Liquidity, spreads and orders

An ETF quotes €99.90 to buy and €100.10 to sell. On screen its price looks like €100, but that number is not available for both operations: whoever sells immediately meets 99.90; whoever buys immediately meets 100.10.

That distance is the bid-ask spread. It does not appear in the TER, it does not arrive as a separate invoice, and it can change while you prepare the order. It is one of the most visible transaction costs — but not the only one.

To understand whether an ETF is genuinely tradable, it is not enough to look at how many units changed hands today. You need at least five elements:

• the buy and sell proposals available;
• the sizes attached to each price;
• the liquidity and the hours of the underlying market;
• the intermediaries’ capacity to hedge and transfer risk;
• the way the order is placed and executed.

The right question, then, is not just “how much does this ETF trade?”, but: at what price, and for what size, can the order be executed — on this venue, at this moment?

An ETF’s liquidity has several layers

For a share, the liquidity observed on the secondary market is already a central part of the answer. An ETF adds another architecture: authorized participants and market makers can connect the units traded on exchange to the underlying portfolio through creations, redemptions and hedges, as seen in chapter 2.

Liquidity should therefore be read on three levels.

These levels communicate, but they are not interchangeable. A rich book does not guarantee it will stay that way through a shock; a low daily volume does not prove, on its own, that a larger order cannot be executed; a liquid underlying does not ensure a tight spread on every listing and at every hour.

Liquidity is not a fixed property of the fund’s name. It is a market condition referring to an instrument, a venue, an instant and a size.

Volume does not mean liquidity

Volume measures what has already been traded over an interval. Liquidity is also about what could be traded now, at reasonable prices and without moving the market excessively.

An ETF may show few trades simply because no investor needed to trade it. If the underlying is liquid and several intermediaries are willing to quote, the book can still offer competitive prices. Conversely, a high volume observed yesterday does not promise the same capacity today, especially when volatility rises or the underlying market closes.

AUM tells yet another story. Assets measure the fund’s economic size; volume measures the trades that occurred; spread and depth describe the conditions immediately tradable. The quantities can be correlated, but none substitutes for the others.

Four snapshots, four different questions:

Bid, ask and spread

The bid is the best visible proposal from those who want to buy. The ask is the best visible proposal from those who want to sell. In normal conditions the bid is below the ask.

Absolute spread = ask − bid

To compare ETFs with different unit prices, the percentage spread relative to the midpoint is more useful:

Midpoint = (ask + bid) ÷ 2
Percentage spread = [(ask − bid) ÷ midpoint] × 100

In the opening example:

• bid = €99.90;
• ask = €100.10;
• midpoint = €100;
• spread = €0.20, that is 0.20%.

Whoever buys 100 units at the ask pays €10 more than the observed midpoint; whoever sold them back immediately to the same book would collect another €10 less. The theoretical cost of the full round trip would therefore be the entire spread, before commissions and market moves.

The midpoint, however, is a reference — not necessarily an executable price. The quoted spread is a photograph of the best proposals; the actual cost depends on the size executed, the levels crossed and what happens while the order travels to the market.

The depth of the book and market impact

The best bid and the best ask show only the first step. The book — the ordered list of buy and sell proposals — can hold different sizes at successive prices. Imagine this sell side:

A market buy order for 100 units could be met entirely by the first level. An order for 400 units, if nothing changes, would consume the first two levels and part of the third:

• 100 units at €100.10;
• 250 units at €100.15;
• 50 units at €100.25.

The average price would be €100.15, above the best ask seen at the start. The difference produced by walking through the levels is a form of market impact.

The example is deliberately static. In reality, proposals can be added, cancelled or modified; part of the liquidity may not be visible; simultaneous orders can arrive first. The book is not a promise, but a photograph revocable under the market’s rules.

Where the spread comes from

The market maker does not quote in the abstract. It tries to estimate the ETF’s value and the cost of hedging or transferring the risk taken on. The differential can include:

• spreads and transaction costs of the underlying securities;
• volatility and the speed at which prices change;
• the cost of currency hedging or hedging through derivatives;
• taxes, market fees, settlement and funding;
• the size of the order and the risk of temporarily holding inventory;
• competition among market makers and other participants;
• the quality and timeliness of the information available.

That is why a bond ETF’s spread does not depend only on the trades visible on the unit. If many bonds in the portfolio trade rarely, the market maker must price in the uncertainty about their valuation and about the real cost of a hedge.

