ETF Liquidity: Why AUM Isn't Liquidity

The three layers of ETF liquidity

Professionals think in layers, and retail screens show only the first.

Layer 1: displayed liquidity. The bid and ask on screen with their sizes. This is what a retail order meets, and it is a fraction of what exists. Market makers show a slice and replenish continuously.

Layer 2: undisplayed market-maker capacity. The size a market maker will actually trade at or near the quote, which it does not display because doing so would reveal its hand. Accessing it requires asking, via a broker's ETF desk or an RFQ platform (Chapter 72), rather than clicking.

Layer 3: implied liquidity from the basket. The true depth. If a market maker can buy the basket, it can create ETF shares to sell you. So the ETF's capacity approximates the basket's capacity, adjusted for how much of the underlying can be traded without moving prices. Providers publish "implied liquidity" estimates using exactly this logic: typically the number of creation units that could be assembled while trading no more than some percentage of each constituent's average daily volume.

Why AUM and volume mislead

AUM measures popularity and viability, not tradability. It matters for closure risk (Chapter 67) and for the fund's ability to spread fixed costs. It says nothing about whether you can buy €100,000 today at a fair price, which depends entirely on what the fund holds.

Volume measures yesterday's interest. A fund that traded nothing yesterday can trade heavily today at a tight spread, because the market maker's inventory is created rather than found. This is the single most counterintuitive fact in the ETF market, and it is why "low volume, avoid", sensible advice for individual stocks, transfers badly to ETFs.

The correlation exists but is spurious. Large, high-volume funds do tend to have tighter spreads, because they tend to hold liquid things and attract more market makers. Confusing that correlation for causation leads investors to reject perfectly tradable niche funds and to over-trust large funds holding illiquid assets. The second error is the more expensive one.

What to look at instead

1. The spread, in percentage terms, at the hour you would trade with the underlying open (Chapters 12, 57). This is the market's own summary of everything above, priced in basis points, and it is the single best liquidity indicator available to a retail investor.

2. The basket's composition. Look through to the holdings: mega-caps, government bonds, and developed-market equities are deeply liquid; micro-caps, frontier markets, off-the-run credit and single commodities are not. The fund inherits its holdings' tradability.

3. Implied liquidity, published by several providers and available on institutional platforms: usually expressed as a daily tradable amount consistent with limited participation in each constituent's volume.

4. Premium/discount history (Chapter 33): a fund whose gap breathes in a wide band is telling you its basket is expensive to convert.

5. For large orders, ask. A quote from an ETF desk reveals Layers 2 and 3 directly (Chapter 72).

Where the argument stops

This chapter is a correction of a myth, not a claim that liquidity is infinite. The honest boundaries:

The basket is the ceiling. If the underlying cannot be traded, neither can the ETF at a fair price. A fund holding illiquid high-yield bonds is exactly as illiquid as those bonds, and its arbitrage mechanism can only pass that reality through: it cannot conjure liquidity that does not exist. This is the essential grain of truth inside the "liquidity illusion" critique (Chapter 87).

Capacity relative to the underlying matters. When a fund grows large relative to its constituents' float: a niche thematic fund holding small caps is the standard case: its own flows move its holdings, in both directions. Size becomes a liquidity problem rather than a comfort, inverting the intuition entirely (Chapter 47).

Stress compresses the layers. In a one-directional panic, secondary-market netting (Chapter 68) does less work, more flow reaches the basket, and Layer 3 does the pricing. The ETF still trades; the price simply reflects what the underlying is genuinely worth that day, which observers sometimes mistake for the ETF failing (Chapters 33, 84).

Halts and closures are real constraints. If the underlying market is shut, the ETF prices a forecast and the spread widens to compensate (Chapter 12).

A worked example

Two European UCITS funds, an institution needs to trade €30m in each.

• Fund A: €180m AUM, average daily volume €400,000. Holdings: 500 US mega-caps. Screen suggests "far too small." The desk quotes NAV +4bp for the full €30m, because the market maker simply buys the basket and creates. Executed in one print.
• Fund B: €4.2bn AUM, average daily volume €9m. Holdings: 60 small-cap companies in a single thematic sector. Screen suggests "plenty of liquidity." The desk quotes NAV +85bp and warns of impact, because assembling the basket means moving the prices of the constituents themselves. Executed over three days.

AUM ratio: 23-to-1 in Fund B's favour. Actual cost ratio: 21-to-1 in Fund A's favour. Everything a screener would have told this institution was backwards, and the reason is that a screener reports the ticker while the cost lives in the basket.