The Spread

A dealer names two prices, and the doubling is the whole of the matter. Ask him what a share is worth and he will not give you a number but a pair: a lower price, the bid, at which he will buy it from you, and a higher price, the ask, at which he will sell it to you. The gap between them is the spread. Why two prices and not one? Because the dealer is not, in truth, pricing the share at all. The share has, at any moment, a single best estimate of its worth — the midpoint between his two quotes, the price at which buyers and sellers stand in rough balance — and that midpoint is the market's verdict on the asset's value. What the dealer prices with his two quotes is something else: not the asset, but the service of trading it at once. The bid and the ask straddle the value like the two banks of a river, and the spread between them is the toll for crossing.

Who pays the toll, and who collects it, is settled by a single distinction — between those who must trade now and those who can wait. One who needs to sell immediately accepts the dealer's bid, taking the lower price for the sake of speed; one who needs to buy immediately pays the ask, the higher price, for the same reason. These are the demanders of immediacy, and they cross the spread. But a patient trader need not cross it. Instead of accepting the dealer's quote, he may post his own — an order to buy a little above the prevailing bid, or to sell a little below the ask — and wait for the world to come to him; and if it does, he has traded at a better price than the impatient, capturing rather than paying the spread. He is a supplier of immediacy, standing ready so that others need not wait. The spread, then, is not a tax the dealer alone imposes; it is a payment that flows from the impatient to the patient, and the dealer is merely the most practiced of the patient, the one whose whole trade is to stand ready.

Cross the spread in both directions and its meaning is laid bare. Buy a share at the ask and sell it back an instant later at the bid, with the value unchanged between, and you are poorer by the spread: you have paid, in that round trip, the full price of immediacy, twice crossing the river for no gain but the speed of it. This round-trip cost is the spread's truest measure, and it is universal, though it wears many disguises. On a heavily traded share it is a matter of pennies, the market so liquid that immediacy is nearly free; on a thinly traded one it widens to a painful fraction. And beyond the exchanges it is everywhere, though rarely posted: the used-car dealer who buys your car for less than he will sell it, the currency booth whose buy and sell rates differ, the house that fetches one price sold in a week and another sold in a year — each is a spread, the price of turning a thing into money now rather than later. Wherever there is a middleman, there is a gap between the two prices he names, and that gap is what this volume is about.

To see what the middleman is for, imagine a world without him. In such a world every buyer must find his own seller and every seller his own buyer — and not just any counterparty, but the right one, who wants exactly what is offered, in the quantity offered, at the moment and the place it is offered. A man with a surplus of grain and a want of cloth must find not merely someone with cloth, but someone with cloth who also wants grain: the double coincidence of wants, the ancient obstacle the sixteenth volume will show money was invented to overcome. Even where money has dissolved that particular knot, the difficulty of finding remains. To sell a house, a bond, a barrel of oil, or an hour of rare skill, one must search — for a willing party, at an agreeable price, without too much delay — and search is costly, in time, in effort, and in the gnawing uncertainty of whether a better counterparty waits just beyond the next inquiry. The middleman exists to spare the world this cost. He makes himself easy to find, stands always ready to trade, and so converts the hard problem of finding a particular counterparty into the easy one of finding him.

What the intermediary bridges, looked at closely, is a set of gaps, and he bridges them in four kinds. He bridges time: the dealer who buys your share today, before any final buyer has appeared, holds it until one does, spanning the interval between your wish to sell and another's wish to buy. He bridges space: the merchant who carries goods from where they are plentiful and cheap to where they are scarce and dear spans the distance between producer and consumer, the oldest form of the trade. He bridges information: the broker who knows who is buying and who selling, and matches the two, spans the gap of mutual ignorance that would otherwise keep them apart. And he bridges want and denomination: the bank that gathers a thousand small savings into one great loan, the wholesaler who buys in bulk and sells in parcels, reconciles the shape in which value is offered with the shape in which it is wanted. In every case the gain is the same in kind. Were a hundred buyers each to search among a hundred sellers, the matchings to be tried would number ten thousand; let all trade instead through a single hub that stands ready to both, and the hundred buyers and hundred sellers need strike but two hundred bargains, each with the middleman. The intermediary collapses a multitude of searches into a single point of contact, and that collapse is the source of his worth.

From this the spread takes its deepest justification. What the middleman charges is, at root, the price of the search he saves — and where he saves more than he charges, the trade is a true exchange, leaving both sides better off than they would have been alone, exactly as the insurer's loading left both better off in the volume just past. The dealer earns his spread honestly when standing ready, holding inventory, and matching the two sides costs him less than it would cost the parties to find one another; and competition among such dealers, as a later plate will show, presses the spread down toward that cost. This is the legitimate core of intermediation, and it is the standing answer to the charge that the middleman is a parasite: he produces nothing one can touch, true, but he produces something real all the same — the meeting itself, the trade that would not otherwise have happened, or would have happened only at greater cost. Yet the justification carries its own warning, which the volume's last plates will press. The spread is earned only to the extent of the search it saves; charged beyond that, it is no longer payment for a service but a toll on a passage the middleman has done nothing to ease — and the same position that lets him save the world its search may let him stand astride a gate he himself keeps shut.

