Bid-Ask Spread and Market Depth: A Practical Guide to Execution Costs

Quoted spreads, displayed depth and order size decide what an execution really costs — a practical look at slippage, timing risk and adverse selection in liquid and thin markets.

What the quoted spread actually tells you

The spread is the gap between the best bid and the best ask: the highest price someone will pay and the lowest price someone will accept, right now. It is the first, most visible component of execution cost, and the one you can measure before you trade.

A market order pays it immediately. Buy at the ask, sell at the bid, and you have crossed the full spread on one side. A round trip — in and out at market prices with no price change in between — costs you roughly the full spread, before any other cost appears.

Spread in basis points, not cents

Cents are misleading because they say nothing about the instrument's price. Convert to basis points (bps): spread ÷ midpoint × 10,000.

A $0.01 spread on a $200 share is 0.5 bps.
A $0.02 spread on a $20 ETF is 10 bps.
A $0.05 spread on a $5 stock is 100 bps — about 1% paid on entry and another 1% on exit.

Only the third example is a real drag on returns, and on a quote screen it looks like loose change. This is why ETF spread comparisons belong in basis points rather than dollars — the same reason fund cost comparisons look trivial in cents but not in percentage terms.

Who sets the spread

Two forces dominate. Inventory risk — a market maker who buys now owns price risk until they sell, so they charge for it. And adverse selection: the possibility that the person on the other side knows something the maker does not. Wider spreads appear where both are larger: volatile names, thin names, and moments when information arrives, such as earnings releases or the first and last minutes of the session.

Where competition between venues and market makers is intense, spreads compress toward the minimum price increment. That is why the tick size regime matters so much — more below.

Market depth: the part most quotes hide

Depth is the size available at each price level. Level 1 tells you the best bid and offer. Level 2 shows stacked price levels and sizes. The practical question is not "what is the spread" but "what is the spread for my order size".

An order smaller than the size shown at the best quote gets the quoted price. An order larger than that walks up (or down) the book, filling at progressively worse prices. The quoted spread then describes only the first slice of your execution.

Displayed size is a lower bound, not the truth

Three reasons displayed depth understates what is actually there:

Hidden and reserve orders. Iceberg-type orders display a small portion and refresh as they fill. The visible book can be a fraction of resting interest.
Odd lots. Quote and trade reporting has historically been built around round lots, so smaller orders did not appear in the consolidated top-of-book quote. A quote showing one round lot can sit in front of a much deeper pool of odd-lot interest.
Off-book and auction liquidity. Dark venues, internalisers and auction mechanisms hold interest that never appears on the public book.

Regulation is moving, slowly, to close part of that gap. Under amendments adopted in September 2024, US round lots are being redefined by price band — tiered to 100, 40, 10 or 1 share depending on the share price — replacing a 100-share convention the SEC describes as more than 120 years old, with the new round-lot definition taking effect in November 2025 and a separate step to disseminate odd-lot quote information scheduled to follow (SEC Chair statement, SEC fact sheet).

The takeaway is practical: a quote showing 200 shares does not mean 200 shares is the market. It means 200 shares is what someone chose to show.

Depth as resilience

Depth also measures how much the price moves when size arrives. A deep book absorbs a 5,000-share order with a few cents of movement; a thin book gaps several percent. Two books can display identical spreads and behave completely differently under the same order — which is why spread alone never describes the cost of size.

Order size and slippage: the square-root law

Execution cost has two parts that behave differently:

Spread cost, proportional to your order regardless of size.
Market impact, the price concession needed to source liquidity, which rises with size.

Market impact is not linear. Across asset classes, venues and time periods, the empirical relationship is approximately:

impact ≈ σ × √(Q / V)

where σ is daily volatility, Q is the order size and V is daily volume. This "square-root law" is described in the market microstructure literature, and a study of roughly 8 million institutional trades in US equities found a clear crossover: impact is close to linear for very small orders and follows square-root scaling as size grows (arXiv:1811.05230). It also underpins commercial trading-cost models; research using over $1.7 trillion of live institutional trades finds the fraction of daily volume traded — and the square root of that fraction — to be the variables with the largest effect on realised costs (Frazzini, Israel and Moskowitz).

Why the square root matters

Square-root scaling means impact grows faster than you would like but slower than linear. Going from 1% to 4% of daily volume multiplies impact by roughly two, not four. Going from 1% to 25% multiplies it by five.

A hypothetical illustration using the formula above, with daily volatility of 2% in both cases:

Order = 5% of average daily volume (ADV). √0.05 ≈ 0.224, so expected impact ≈ 0.45%.
Order = 0.5% of ADV. √0.005 ≈ 0.071, so impact ≈ 0.14%.

In both cases the quoted spread might be 2–3 bps. The impact estimate is many times larger. For anyone trading size relative to a stock's volume, impact dominates the spread — and for the smallest orders the reverse is true.

What to compare your order against

Average daily volume (ADV) is the reference. Rules of thumb used by execution desks:

Under 1% of ADV: spread is usually the main cost.
1–10% of ADV: impact becomes visible; splitting the order across the day helps.
Above 10% of ADV: expect to be the market in that name; realistic expectations and multi-day horizons are needed.

These thresholds are conventions, not regulation. The underlying point is that cost scales with your order relative to what the instrument normally trades.

