask spread over the most recent 30 calendar days, along with historical premium and discount information (SEC small entity compliance guide, effective 23 December 2019). That single number is the fastest screening filter available: it is fund-specific, rolling, and comparable across issuers.

In the EU there is no equivalent uniform fund-website requirement, so European investors typically source spread data from trading venues and market-data vendors instead.

Practical filter: compare an ETF's published 30-day median spread with those of its closest competitors before looking at anything else. A persistent gap is a permanent, repeatable cost you pay on every round trip.

Step 2: Compute the Amihud illiquidity ratio

Spread tells you the cost of small trades. It tells you nothing about what happens when your order is a meaningful share of the day's activity. That is the price-impact problem.

The Amihud illiquidity ratio, introduced in Illiquidity and Stock Returns (Journal of Financial Markets, 2002), measures exactly that. It is the average daily ratio of the absolute return to dollar volume (Amihud, 2002):

ILLIQ = average over days of |R_t| / DollarVolume_t

where R_t is the day's return and DollarVolume_t is that day's traded value. In published work the ratio is normally multiplied by 10^6 for readability.

How to read the number

The ratio is the daily price response per unit of traded volume — a rough estimate of the price move a dollar of order flow buys. Higher means less liquid. Because it uses only daily returns and dollar volume, you can build long time series from any market data source, which is why the measure spread well beyond US equities.

For an ETF, run the calculation in three windows and compare them:

  • Full history (3–5 years): the structural liquidity baseline.
  • Trailing 60 trading days: current conditions.
  • Stress episodes: the weeks you would actually need liquidity, for example a volatility spike. Research on US exchanges showed spreads broadly doubling during the sharp market stress of April 2025 before settling back (Morningstar). A fund whose Amihud ratio stays flat through those windows has a genuinely resilient market, not just a tight quote on quiet days.

Where the measure misleads on ETFs

Three cautions matter when you apply a stock-liquidity measure to a fund:

  • Volume is not investor demand. ETF volume is inflated by market makers trading the same shares repeatedly, so dollar volume overstates the flow available at a given price.
  • The basket matters more than the ticker. An ETF's true depth comes from its underlying holdings and the cost of assembling a creation basket. A thinly traded-looking ETF holding the largest, most liquid names can still absorb size.
  • Low-volume days distort the average. Because the denominator is dollar volume, a single quiet day can dominate the ratio. Use medians alongside means, and require a minimum activity threshold before trusting the figure.

For a fuller set of impact and slippage measures, see Transaction Cost Analysis for Retail Traders: Tools and Metrics.

Step 3: Read creation and redemption activity

The primary market is what makes an ETF's liquidity larger than its secondary-market volume. When creation baskets are cheap to assemble and easy to hedge, market makers can quote tighter and bigger than the tape suggests.

Where the data comes from

In the US, creation and redemption processing runs through the National Securities Clearing Corporation. DTCC publishes daily Portfolio Composition Files showing each fund's creation-unit basket, and its ETF Kinetics service provides underlying basket creation volumes, redemption volumes and fund flow information on a T+3 basis (DTCC ETF Kinetics, DTCC ETF Portfolio Data). Annual gross creation and redemption volumes by AP are also visible in Form N-CEN filings.

Rule 6c-11 recordkeeping adds a useful audit trail: funds must retain, for each basket exchanged with an AP, the holdings and weights, whether the basket was custom, any cash balancing amount, and the identity of the AP (17 CFR 270.6c-11). That record is not published in real time, but it underpins the premium/discount history funds do disclose.

The signal to watch: AP concentration

A Federal Reserve analysis of N-CEN filings found that ETF primary-market activity is concentrated in about four active APs per fund, with concentration especially high for fixed income and international equity ETFs (Federal Reserve FEDS Notes, January 2023).

The practical reading:

  • Many active APs, two-way flow: the arbitrage mechanism is competitive, deviations from net asset value (NAV) close quickly, and quoted size holds up.
  • Few APs, one-directional flow: creations or redemptions are being absorbed by a small group. Deviations from NAV persist longer, and a bout of one-way investor flow can widen the market well beyond the 30-day median spread.