The first plate called the spread the price of immediacy; this plate asks what immediacy is, and gives it its market name — liquidity. An asset is liquid when it can be turned into cash quickly, in quantity, and without moving its price; illiquid when it cannot. Cash itself is perfectly liquid; a traded share nearly so; a house or a private business barely at all. Liquidity is not a property of the asset alone but of the market around it — of how many stand ready to take the other side, how deep their willingness runs, how fast they replenish once a trade has drawn them down. And like immediacy, of which it is only the fuller name, liquidity is bought and sold. Those who demand it — who must trade now — pay the spread; those who supply it — who stand ready so that others need not wait — earn it. To understand the spread, then, is to understand the market in liquidity, and to see that market plainly one must look at the mechanism in which it lives: the order book.

Modern markets keep a book of orders, and the book makes the trade in liquidity visible. Into it flow two kinds of order. A limit order is an offer to trade at a stated price or better, posted and left to wait — an order to buy at no more than so much, or sell at no less; it adds itself to the book, joining the queue of the patient, and supplies liquidity to whoever later wants it. A market order is an instruction to trade at once, at whatever price the book offers — it does not wait but reaches into the book and takes; it demands liquidity, and pays the spread to get it. The book thus has two sides: a stack of bids below, the standing offers to buy, descending from the best bid; and a stack of asks above, the standing offers to sell, ascending from the best ask. Between the best bid and the best ask lies the spread, the no-man's-land where no trade rests. And the quantity standing at each price is the market's depth — how much can be traded before the price must move. The figure shows the book at a moment, with all of these upon it.

The picture lets us name liquidity's three dimensions, for it is not one thing but three, and a market may be rich in one and poor in another. The first is tightness — the narrowness of the spread itself, the cost of a small trade done at once — and it is the dimension the rest of this volume chiefly pursues. The second is depth — how large a trade the book can absorb before the price moves against it, measured by the quantity standing near the best quotes — so that a market may have a tight spread for small orders yet little depth for large ones, immediacy cheap for the minnow and dear for the whale. The third is resilience — how quickly the book refills after a large trade has swept through it, the patient returning to replace the liquidity consumed — for a market that empties and does not refill is liquid only until it is tested. Together these make immediacy what Harold Demsetz, who first thought to price it, treated as a predictable thing: the assurance that one may trade now, in size, at a price near the last. That assurance is supplied by the patient who post and wait, and consumed by the impatient who take; and because the patient bear risk and trouble to supply it, while all may consume it, liquidity has something of the character of a common resource — abundant when many tend it, and prone, as the thirteenth volume warned of commons, to vanish exactly when the crowd that relied on it most needs it to remain.

Of the three costs that set the spread, the first to examine is the one the volume just past prepared us for: the cost of bearing risk. A dealer who makes a market cannot wait, as the patient do, for a matching order to appear before he trades; that is the whole of his service — he trades now, taking the other side at once, and worries afterward about finding someone to take it off him. So when you sell to him, he buys, and is left holding the asset; when you buy from him, he sells, and is left short, owing an asset he must later acquire. Either way he carries a position he did not want for its own sake, only as the residue of having served you, and he carries it through an interval — seconds, hours, days — in which its price may move. That is inventory risk, and it is nothing other than the price of risk of the fourth volume, now seen from the side of the bearer rather than the shedder. The insured pays a premium to be rid of a risk; the impatient trader pays a spread to be rid of one too — the risk of holding, of waiting, of not yet having traded — and the dealer, like the insurer, is paid to take it on.

How the dealer manages this risk shapes his quotes, and the management is ingenious. He does not hold a fixed view of the asset's value and quote a fixed spread about it; rather, he shades his quotes according to the inventory he happens to carry, as Garman, and Ho and Stoll, taught us to see. Suppose he has bought heavily and runs long, holding more than he wishes: he is overexposed, anxious to sell and reluctant to buy more, and so he lowers both his bid and his ask together — the lower ask to tempt buyers to relieve him, the lower bid to deter further sellers — sliding his whole quote down until his inventory eases. Suppose instead he has sold heavily and runs short: he raises both quotes, the higher bid to attract sellers who will cover him, the higher ask to deter buyers. His quotes thus drift with his position, always leaning against it, always nudging his inventory back toward the level he wishes to hold. The figure traces this: how the dealer's bid and ask slide as his inventory moves from short to long, the spread between them riding up and down upon the true value like a boat upon a swell.

