Why Anyone Quotes At All
A quoted spread is a fixed cost plus two risks — the position the quoter did not want, and the customer who knows more than they do. The cost sets the floor; the two risks are why the spread moves, and almost every behaviour in the rest of this course is one of them being managed. On the example below, an unchanged company goes from a penny wide on 2,000 shares to four cents wide on 300 in a single afternoon, turning a $10 round trip into more than $40 before any news has arrived.
Prerequisites: Lesson 1 for inventory risk, and lesson 3 for adverse selection. Both have been named. This lesson turns them into something you can predict with.
Consider the job honestly. You must post a price at which you will buy and a price at which you will sell, and you must post both before knowing which one somebody is going to take. You do not get to choose your customers. Whoever arrives, arrives.
Part of what you charge for that is dull and roughly fixed: exchange fees, clearing, the technology, the people. It sets a floor under the spread and it does not change much from one hour to the next, so it explains why spreads exist but not why they move. The two things that make a spread move are risks, and they are worth understanding separately.
The first risk: you end up owning things
Somebody sells you 5,000 shares at your bid. You now own 5,000 shares of something you have no opinion about, bought purely because you said you would. Until you sell them, every tick against you is your loss, and the penny you were trying to earn is nothing against a move of thirty.
So you do the obvious thing: you make it more attractive for the next person to buy from you, and less attractive for them to sell to you. You quoted 50.02 bid and 50.03 ask; now you quote 50.01 and 50.02.
Look at what just happened on the screen. The price fell from 50.02 to 50.01. No news arrived. Nobody sold with conviction. One participant is managing a position they never wanted, and that alone moved the quote. A meaningful share of the small movement in any market is this, and it means nothing at all — a fact worth holding on to when you reach the lessons about reading intent from price.
The second risk: some of your customers know things
The second risk is worse than the first, because you cannot hedge it and you cannot see it coming.
Whoever sold you those 5,000 shares had a reason, and most people who take your quote are trading for reasons unrelated to what the thing is worth — a fund with money to put to work, a pension rebalancing, somebody paying a tax bill. Against them you earn the spread and nothing bad happens. But a minority arrive precisely because they have worked out that your price is wrong, and against those you lose every time, by construction. That is adverse selection, and it is the cost of standing still.
Your spread has to be wide enough that what you earn from the uninformed majority covers what you lose to the informed minority. That is the whole business, in one sentence, and it produces a set of predictions you can check yourself this week.
What the model predicts
If a spread is the price of those two risks, it must widen whenever either rises. Before a scheduled announcement, the chance that the next person to arrive knows something goes up, and spreads widen — often hours before, with no change in price at all. In the first minutes after the open, a night of information has not yet been absorbed, and spreads are at their widest of the day. In a thinly traded instrument, there is not enough uninformed flow to pay for the informed, so spreads are structurally wide and stay wide. When volume dries up, the same thing happens intraday.
Two of those are free to act on, which makes them the cheapest saving in this module — the scheduled release is the four-cents-on-300 case priced out below, and the hour is the same effect without an announcement to blame it on. The hour is one: the same instrument quotes far wider in a thin overnight session than in the middle of a liquid one, while the edge you are trading is usually identical either way — so the difference between the two is a cost you remove by choosing when to press the button rather than by trading better. The scheduled release is the other: the widest quote of the day is the one printed in the seconds around it, and a market order sent then pays a multiple of what the same order pays a few minutes later. Neither multiple is a figure to take from a lesson, because it belongs to your instrument and your venue — record your own at the hours you actually trade, the way problem 2 records it around an announcement. The direction, though, is not in question.
And the one that matters most. A market maker who has just been run over — who bought and bought as price fell straight through them — will quote wider, quote smaller, or stop quoting altogether. They are largely allowed to. Some venues designate market makers who carry a quoting obligation, but an obligation to show a price has never been an obligation to show a good one, and a quote widened to the edge of the rule is a withdrawal in all but name.
So liquidity is not a property of an instrument. It is a decision, made continuously, by people who can stop making it — and they stop under exactly the conditions that made you want to trade. The 7,300 shares you counted in lesson 2’s book were counted on a calm afternoon, and 900 of them stood at the quote. Neither figure is a promise about Thursday.
