The Bid–Ask Spread: Why It Exists

Last lesson the spread was just “the gap in the middle of the book.” Now we interrogate it: why is there a gap at all, who keeps it open, and what determines whether it’s a single cent or a brutal dollar? The spread is the most-watched number in microstructure because it’s the price of immediacy — the toll you pay to trade now instead of waiting — and, viewed from the other side, the market maker’s paycheck. Decode it and you can read a market’s health off one number.

The spread as the price of immediacy

Analogy. A 24-hour convenience store charges more than the supermarket across town. You’re not being robbed — you’re paying for immediacy and availability: the store stocks shelves, pays staff, and bears the risk of unsold inventory so that you can buy milk at 3 a.m. without waiting. A market maker is that convenience store for shares. The spread is the markup for standing ready to trade with you instantly, any time, on either side.

Definition. The bid–ask spread is best ask minus best bid. Crossing it — buying at the ask, selling at the bid — is the implicit cost of demanding immediate execution. Whoever quotes both sides (a market maker, or any resting limit order) earns that spread as compensation; whoever crosses it (a taker) pays it. The spread is therefore a transfer from the impatient to the patient, and its width is set by how costly and risky it is to be the patient one.

Worked example — the round trip, again, in cold numbers

A stock quotes bid $49.95 / ask $50.05 — a $0.10 spread. You buy 1,000 shares and instantly sell them, with zero price movement:

You lost $100 (1,000 × $0.10) while the “price” did nothing — it went to whoever quoted both sides. Halve the spread to $0.05 and the same round trip costs $50; widen it to $0.60 and it costs $600. The spread is a real, recurring tax on trading, and it’s the same arithmetic whether you’re a retail tapper or a pension fund.

Quoted, effective, and realized spread

Practitioners don’t trust the quoted spread alone, because you don’t always pay the full posted gap — and the maker doesn’t always keep it. Three flavours matter.

Definition.

• Quoted spread = best ask − best bid. The headline number on the screen.
• Effective spread = 2 × |trade price − mid price at the time of the trade|. What you actually paid relative to the midpoint. It can be smaller than the quoted spread if your order got price improvement (filled inside the quotes), or larger for a big order that walked the book.
• Realized spread = 2 × |trade price − mid price a short while later| (e.g. 5 minutes after). What the maker actually kept after the price moved. If the price drifted against the maker (because the trade was informed), the realized spread is smaller than the effective spread — the difference leaked away to adverse selection.

Analogy. Quoted spread is the sticker price on the menu. Effective spread is what you actually paid after a coupon (price improvement) or an upsell (walking the book). Realized spread is what the restaurant kept after the cost of the meal — and sometimes the customer who ordered knew the kitchen was about to raise prices, leaving the restaurant worse off than the bill suggested.

Worked example — effective vs realized

Mid price is $100.00; quoted spread is $0.10 (bid $99.95 / ask $100.05). You buy at $100.04 (a hair of price improvement inside the ask).

• Effective spread = 2 × |$100.04 − $100.00| = $0.08 — less than the $0.10 quoted, because you filled inside the quote.
• Five minutes later the mid has risen to $100.03 (your buy was mildly informed; the price drifted your way).
• Realized spread (from the maker’s view) = 2 × |$100.04 − $100.03| = $0.02 — the maker kept only 2 cents of the 8 it appeared to earn. The other 6 cents bled away because the price moved against the maker after the fill.

That leakage — the gap between what the maker appears to earn (effective) and what it keeps (realized) — is the fingerprint of adverse selection, and it’s the deepest reason spreads exist. Hold that thought; we have a whole lesson on it.

Why the spread exists: three risks in one number

Here’s the heart of the lesson. A market maker doesn’t pick the spread out of greed; it sets the spread to break even on average against three distinct costs. Decompose the spread and you find:

1. Order-processing cost — the roughly fixed per-trade cost of running the operation: exchange and clearing fees, technology, back-office. It barely moves with market conditions.
2. Inventory-holding cost (risk) — the maker ends each trade holding a position (long if it just bought, short if it just sold). Until it can offload that position, the price can move against it. This cost grows with volatility and with how long the maker must hold.
3. Adverse-selection cost — the risk that the counterparty knew something. If informed traders are picking off the maker’s quotes, every fill is slightly more likely to be on the wrong side of an imminent move. This cost grows fastest when informed trading is heavy, and it’s why spreads explode around news and earnings.

The interactive below stacks these three into a single quoted spread (in cents per share) and lets you flip the market between calm, normal, and volatile. Watch the bar grow — and watch which slices do the growing.

The lesson the chart hammers home: when markets get scary, the spread widens almost entirely because the inventory and adverse-selection slices balloon. Order processing — the “fee” part — barely budges. So a widening spread isn’t the maker getting greedy; it’s the maker pricing in more risk. The spread is risk compensation wearing the costume of a fee.

The decomposition, as arithmetic

Suppose a maker estimates, per share: order-processing 1¢, inventory risk 2¢, adverse selection 3¢. The half-spread it needs is 1 + 2 + 3 = 6¢, so it quotes 6¢ below mid on the bid and 6¢ above on the ask — a 12¢ quoted spread. If volatility doubles the inventory slice to 4¢ and informed trading pushes adverse selection to 7¢, the half-spread becomes 1 + 4 + 7 = 12¢ and the quoted spread doubles to 24¢ — even though the fixed processing cost never moved. That’s exactly the calm-to-volatile jump in the chart.

What makes spreads wide or tight

Reading the decomposition backwards gives you a checklist of what drives a spread:

Misconception: “spreads are wide because exchanges or makers are gouging.” Competition between makers relentlessly compresses spreads toward the sum of real costs and risks. Where you see a persistently wide spread, the usual culprit is genuine risk — thin liquidity, high volatility, or a real danger of informed counterparties — not a cartel. The convenience store charges more because 3 a.m. milk is genuinely costlier to provide.

Putting it together

The bid–ask spread is the price of immediacy: a transfer from impatient takers to patient makers, and the maker’s compensation for three real risks — fixed order-processing cost, inventory risk that grows with volatility, and adverse-selection cost that grows with informed trading. Quoted is the sticker, effective is what you actually paid versus the midpoint, and realized is what the maker kept after the price moved — and the gap between effective and realized is adverse selection leaking away. Competition compresses spreads toward the sum of real costs, so a wide spread is a risk signal, not a swindle.

Key Takeaways