What the figure shows about a single dealer, the market enforces upon all: the spread must be wide enough to pay for the inventory risk borne, or no one will bear it. A dealer who quoted too tight a spread would be flooded with the trades others shunned, accumulate inventory he could not safely carry, and be ruined by the first adverse move; the spread he survives on is the one that compensates him, on average, for the price risk he shoulders between the trade that gives him a position and the trade that clears it. And here the kinship with the premium is exact. The insurer's loading, the volume past showed, was in part a charge for the capital he must hold against the risk he has assumed; the dealer's spread is the same charge in another costume — payment for standing between, for holding what waits, for converting the impatient trader's risk of not-yet-having-traded into his own risk of having-traded-too-soon. The wider the swings in the asset's price, the more this protection must cost, which is why spreads gape in volatile markets and on volatile assets: when the thing in inventory is dangerous to hold, immediacy in it is dear. The first reason for the spread, then, is simply that someone must carry the asset across the gap of time, and carrying is risk, and risk has a price — the price the fourth volume named.

The deepest reason for the spread is the subtlest, and to reach it we must set the first reason aside entirely. Grant the dealer a costless trading desk, so that order-processing is free; grant him the power to clear every position instantly, so that he bears no inventory risk at all. Even so — and this is the discovery that remade the study of markets — a spread will open, for a reason of pure information. The reason is the malady of the twelfth volume, adverse selection, here wearing the dress of the trading floor. Among those who come to the dealer to trade, some know more than he does about what the asset is truly worth — an insider, an analyst, a trader who has seen the news a moment sooner — and against these informed traders the dealer cannot win. When the informed buy from him, it is because they know the price is headed up, and he has sold too cheap; when they sell to him, it is because they know it is headed down, and he has bought too dear. He loses to them as surely as the insurer lost to the sick who bought eagerly, and for the same reason: the very eagerness to trade is a signal, and it is a signal of bad news for whoever takes the other side.

Glosten and Milgrom showed precisely how this lone fact builds a spread, and the construction is beautiful. The dealer cannot tell the informed from the uninformed — they arrive at his window alike — but he knows the direction of a trade is itself evidence. The mere fact that someone wishes to buy makes it a touch more likely the asset is worth more than he thought, for an informed trader would buy only on good news; the fact that someone wishes to sell makes it likelier the asset is worth less. A dealer who would not be systematically fleeced must therefore price not against the asset's unconditional value but against its value conditional on the direction of the trade he is about to make. He sets his ask — the price at which he sells — above the unconditional expected value, high enough to cover the chance that his buyer knows something good; and he sets his bid — the price at which he buys — below it, low enough to cover the chance that his seller knows something bad. The gap between these two conditional expectations is the spread, and it has appeared without a single penny of processing cost or inventory risk. The figure shows it widening as the share of informed traders grows: where none are informed, the conditional values collapse to the unconditional one and the spread closes; where many are, the dealer must protect himself fiercely, and the spread yawns.

This adverse-selection spread is the toll the uninformed pay for the presence of the informed among them — and they pay it without ever meeting an informed trader, simply by trading in the same market. The dealer, unable to tell who is who, must charge everyone the wider spread the danger of the few requires; so the patient saver rebalancing his pension, who knows nothing and threatens no one, pays the same protective margin as covers the insider, his immediacy made dearer by company he never sees. The connection to the twelfth volume is therefore exact: the spread is, in this part, the market's quarantine against asymmetric information, the price all pay because some know more. But the story has a second face, and it is more flattering. The informed trader, in picking off the dealer, also moves the price — buying when value is high drives the price up, selling when it is low drives it down — and so drags the market's price toward the truth he privately knows. This is price discovery, and Albert Kyle showed how to measure its cost: the depth of a market, how far an order moves the price, is set by exactly how much informed trading the dealer fears. The spread and the price impact are thus two faces of one thing — the friction by which private knowledge is forced, at a cost, into the public price. The uninformed who pay the adverse-selection spread are, in a sense, paying for the accuracy of the very prices they trade at.

We may now gather what the central plates have separated. The spread the dealer charges is not one thing but the sum of three, each a distinct reason that standing ready costs him something. There is the order-processing cost — the plain expense of the apparatus of trading, the clerks and screens and exchange fees, the cost of doing the work — which is real but small, and, being a cost of operation rather than of risk or information, roughly fixed per trade. There is the inventory cost, the price of the risk the dealer bears while he holds what he has bought and waits to sell it, the fourth volume's premium borne by the market-maker. And there is the adverse-selection cost, his expected loss to the better-informed, the twelfth volume's malady become a charge on every trade. The half-spread, the distance from the midpoint to either quote, is the sum of these three; the full spread is twice it. To take a spread apart is to find these three components stacked within it, and the figure stacks them.

The decomposition is not merely a conceit; it can be measured, and the measuring has occupied a generation of students of market microstructure. The three components leave different fingerprints. The order-processing cost is paid on every trade alike and does not depend on what follows. The inventory cost reveals itself in the way a dealer's quotes drift with his position and then revert. The adverse-selection cost shows itself in what happens to the price after a trade: if a purchase from the dealer is, on average, followed by a rise — if the buyer tended to know something — then part of the spread was protection against him, and that part is permanent, for it reflects a true revision of the asset's value, not a temporary swing that will reverse. By such signatures the parts can be told from one another, and what the data show is instructive. The order-processing cost is small and steady. The variation in the spread, across assets and across time, comes overwhelmingly from the other two — from how much risk the asset is to hold and how much informed trading swirls around it. The figure sets a liquid market beside an illiquid one to show this: the processing layer scarcely changes, while the inventory and adverse-selection layers swell, and it is their swelling that makes an illiquid market dear.