Reading a spread as a number
Take a stock that normally quotes a penny wide on 2,000 shares a side. Earnings are due after the close. By mid-afternoon it quotes four cents wide on 300 shares.
Nothing about the company changed today. What changed is that quoting has become dangerous: whoever arrives next is more likely than usual to know something, so the people quoting have raised their price for standing there and reduced how much of it they will do.
Price that as a trader. Your round trip was $0.01 a share; it is now $0.04. On 1,000 shares that is $10 becoming $40. Worse, only 300 shares are available at the quote, so a 1,000-share order walks levels as well — the cost is not four times, it is more.
The useful part is that it is visible before anything happens. A widening spread on unchanged price is telling you something specific: the people whose job is to be there have decided the next few hours are expensive. You do not need to know what they know to take the same view of your own costs.
What this does not settle
This is a model of one market maker with one position. Real firms quote thousands of instruments at once, hedge across them, and net exposures in ways that make any single quote a poor guide to what they hold. When you see a quote shade down, you are seeing the output of a system, not one trader’s feelings about your stock.
Nor does the model tell you when inventory is being worked off rather than information being acted on. Both look like price moving. Distinguishing them is what the whole of Module 4 exists for, and it cannot be done from the quote alone — which is precisely why the rest of this module goes looking for other evidence.
There is also no competition anywhere in this model. One quoter sets a price for a risk, and that is not how a quote is arrived at: several firms quote the same instrument and undercut one another, so the spread you see is the narrowest any of them will accept rather than the price this one wanted. The risk story explains the floor and the movement. It does not explain the level.
The freedom to stop quoting is also not uniform. Obligations differ by venue, by instrument and by the class of firm, and a designated market maker on one exchange operates under rules that do not apply to a proprietary firm quoting the same name elsewhere. “They can simply stop” is the right instinct and the wrong sentence.
And every prediction above carries a direction and no magnitude. That spreads widen before a scheduled release is checkable; that they widen by enough to change what you should do is not, until somebody supplies the number. Nobody can supply it for you, because it belongs to your instrument, your venue and your hour, which is what problem 2 is for and why this lesson declines to print a multiple it would only be inventing.
A spread is not a fee. It is a price somebody set for a risk they are taking, and it moves when the risk does.
Problems
- Shading. A market maker quotes 20.10 bid, 20.12 ask, 1,000 shares each side. They are hit for the full 1,000 on the bid twice in a minute. Describe what happens to their quote and say why, then state what a chart of that minute would show and what a reader of that chart would probably conclude.
- The prediction. Pick a liquid stock with a scheduled earnings date. Record the quoted spread and the size at the quote at the same time each day for the three days before it, and again on the morning after. Did the spread widen before the announcement rather than after? By how much?
- The uninformed majority. The model says the spread must cover losses to informed traders out of profits from uninformed ones. Using that, explain why an instrument traded almost exclusively by professionals can have a wider spread than a household name with the same daily volume.
Sources. Lawrence R. Glosten and Paul R. Milgrom (1985) for the adverse-selection half. Thomas Ho and Hans R. Stoll, “Optimal Dealer Pricing under Transactions and Return Uncertainty” (Journal of Financial Economics 9, 1981), for the inventory half and the shading described above. Larry Harris, Trading and Exchanges (Oxford, 2003), chapter 13, for both together in practitioner form.
That is the mechanism complete: a price, a book, a fill, a spread, and the person who sets it. The next lesson turns to what your screen does with all of it, and to the five numbers a candle keeps out of everything that happened in its interval. Which five they are is what decides the questions a chart can answer.
What a Fill Actually Is
Adverse selection from your side of the trade rather than the quoting side.
Read Lesson →The Spread Is the Price of Immediacy
What the spread charges for, and how to read it changing.
Read Lesson →The Order Book Is Theater
Why displayed size is an advertisement, and the tests that separate it from a fact.
Read Lesson →The Market Maker’s Business
The same two risks, run at scale by a firm rather than a person.
Read Lesson →Educational only. Trading involves substantial risk of loss. Not financial advice. Past performance does not guarantee future results.
💬 Discussion (0 comments)
Loading comments...
Ready to Trade with Signal Pilot?
Apply your trading education with professional indicators and real-time market analysis tools.
Back to Signal Pilot →