The same ETF can show a different spread on two venues, or at two nearby moments. It is not a contradiction: they are markets with their own orders, participants, hours and trading currencies.

When the underlying market is closed

A European ETF can give exposure to US or Asian equities, or to bonds traded mostly outside the hours of the exchange where the unit is listed. Trading in the ETF can continue even when part of the underlying is closed.

At that moment the price does not automatically become wrong. Market makers can use futures, currencies, related securities and other information to estimate current value. Uncertainty rises, though: the hedge can be less direct and the spread can widen.

When the hours overlap, the availability of prices and hedging instruments can improve. It is not a universal rule: auctions, macroeconomic announcements, shocks and volatility spikes can worsen conditions even during the overlap.

The correct formulation is not “never trade when the underlying is closed”, but: before the order, check whether the market that determines the value, and the instruments used to hedge it, are open and liquid.

This is general information about execution quality, not a recommendation to buy or sell any specific ETF.

Open, close and auctions

In the opening minutes the market must absorb the news accumulated during the close and align many new proposals. Towards the end of the session, flows, rebalancings and orders tied to closing prices can concentrate. Opening and closing auctions follow different rules from continuous trading.

Hence the simplistic advice to always avoid the open and the close. It is too absolute. In some markets the auction concentrates a lot of liquidity and produces a robust price; in other cases a retail investor may not know the final price before the matching. The useful question is which market phase the order is using, and with which price protections.

To read a data point correctly you need at least:

• the phase of the session;
• the timestamp;
• the trading venue;
• whether the relevant underlying is open or closed;
• the volatility of the moment;
• the size attached to the prices shown.

One class, several listings

The same ISIN can be traded on several exchanges, with different tickers and currencies. The liquidity observed can therefore be fragmented.

Comparing only the volume of a single venue risks ignoring the trades that happened elsewhere. Mechanically adding up all venues can be just as misleading if the data have different hours, definitions or quality. A consolidated volume requires consistent venue coverage and documented rules.

The trading currency does not, by itself, change the portfolio’s currency exposure, as seen in chapter 3. It can, however, add a conversion cost applied by the broker to the individual investor, if the order’s currency differs from the account’s.

Ticker, currency and exchange identify the listing; the ISIN identifies the class of the instrument. Evaluating an execution requires both levels.

Market order and limit order

The order type changes the risk the investor takes on. Exact names and features can vary across brokers and venues; read the applicable documentation before using them.

Market order. It asks to execute against the best proposals available, without setting a maximum buying price or a minimum selling price. It offers priority to execution, not certainty on price. If the size exceeds the first level, the book changes quickly or liquidity is scarce, the average price can drift from the last price seen.

Limit order. It sets the maximum price accepted to buy, or the minimum price accepted to sell. It protects against crossing the limit, but it does not guarantee execution. The order can remain unfilled, or be filled only in part. Its behaviour depends on where it sits relative to the proposals on the book.

The marketable limit order. A limit order does not have to be placed to wait for a better price. It can also be set to meet immediately the proposals already on the book, while still keeping a maximum buying price or a minimum selling price. In that case it is called a marketable limit order:

• to buy, the limit is at or above the best available ask;
• to sell, the limit is at or below the best available bid.

Take the book from the opening example: best bid at €99.90 and best ask at €100.10. A buy order with a limit at €100.12 can be executed immediately at €100.10, if that proposal is still available. If the requested size exceeds what is offered at the first level, the order can also meet the next levels — but it cannot buy above €100.12.

The limit is not necessarily the price paid: it is the worst price the investor declares acceptable. If better prices are available on the book, the order may be executed at those prices without crossing the stated threshold.

The same principle works when selling. With a bid at €99.90, an order with a limit at €99.88 can be executed at €99.90 or at a higher price, but not below €99.88.

This protection is particularly relevant when the spread is wide or unstable, the book is shallow, the size exceeds what is available at the first level, volatility is high, or the underlying market is not fully open.

It does not, however, guarantee a good price in any absolute sense. A limit can be set badly, refer to a stale quotation or sit very far from the ETF’s current value. Setting it correctly requires observing bid, ask, sizes, venue and timestamp — not just the last traded price.

Rule of thumb. If immediate execution is indispensable, the market order privileges speed. If there is a price beyond which you would rather not complete the operation, that price should become the order’s limit.