Over all three components plays the force the next plate will treat at length: competition. For the spread the dealer charges need not equal the cost he bears; the difference, where it exists, is his profit — a markup over cost that his market power, if he has any, allows him to keep. Competition among dealers presses this markup toward zero, driving the spread down until it just covers the three costs and no more. But mark precisely what competition can and cannot do. It can compete away the dealer's rent, the part of the spread that was pure profit; it cannot compete away the costs themselves. No rivalry, however fierce, can make it costless to bear inventory risk or to trade against the informed; those costs are features of the asset and its market, not of the dealer's greed, and they set a floor beneath the spread that competition approaches but cannot pierce. This is why the spread is so revealing a number. Stripped by competition to its costs, it becomes a barometer of those costs — a reading, taken from the price of immediacy, of how risky the asset is to hold and how thick with private information its market is. A wide competitive spread does not indict the dealer; it reports that the thing traded is dangerous to carry, or swarming with those who know more. The spread, honestly competed, tells the truth about the market beneath it.

The spread, the last plate showed, has a floor of costs that competition cannot pierce; this plate asks what holds it above that floor, and what presses it down. What holds it up is market power; what presses it down is rivalry, and the form the rivalry takes decides the spread as surely as the costs do. Begin with the simplest case. A single dealer, alone in a market — the old specialist, the sole broker in a country town — faces no rival to undercut him, and so may set his spread well above his costs, the surplus his monopoly profit, the toll of a gatekeeper who keeps the only gate. This is the market power of the tenth volume, exercised upon the price of immediacy. Now admit competitors. When many dealers quote, a buyer takes whichever ask is lowest and a seller whichever bid is highest, so each dealer who hopes to trade must better his rivals' quotes — and the logic, which the volume on power will name for Bertrand, drives the spread down toward cost, each dealer shaving his margin to win the order until little is left to shave. The number of dealers and the keenness of their rivalry thus set the spread between its floor of cost and its ceiling of monopoly; competition is the force that walks it down from the one toward the other.

But competition among dealers carries a peculiar hazard that keeps the spread from ever quite reaching its floor — the winner's curse, met before in the volume on the auction and the bid. Consider what it means to win a trade by quoting the best price. The informed trader, who knows where value lies, shops for the keenest quote as eagerly as anyone — more eagerly, for his edge is largest where the price is most favorable to him — so the dealer who posts the tightest quote is precisely the one most likely to find himself trading against someone who knows better. To win is to be selected against. A dealer who forgot this and competed his spread down to mere processing cost would discover that his winning trades were disproportionately the poisoned ones, and would bleed to death on the very orders he had fought to capture. Knowing it, dealers hold back, refusing to tighten past the point where the adverse selection on the marginal won trade eats the margin. The spread that competition produces is therefore not the naive floor of processing cost but a higher one, set where the keenest safe quote sits — competition pressing down, the winner's curse pressing back, the spread resting where the two balance.

The modern market complicates this contest with a practice that turns the adverse-selection logic into a business model: the sorting of order flow by how informed it is. Retail orders — a saver buying a few shares, a household rebalancing — are, on the whole, uninformed, and therefore safe to trade against; a dealer who could trade only with them would face almost no adverse selection and could quote a very tight spread. So such flow is valuable, and a market has grown up to capture it. Wholesale dealers pay retail brokers for the right to fill their customers' orders — payment for order flow — and internalize those orders, executing them in-house at a hair better than the public quote, creaming the harmless retail stream and leaving the dangerous, possibly-informed flow to the public exchanges. The retail trader gains a small price improvement and is content; but the public exchange, drained of the safe flow that once diluted its danger, faces a more adversely-selected remainder and must widen its quoted spread in turn. The spread the small investor sees grows tight even as the spread that sets the public price grows wide, and the market fragments into a sheltered channel for the uninformed and an exposed one for everyone else — a sorting that lowers the price of immediacy for some by raising it for others, and whose net worth to the whole is among the most disputed questions in the regulation of markets.

So far the spread has been a cost of trading — a toll paid at the moment of exchange. But a toll that must be paid every time one trades does not stay at the threshold; it seeps back into the value of the thing traded. An asset that is costly to sell is, for that reason alone, worth less than an otherwise identical asset that is cheap to sell, for whoever buys it knows he, too, will one day pay the spread to be rid of it. Illiquidity is a defect in an asset as surely as a flaw in its earnings, and the market marks the price down for it — an illiquidity discount, the haircut the title names. Seen from the other side, the same fact appears as a reward: because the illiquid asset is cheaper, it offers a higher return to whoever holds it, a liquidity premium paid to compensate for the spreads he will someday cross. The price of immediacy, paid at the moment of trade, thus has a mirror-twin paid across the whole life of the holding — and the two are the same money, the spread looked at as a cost when one trades and as a return when one waits.