Custom baskets are worth understanding here. Rule 6c-11 lets funds accept baskets that are not pro-rata slices of the portfolio, subject to written policies and procedures and adviser review — flexibility the SEC expected to support more efficient arbitrage and narrower spreads, but which also makes basket composition less predictable to outside observers (SEC press release, 2019).

iNAV: useful, with caveats

Intraday indicative value (iNAV) is calculated from the fund's Portfolio Composition File and published roughly every 15 seconds during trading hours (Solactive, Tradeweb methodology). It is your live reference for premium and discount. But when part of the basket sits in a closed market, or a bond price comes from a model rather than a trade, the iNAV can be stale or indicative at best — Invesco warns against treating it as the sole execution reference (Invesco). For cross-border funds, watch the overlap window when both the ETF listing venue and the underlying market are open; that is where pricing is most reliable.

Putting the three metrics together

Layer What it captures Typical source Main caveat
30-day median bid-ask spread Round-trip cost at small size ETF website (US, Rule 6c-11); venue and vendor data in the EU 30-day window hides stress days
Amihud illiquidity ratio Price impact per dollar of volume Computed from daily returns and dollar volume Distorted by market-maker volume and quiet days
Creation/redemption flow and AP count Primary-market capacity, NAV convergence DTCC PCF and flows, Form N-CEN, fund premium/discount history Not real-time; custom baskets reduce transparency

A workable workflow: screen on the published median spread, confirm with a rolling Amihud ratio, then check whether the fund's AP base and flow pattern can support the size you intend to trade. If a fund screens well on the first two but shows one-directional flow through a handful of APs, size down or work the order. If you would rather compare the whole holding cost, see TER vs Total Cost of Ownership: What an ETF Really Costs.

Risks and costs

  • Spread is a cost on every round trip. In the US, a fund showing a 0.12% median 30-day spread costs roughly 0.12% of notional to enter and exit at that size — a figure worth comparing against the annual expense ratio for anyone trading frequently (Morningstar).
  • Impact scales non-linearly. Amihud ratios and depth estimates are averages; your specific order can move the market more than the average implies, especially when a large share of daily volume is market-maker round-tripping.
  • Spreads widen when you need liquidity most. Mid-April 2025 data for US large-cap active ETFs showed spreads roughly doubling at the widest quartile (Morningstar). Liquidity measured in calm markets is not the liquidity you will trade against in a drawdown.
  • Primary-market fragility. With primary activity concentrated in a few APs per fund (Federal Reserve), a fund can lose creation capacity quickly, at which point it behaves more like a closed-end fund: prices decouple from NAV and the discount closes on someone's book.
  • Operational and data costs. Basket-level data from DTCC is a paid subscription, and computing robust impact measures requires clean daily volume and return series. Free spread data does not extend to basket composition.
  • Jurisdiction and product differences. The US website-disclosure regime is a Rule 6c-11 feature; European UCITS ETFs and other structures are not covered by the same requirement, and settlement cycles, listing venues and share-class conventions differ. Rules described here are not universal.
  • Measure risk. Any metric built on historical data describes past conditions. If a fund's underlying holdings, index methodology or share class structure change, recompute rather than carry the old number forward. Related reading: ETF Creation and Redemption: Why Prices Diverge From NAV and Corporate Bond ETF Liquidity: Spreads, Premiums and Stress Risk.

Bottom line

ETF liquidity metrics answer three different questions, and mixing them up is the most common analytical error. The published median 30-day bid-ask spread tells you the round-trip cost of a normal-sized trade. The Amihud illiquidity ratio tells you how much price moves per dollar of volume, and it works for any market where you have daily returns and dollar volume — provided you correct for the way ETF volume overstates true demand. Creation and redemption activity, plus the number of active APs behind a fund, tells you whether the primary market can absorb or replenish shares when the secondary market tightens.

Score the three separately, then combine: the funds that hold up are the ones with a tight published spread, a stable impact ratio across both calm and stressed windows, and a broad, two-way AP base. This framework is a research method, not a recommendation — apply it to the specific instruments you are evaluating and check the current data yourself before trading.

This article is for information and education only. It is not investment advice and does not take into account your personal situation. Read the full disclaimer.

Follow the shortlist

Comparing ETFs on cost, liquidity and structure? Our global ETF and stock research watchlist is tracked daily. Premium members get the full list, the reasoning behind every entry and updates when entries change.

See premium access

Information and education only, not investment advice.

Related