Timing risk, adverse selection and the cost of waiting

The trade-off when you switch from market orders to limit orders is not "free". You hand over a different risk.

Market orders pay spread and impact, guarantee a fill, and leave you exposed to price movement between decision and execution — timing risk.

Limit orders avoid paying the spread, but:

You may not get filled, and the price may run away from you (opportunity cost).
When you are filled, it is disproportionately when the market is about to move against you. That is adverse selection — resting bids get hit right before prices fall.

This asymmetry is why limit orders look cheaper in average-cost statistics than they feel in practice. The orders that fill and the orders that get cancelled are not a random sample.

Timing risk also scales with volatility and with how long your order works. An order worked over a full day carries more price risk than one executed in the closing auction. For portfolios moved in size, auctions and closing price mechanisms frequently concentrate liquidity and reduce impact — at the cost of giving up control over the exact print price.

How costs differ across markets

Minimum price increments — "ticks" — set a floor under spreads, and the rules differ by jurisdiction.

In the US, Rule 612 of Regulation NMS has long imposed a $0.01 minimum pricing increment for quotes and orders in NMS stocks priced at or above $1.00, which prevents the spread for those stocks from being narrower than one cent (SEC fact sheet). In 2024 the SEC adopted a second, smaller increment for the most tightly quoted stocks, alongside a cut in the maximum access fee that exchanges may charge to reach a protected quotation, to $0.001 per share (SEC press release 2024-137).

Time-weighted average quoted spread	Minimum pricing increment
$0.015 or less	$0.005
Greater than $0.015	$0.01

Both changes were originally due in November 2025, but the SEC has since granted temporary exemptive relief twice, and in June 2026 pushed compliance for the new tick sizes and access-fee caps to November 2027 (SEC exemptive order, November 2025, SEC Chairman statement, June 2026). Until then the one-cent tick still applies. Once in force, the half-cent tick will narrow the achievable spread for the most liquid names, where the penny tick rather than competition is the binding constraint.

In the EU, the tick size regime in Commission Delegated Regulation (EU) 2017/588 calibrates minimum ticks to both the price of the order and the instrument's liquidity, measured by average daily number of transactions (ADNT) on the most relevant market (EUR-Lex). A worked example published by Nasdaq: a share with an ADNT of 500 falls into the 80–600 band, giving a minimum tick of 0.02 at a price of €19.90 and 0.05 at €20.00 (Nasdaq Nordic notice) — a tick-driven cost jump of more than 100% at a price threshold, with nothing changing in the underlying business.

That threshold effect is the point. Tick regimes can make a 10 bp spread a legal impossibility in one instrument and the floor in another.

Across segments, spread levels differ enormously. The most heavily traded large caps quote in fractions of a basis point, while small-cap and thinly traded names commonly quote spreads many multiples wider, with quote sizes of a few hundred shares. In such instruments a modest order can be several percent of daily volume — which is where the square-root math above stops being academic.

A pre-trade checklist

Convert everything to basis points. Spread, impact estimate and fees on one scale.
Look past level 1. Check total depth within 10–20 bps of the midpoint, not just the best quote.
Compare to ADV. If the order exceeds 1% of daily volume, plan the execution.
Check the clock. Avoid the first and last minutes unless you want that volatility; midday books are usually deepest relative to flow.
Match the order type to the cost. Market, limit and stop orders carry different cost profiles; a wide spread plus small size usually favours a resting limit order.
Account for fees separately. Commissions, exchange access fees, FX conversion and settlement charges sit outside the spread but inside your total cost of ownership.
For funds, compare to NAV. A quoted ETF spread is only part of the story; creation and redemption mechanics, premiums and discounts, and the liquidity of the underlying basket all matter.

Risks and costs

Spread cost is paid on entry and exit; it is close to a certainty, not a risk, and it widens when volatility rises, around the open and close, and on days when correlated markets are closed.
Market impact is the largest cost for size, and it is not observable in advance — only estimated. Estimates can be badly wrong in stressed markets.
Timing risk: the price can move before or during execution, in either direction. Working an order longer reduces impact but increases exposure.
Adverse selection: resting limit orders are filled preferentially when the market is about to move against the resting side. A limit order can also fail to fill entirely, leaving the original intention unexecuted.
Fees: exchange access fees, commissions, currency conversion and custody charges are separate from the spread, and are easy to miss when comparing products. See the full cost breakdown for ETFs.
Liquidity illusion: displayed depth can evaporate. Quote sizes are not contractual promises, and depth that exists in calm conditions may be withdrawn precisely when you need to trade.
Thin markets: in small caps and illiquid instruments, spread and impact can combine to make a round trip cost several percent, which must be earned back before any gain.
Measurement error: published effective-spread figures are usually computed against the quote midpoint, which does not capture every execution and can overstate or understate cost depending on how your fill compares with the midpoint. Read execution-quality data with care.
Jurisdiction differences: tick rules, fee caps and disclosure requirements are set nationally (in the US, in the EU) and do not transfer between markets.

Bottom line

The quoted spread is the visible, measurable part of execution cost, and for very liquid instruments it is close to irrelevant in basis points. The expensive components are depth, size and time. Convert spreads to basis points, compare your order to average daily volume, treat displayed size as a floor rather than a promise, and remember that avoiding the spread with a limit order means accepting fill risk and adverse selection instead. Execution cost is a design decision made before you click send, not a detail settled afterwards.