Yakov Amihud and Haim Mendelson drew the relation between the two, and it is the subject of the figure. Range assets along an axis by their illiquidity — by the spread one must cross to trade them — and plot the return the market demands of each. The line rises: the more illiquid the asset, the higher the return it must offer to be held, the liquidity premium climbing with the spread. But the line rises at a decreasing rate, bending toward the flat, and the reason for the bend is the deepest point of the plate. The spread is paid only when one trades; an investor who holds for a long time pays it seldom, spreading its cost thin across many years, while one who trades often pays it again and again. So the patient demand only a small premium to hold an illiquid asset — they will rarely cross its wide spread — while the impatient demand a large one. The market sorts itself accordingly: illiquid assets gravitate into patient hands, who mind their illiquidity least, and the premium any asset must pay is set by the most patient hand willing to hold it, not the least. The curve bends because, as illiquidity rises, the clientele shifts toward longer horizons that amortize the spread, and the extra premium each further degree of illiquidity commands grows smaller.

This sorting by horizon is the quiet logic by which a great deal of the world's capital is arranged. The illiquid — the private business, the building, the venture stake, the loan that cannot be sold — flows to those who can hold it untroubled: the endowment with no call on its funds for a generation, the pension matching long liabilities, the patient family fortune. They are paid the liquidity premium for a patience that costs them little, and in being paid it they perform a service, lending their long horizon to assets that would otherwise go unfunded or underpriced. But there is a subtler thing the market prices than the mere level of an asset's illiquidity, and it returns us to the fourth volume. What investors fear most is not that an asset is illiquid on average but that it will turn illiquid exactly when they must sell — that its spread will gape in the crisis, when cash is dear and buyers have fled. This is liquidity risk, and like every risk that moves with the whole rather than washing out in the crowd, it commands a premium of its own: assets whose liquidity evaporates in bad times must pay more than assets whose liquidity holds, over and above any premium for their average illiquidity. The liquidity premium is thus kin to the term premium of the third volume and the risk premium of the fourth — another of the rewards the market pays for bearing what cannot be diversified or foreseen, and another component folded silently into every required return.

The intermediary of the central plates stood between a buyer and a seller of a single asset, bridging the gap of time that separated their wishes. The grandest intermediary of all bridges a deeper gap, and earns the most consequential spread in the economy for doing so. The bank stands between savers and borrowers, and the gap it spans is one of maturity. Savers want their money liquid — near at hand, withdrawable, lent only for a moment if at all; borrowers want money for a long time — to buy a house over thirty years, to build a factory that pays back over twenty. Left alone, the two could scarcely meet: no saver would tie up his money for thirty years, no borrower could promise to repay on a day's notice. The bank reconciles them. It borrows short, taking the savers' deposits, payable on demand; and it lends long, making the borrowers' mortgages and business loans; and it pockets the difference between the rate it earns on the long loans and the rate it pays on the short deposits — the net interest margin, which is the bank's spread, the price of its intermediation between the patient borrower and the impatient saver.

The bank's opportunity is the shape of the yield curve, which the third volume drew and the figure recalls: the rate on money lent rises, ordinarily, with the term of the loan, so that long money commands a higher rate than short. The bank lives in the gap between the two ends of that curve. It funds itself at the short end, where the rate is low — for it pays its depositors little, their money being liquid and safe — and it lends at the long end, where the rate is high, earning the term premium the third volume described. The net interest margin is the harvest of this spanning, the slope of the yield curve turned into the bank's income. But the margin is not a free lunch, for in earning it the bank performs a genuine and perilous transformation. It does not merely pass the savers' money through to the borrowers; it changes the money's character on the way, turning a thousand liquid, short, withdrawable deposits into a few illiquid, long, locked-up loans — maturity transformation, and with it liquidity transformation, the manufacture of liquidity itself. The bank tells every depositor he may have his money whenever he likes, and tells every borrower he need not repay for years, and both believe it, and ordinarily both are right.

The transformation is a true creation — the bank manufactures liquidity, giving the saver the liquidity he craves and the borrower the patience he needs, a service neither could have had alone, and one on which the whole modern economy is built. Douglas Diamond and Philip Dybvig showed precisely what the bank provides: insurance against the saver's own uncertain need for cash, a pool in which the many who will not need their money early quietly fund the few who will, exactly the risk-pooling of the fourth volume applied to liquidity rather than to loss. But the same construction that creates the liquidity contains the seed of its destruction, and the seed is the run. The bank's promise — that all may withdraw at once — is one it cannot actually keep, for its assets are locked in long loans that cannot be called in a day; the promise holds only so long as not everyone asks. Let the depositors come to fear that others will withdraw, and each has reason to rush for the door before the bank's cash runs out, and the rush itself empties it — a self-fulfilling panic, the catastrophe of correlation from the fourth volume in its most explosive form, in which the very liquidity the bank created vanishes precisely because everyone wants it at once. The net interest margin is the price the bank earns for spanning maturity; the run is the disaster it courts in earning it; and the sixteenth volume, on money, must take up how the state — through deposit insurance and the lender of last resort — props up a promise that is, by its nature, impossible to keep alone.