Stop and stop-limit orders. A stop order activates when a threshold is reached and, under the applicable rules, can turn into a market order. In a fast move, the execution price can be very different from the threshold. A stop-limit adds a price limit: it reduces the risk of executing beyond that level, but increases the risk of not executing at all. On ETFs with a thin book, or during market gaps, the distinction is substantial.

Duration and conditions. Orders valid for the day, until cancelled or until a date; full or partial execution; minimum quantity; participation in auctions: each condition changes the order’s behaviour. The broker’s labels are not enough if you do not know the rule of the venue the order will be sent to.

Describing protections without giving a recommendation

Saying “always use a limit order” would be wrong for two reasons. First, it turns an educational principle into a prescription independent of goal, urgency, size and market. Second, it hides the cost of missed execution.

The correct distinction is this:

There is no order type that simultaneously removes price risk, missed-execution risk and time risk. There is a choice between different risks, to be understood before sending the order.

The transaction cost is more than the spread

The TER reduces the fund’s assets over time. Transaction costs hit the purchase or the sale instead, and they also depend on the investor, the broker and the moment.

An ETF with a lower TER is not necessarily the cheaper one for a small or occasional operation; a momentarily wider spread does not prove it will be more expensive over a long horizon. One-off and recurring costs must be separated, then related to amount and duration.

Chapter 6 will build the total cost: TER, tracking difference, the portfolio’s internal costs and the costs borne by the investor. Here the point is simpler: the price shown does not necessarily coincide with the average price obtainable for the desired size.

Quoted price, executed price and implementation shortfall

To evaluate an execution, comparing it with the last trade is not enough. That price could be old, refer to a minimal size or to another market phase.

Professional analysis also uses the implementation shortfall: the difference between the theoretical result associated with a decision price and the result actually obtained, including executions, commissions, delays and unexecuted quantities.

For a single retail order there is no need to turn it into a complex model. It is enough to keep:

• the time of the decision and of the submission;
• venue and currency;
• the bid, ask and available sizes observed;
• the order’s type, limit and duration;
• the quantity requested and executed;
• the prices of the individual executions and the commissions.

Without these elements, judging the broker or the order’s quality after the fact, on the basis of a single closing price, is fragile.

How to read the data on Rebalix

Rebalix distinguishes the instrument from its listings. The information comes from the ESMA register, the exchanges’ official lists and the issuers. For each venue, the fund page can show the exchange’s name derived from the MIC, the ticker, the trading currency when it comes from the official list — otherwise “—” — and the trading start date. Non-EU listings, for example those in the United Kingdom, are not included.

Rebalix does not currently show prices, spreads or book depth in real time, and does not calculate a premium or discount from temporally incompatible observations. A future trading data point will have to state at least source, venue, timestamp, currency, the nature of the price and the coverage of venues.

The absence of a “most liquid ETFs” ranking avoids presenting volume, AUM or a single snapshot of the spread as a universal verdict. Liquidity should be observed at the moment, and on the venue, where the order could be executed. The criterion is stated in the methodology.

A checklist before interpreting an order

• Right instrument: are the ISIN and the class the ones you want?
• Right listing: are the exchange, ticker and currency identified?
• Right time: is the data current, and is the venue in continuous trading or in auction?
• Observable underlying: are the relevant markets and hedging instruments open?
• Relative spread: what is the percentage distance between bid and ask?
• Size: how many units are available at the first level and at the next ones?
• Order type: which risk does it privilege — price, execution or time?
• Full cost: are commission, currency conversion, spread and possible impact consistent with the amount?

The checklist does not decide whether to buy. It only prevents confusing the ETF’s name with the concrete conditions of its trading.

Eight mistakes to avoid

• “Few trades means an illiquid ETF.” Past volume does not measure the full capacity of the primary market and the underlying.
• “High AUM means an always-tight spread.” Fund size and executable conditions are not the same quantity.
• “The best ask holds for any size.” It holds only for the size available at that level, as long as the proposal stays on the book.
• “The last price is what I will pay.” Execution meets the current bid and ask, not necessarily the last trade.
• “The same ISIN has the same liquidity everywhere.” Each listing has its own venue, currency, hours and book.
• “A limit order guarantees a better price.” It only guarantees not to cross the limit; it may not be executed.
• “When the underlying is closed, the price is false.” It is a market estimate with possibly higher uncertainty, not automatically an error.
• “The TER includes the cost of the order.” Spread, commissions, currency conversion and impact belong to another level.