Bring the market-maker into the present, and he is no longer a man at all but a program, and the spread is set in millionths of a second by machines racing one another in the dark. The electronic market has, on its face, been a triumph for the price of immediacy: automated market-makers, competing fiercely and bearing tiny costs, have driven quoted spreads on liquid stocks to a thinness no human dealer could match, and the small investor today crosses a spread a fraction of what his grandfather paid. If the spread is the price of immediacy, immediacy has rarely been cheaper, and the machines may seem an unambiguous gift. But the same speed that narrows the spread has bred a contest of its own, and the contest raises, in a new and literal form, the question the fifth plate left open — who the informed really are, and whether the spread protects against the discovery of value or merely against being outrun.

The trouble is that a quote, once posted, is a sitting target until it is updated, and updating takes time — a few millionths of a second to see that the world has changed and pull the stale offer back. In that sliver the market-maker is exposed. When news arrives and the asset's true value jumps, every standing quote is suddenly mispriced — the maker's ask now too cheap, his bid now too dear — and a trader faster than the maker can buy the cheap ask or hit the dear bid before the maker reacts, taking a near-certain profit from a quote the maker would have withdrawn had he only been quick enough. This is sniping, and it is adverse selection of the fifth plate's kind with a crucial twist: the trader who picks off the stale quote need know nothing about the asset's value at all, only that the value has moved and that he is faster. Speed has become a form of information — information not about what a thing is worth but about the fact that its price is about to change — and against it the maker is as helpless as against any insider. The figure shows the moment: the value jumps, the maker's quote lags, and in the gap between them the swift collect what the slow have left exposed.

Two consequences follow, and they frame the modern anxiety about the speed of markets. The first is an arms race. Since being a little faster than one's rivals is worth a great deal — the difference between sniping and being sniped — market-makers and traders pour fortunes into speed: into faster cables, microwave towers strung in straight lines across continents, machines placed as near the exchange as money can buy them, all to shave microseconds from a reaction time. Eric Budish, Peter Cramton, and John Shim argued that this race is the product of a flaw in the market's very design — the running of trades continuously, instant by instant, which makes the tiniest speed advantage decisive — and that batching trades together, clearing them in periodic auctions a fraction of a second apart rather than one by one, would render raw speed nearly worthless and call the race off, narrowing the spread to its true informational cost. The second consequence is fragility. The liquidity the machines supply is liquidity only in fair weather: when prices move violently, the makers, fearing they are being sniped wholesale, withdraw their quotes all at once, and the liquidity that seemed so abundant evaporates in an instant — the flash crash, in which a market deep and tight one moment is empty and broken the next, the spread blowing out to nothing as every machine flees the same danger together. So the oldest question about the intermediary returns in silicon form. Is the high-speed trader a market-maker, the descendant of the dealer who narrows spreads and provides immediacy — the stabilizing speculator Milton Friedman defended, who buys cheap and sells dear and so smooths the price? Or is he a predator who fattens on the slow and flees when needed — the destabilizing speculator Keynes feared, who turns the market into a casino and the serious business of valuation into a guessing game about the guesses of others? The machines have made immediacy cheaper than ever in calm times; whether they have made it dearer, or merely more treacherous, in the times that matter most is the contest of the modern spread.

Every plate so far has priced the middleman's service; this one asks the question the pricing has circled from the start, and that the whole moral history of commerce has turned upon. Is the spread a fair price for a service rendered, or a toll extracted by one who has wedged himself into a passage he did nothing to open? The suspicion is ancient and nearly universal. The middleman makes nothing one can hold — he grows no grain, forges no tool, raises no wall — and to the moral eye of most times and places this has seemed damning: he buys cheap and sells dear, pocketing a difference he did not earn by any labor, living off the trades of others like a toll-keeper on a road built by other hands. The medieval law made crimes of his very acts, punishing the forestaller who bought goods before they reached the market and the regrater who bought to sell again at a higher price in the same place; the moralists condemned the usurer who sold the use of money and the engrosser who cornered a commodity; and the modern temper, which tolerates the merchant, still recoils at the scalper who buys tickets only to resell them, the speculator who profits from a famine's prices, the arbitrageur who seems to conjure money from nothing but being in two places at once. Beneath all these lies one intuition: that value is created by making and growing, and that the man who merely moves, holds, or times what others made is a parasite upon their industry.

The whole of this volume is the economist's answer to that intuition, and the answer is that the middleman does produce something real — not a good, but a service, and a valuable one. He produces the meeting that would not otherwise have happened: the liquidity that lets a man sell when he must, the immediacy that spares him the wait, the bearing of the risk he wished to shed, the search he need no longer undertake, the bridging of the gaps in time and space and information and maturity that the second and ninth plates set out. The arbitrageur who buys where a thing is cheap and sells where it is dear is not conjuring money from the air; he is carrying value from where it is abundant to where it is wanted, equalizing prices across the gap and bearing the risk of the passage, performing in space exactly the service the merchant has always performed. The speculator who buys in the glut and sells in the dearth, reviled in every age, is — when he guesses right — moving goods across time from the season of plenty to the season of want, smoothing the price and softening the famine, the stabilizing service Milton Friedman insisted the successful speculator must perform, since to profit he must buy when prices are low and sell when they are high, which is to say he must push them toward the middle. The spread is the honest price of these services, and to call it theft is to misunderstand what trade is and why the meeting of buyer and seller is itself a thing of value.

And yet the suspicion cannot be simply dismissed, for the same position that lets the middleman serve lets him extract, and the two can wear the same coat. The line the volume has sought runs here, between intermediation that creates value and intermediation that merely captures it, and the line is real though it is devilishly hard to draw. The middleman extracts rather than serves when he adds nothing and interposes everything — the rent-collector of the second volume in a new guise, levying a toll on a passage whose scarcity he guards rather than eases. He extracts when he manufactures the bottleneck he profits from: the dealer who corners the market, the platform that locks in its users and then raises its take, the broker who keeps buyer and seller apart precisely so they will still need him. He extracts when his service is pure transfer and no creation — the latency arbitrageur of the last plate, whose microsecond edge produces no grain of price discovery and merely shifts the spread from slower hands to faster, a private gain matched by an equal social loss. He extracts, the scalper's critics say, when he intercepts a surplus the maker meant for someone else, buying the underpriced ticket the artist priced low for his true audience and reselling it to the highest bidder, capturing a value he created none of. In each case there is a spread, and in each case the spread is a toll and not a wage — payment not for a service rendered but for a position held.

The contest, then, resolves into neither slogan. It is false that the middleman is a parasite, for most intermediation is the productive bridging of real gaps, and a world without middlemen would be a world of crippling search, frozen illiquidity, and unborne risk. It is equally false that every spread is an earned wage, for some intermediation creates nothing and merely levies, and the history of commerce is littered with toll-keepers who served only themselves. The truth is that the spread is sometimes the one and sometimes the other, and that telling which requires asking, of each intermediary in turn, the single question this volume has pressed from its first page: what would have happened without him? Where his absence would have meant a trade forgone, a risk unborne, a price less true, a search longer — there his spread is the price of a service, and is earned. Where his absence would have changed nothing but the destination of the surplus — there his spread is a toll, and is extracted. The moral suspicion of the middleman, so easily mocked by the economist, is at bottom a rough and serviceable intuition that the second kind exists; the economist's defense, so easily caricatured by the moralist, is the harder-won recognition that the first kind predominates. Between them lies not a verdict but a method — to judge each spread by the service beneath it, and to keep, by competition and by law, the middleman honest enough that his toll stays a wage.

The volume set out to price the most overlooked and most resented of the foundational prices — the middleman's cut, the gap between the two prices a dealer names — and eleven plates have traced it from the trading floor to the moral imagination. The account may now be gathered. The spread is the price of intermediation: the toll the middleman takes for standing between those who would buy and those who would sell, and bridging the gaps — of time, of space, of information, of maturity — that would otherwise keep them apart. Most fundamentally it is the price of immediacy, the cost of trading now rather than waiting for a willing other, paid by the impatient who cross it and earned by the patient who stand ready. Its width is set by three costs that standing ready imposes — the plain cost of processing the trade, the cost of bearing the risk of inventory held while a buyer is found, and the cost of trading against those who know more — and pressed by competition toward those costs, so that the spread, honestly competed, is a true reading of the risk and the information in the market beneath it.

The deepest discovery of the volume is that the spread is not a new and separate price but the meeting-place of the prices that came before, which is why it is the right note on which to close their group. Its inventory component is the price of risk of the fourth volume, the insurer's burden borne now by the market-maker, who is paid, exactly as the insurer is paid, to carry what others wish to shed. Its adverse-selection component is the malady of the twelfth volume, the asymmetry between the informed and the ignorant, become a toll levied on all because some know more. Its mirror in asset prices, the liquidity premium, is kin to the term premium of the third volume, another reward the market pays for bearing what cannot be foreseen. And the force that disciplines it is the competition whose absence the tenth volume will call power. The spread thus gathers into a single number the risk, the information, the time, and the rivalry that the other prices treat severally — and to understand it is to see the prices of this group not as four or five separate things but as facets of one subject, the cost a market charges for its most basic services.

The achievement this records deserves to be named plainly, for the middleman's cut, despised for most of history as the parasite's skim, has been made by the study of markets into a precise and illuminating object of science. To have seen, with Demsetz, that immediacy is a service with a price; to have shown, with Glosten and Milgrom, that a spread arises from information alone, and with Kyle, how private knowledge moves price at a measurable rate; to have priced illiquidity, with Amihud and Mendelson, as a discount on value and a premium on return — these turned a moral grievance into a working theory of how prices are made, trade by trade, in the only place prices are ever actually made, which is the meeting of a buyer and a seller across someone's book. And the thing the theory describes is itself among the great achievements of economic organization. Intermediation is the circulatory system of the market economy: the liquidity it manufactures is what lets capital flow to its uses, savers lend to borrowers, risk pass to those who will bear it, and goods reach the hands that want them most. A world that could not trade now, that had to search anew for every counterparty and hold every risk it could not place, would be a poorer and more frightened world; the middleman, and the spread that pays him, are much of what spares us it.

Yet the volume has been, as much as a vindication, a catalogue of where intermediation strains and turns against those it serves. The liquidity the market provides is a fair-weather friend, abundant in calm and vanishing in the storm, so that the spread is narrowest when immediacy can best be done without and gapes precisely in the crisis when it is most needed — liquidity risk, the engine of every panic. The maturity transformation by which the bank earns its margin is the same mismatch that breeds the run, a fragility finance has learned only to move and not to cure. The speed of the electronic market, which narrows the spread in tranquil times, funds an arms race that may be pure waste and a sniping that preys on the slow, and supplies a liquidity that evaporates in the flash crash. And over all hangs the volume's central and unresolved question — whether a given spread is the wage of a service or the toll of a position — which economics can frame with care but cannot finally answer, because the service is a counterfactual no one can observe and the fairness of the middleman's share is a moral question no model contains. The spread is an honest price and a powerful instrument; it is also, in its every component, a site of contest that the pricing illuminates without ending.

With this volume the group of the prices is complete, and the foundation of the table is laid. Five prices have been drawn, each a verdict on a different hidden contest. The wage is the price of labor, and the contest beneath it is over what an hour of a life is worth. The rent is the price of land and of all that is scarce, and the contest is over who shall keep the surplus that scarcity throws off. The rate is the price of time, and the contest is between the present and the future, the patient and the impatient. The premium is the price of risk, and the contest is between the one who sheds a danger and the one who bears it. And the spread is the price of intermediation and of immediacy, and the contest is between the value the middleman creates and the toll he may extract, between the patient and the impatient, the informed and the uninformed, who meet across his book. Labor, land, time, risk, and the meeting itself — these are the elementary services a market charges for, and The Prices has now named the charge for each.

To complete the prices is to complete the foundation, for everything the rest of the table will treat presupposes them. The quantities a market produces, the failures it suffers, the institutions that govern it, the aggregates it sums to, and the behaviors that move it — all are built upon prices, and could not be reasoned about until the prices themselves had been denaturalized. That has been the work of this group: to take the five numbers a market quotes for its most basic services and show that not one of them is the simple fact it appears to be — that each is the visible settlement of an invisible struggle, a verdict rendered on a contest the price itself conceals. The wage that looks like the value of work, the rent that looks like the cost of land, the rate that looks like the price of a loan, the premium that looks like the cost of cover, and the spread that looks like the middleman's skim are each, on inspection, a truce in a hidden war — and to read a price as a verdict rather than a fact is the habit of mind this group has tried to teach, and that the groups to come will need.

The debts of this volume run outward across the table. It hands on, most immediately, to The Margin, the sixth volume and the opening of The Quantities, where the difference between what a thing costs to provide and what it fetches — the markup the spread has been a special case of — becomes the subject in its own right. It hands on to The Money, the sixteenth, where the bank's manufacture of liquidity, the run that threatens it, and the double coincidence of wants the second plate invoked are taken up as the nature of money itself. It hands on to The Asymmetry, the twelfth, for the informed trader whose presence widens every spread; to The Rate, the third, with which its liquidity premium composes the return on every asset; to The Power, the tenth, for the market power that lets a spread swell into a rent; and to The Commons, the thirteenth, for the liquidity that is tended by the many and consumed by all and apt to vanish when most relied upon. And it stands beside The Premium, the volume just past and its nearest sibling, with which it shares the deepest of its themes — that to stand between, to hold, to bear what waits, is a service, and that the price of risk and the price of immediacy are two namings of the cost of carrying the burden of an uncertain future on another's behalf.

The spread, in the end, is the price of the crossing — of the meeting the middleman makes possible between a buyer and a seller who would not, or could not, have met alone. It is the price of being able to trade at all: to turn a possession into money, or money into a possession, now, at a price one can know, without searching the world for a willing other or bearing alone the risk of the wait. That is the quiet service on which every market rests, and the spread is its honest price — save where it is not a price but a toll, levied by one who has wedged himself into a passage he keeps narrow on purpose. Which it is, in any given case, the volume has insisted cannot be settled by a slogan but only by looking beneath the spread to the service, or the absence of service, that it pays for. The prices are complete; the contest each conceals goes on. The argument remains open.