The last two plates turned the kinder faces of the dealer's gap: the immediacy he sells, which is a service, and the inventory he holds, which is a risk he is fairly paid to bear. This plate turns the third and darkest, which is neither a service he renders nor a risk he chooses but a wound he cannot stanch — the loss he takes, again and again, to the people who trade with him precisely because they know something he does not. For the dealer stands ready to trade with all comers, and he cannot see into them. Most who come to him trade for reasons that have nothing to do with any private knowledge: they need cash, they are rebalancing a portfolio, they are following a rule, they are, in the market's unlovely word, noise. But some who come to him are informed — they have done the research, or caught the rumour, or seen the mispricing, and they know which way the price is about to move. And the dealer cannot tell the one kind from the other, for an order to buy looks exactly the same whether it comes from a man who needs cash or a man who knows the price is about to rise. This blindness is the source of the spread's third component, the one Glosten and Milgrom taught the discipline to see, and it behaves unlike either of the others — for it is not a charge the dealer keeps, but a tax he is forced to levy on the innocent to cover what the knowing take from him.
The trouble begins with a simple asymmetry in when the two kinds of trader choose to act. The uninformed trade at all times and in all directions, their buying and selling scattered without regard to where the price is headed, for they trade on needs and whims unrelated to the thing's future value. The informed do not trade so evenly. They trade only when they have an edge, and always in the direction the edge points: they buy from the dealer when, and only when, they know the price is about to rise, and they sell to him when, and only when, they know it is about to fall. This means that every trade the dealer does against an informed counterparty is a trade he loses. He sells at his ask to a buyer who knows the thing is worth more than the ask, and watches it rise; he buys at his bid from a seller who knows it is worth less than the bid, and watches it fall. He is, against the informed, always and necessarily on the wrong side, for the informed have chosen their side precisely because it is the winning one. And he cannot decline these trades, because he cannot identify them: the informed buyer's order is indistinguishable, at the moment it arrives, from the uninformed buyer's. The dealer is thus a creature condemned to be picked off, knowing he will be picked off, and unable to see the hand that does it until the loss is already taken.
What can such a dealer do? He cannot refuse the dangerous trades, for he cannot spot them; he can only price for them. This was the insight of Glosten and Milgrom, who showed that a dealer facing a crowd of which some known fraction is informed will set a spread sufficient to recover, from his gains on the uninformed, exactly what he loses to the informed — and that this spread will exist even if his inventory is never at risk and his costs of doing business are nil. It is a spread made of pure information, or rather of the lack of it: the dealer widens his quote not to cover any cost or risk of his own but solely to offset his certain losses to those who know more than he does, charging every trader a little extra because some traders are dangerous and none can be told from the rest in advance. The wider the danger — the larger the share of informed traders, or the greater their edge — the wider the spread must grow. And the same logic teaches the dealer to read the order flow itself as a signal: because a buy order might come from someone who knows the price will rise, each buy nudges his estimate of the thing's value upward and each sale nudges it down, so that the trades of the informed, even unidentified, drag the price toward the value they alone can see. The spread is the dealer's shield against the informed; the drift of his quotes is the informed quietly winning anyway.
The informed trade only when they are right, and the dealer cannot tell them from the rest; so he loses to them without fail, and recovers only by charging the innocent for the company of the dangerous.
Now mark how strange this third component is, and how unlike the other two. The immediacy the dealer sells is his to keep: it is the wage of a service, and the dealer is richer for it. The inventory risk he bears is his to keep too: it is the premium for a hazard, and the dealer earns it. But the adverse-selection spread is not the dealer's to keep at all. He collects it from the uninformed only to pay it straight out again in his losses to the informed; on the whole of it he merely breaks even, the gains from the innocent exactly cancelling the losses to the knowing. The spread that adverse selection forces him to set is therefore not a profit he extracts but a transfer he routes — money taken from the pockets of the uninformed and delivered, by way of his unavoidable losses, into the pockets of the informed, with the dealer functioning as nothing more than the conduit through which the transfer passes. He is not enriched by it; he is merely the tollgate at which the tax is collected. And the tax is real: every uninformed trader who crosses the dealer's spread pays a little more, or receives a little less, than the thing is worth, and the margin of his loss is precisely the margin by which the informed, somewhere in the same crowd, are being paid for knowing what he does not.
This recasts the volume's whole question, for here the extraction is real but the middleman is innocent of it. Through every prior plate, the question has been whether the dealer earns his cut or merely levies it — whether the spread is his honest wage or his unearned toll. The adverse-selection component answers: neither, for he keeps none of it. The extraction is genuine, but its agent is not the man in the middle; it is the informed trader, reaching through the dealer's spread into the uninformed trader's pocket. The dealer is not the predator here but the predator's unwilling instrument, the channel cut by the informed through which they drain the uninformed, and the spread he sets is not the wall he builds to rob the customer but the wall he builds to survive being robbed himself. The volume's seam — between the earned and the extracted — runs, in this plate, not through the middleman at all but through the crowd of traders he serves, dividing the informed who gain from the uninformed who lose. It is the first time in the volume that the cut has been shown to enrich someone other than the one who keeps it, and it forces a question the volume has not yet had to ask: when the extraction is real but the middleman blameless, who is the extractor, and is what he takes a theft?
The dealer keeps none of it: the spread the informed force him to set is collected from the uninformed and paid straight out in his losses — a tax the ignorant pay the knowing, with the dealer only the gate it passes through.
The question has no clean answer, because the informed trader is two things at once, and cannot be the one without being the other. Seen from the uninformed trader's side, he is plainly a predator: he wins exactly what they lose, and they lose only because he knew what they did not. But seen from the market's side, he is the agent of its wisdom, for his trade is the very mechanism by which what he knows becomes what the price reflects. When he buys on his knowledge that the thing is undervalued, his buying lifts the price toward its true worth; when he sells on his knowledge that it is overvalued, his selling lowers it; and a market in which the informed could not profit would be a market in which no one troubled to discover anything, and prices would drift, uncorrected, ever further from the truth. The same trade that fleeces the ignorant is the trade that makes the market wise, and the profit that lures the informed to prey upon the uninformed is the identical profit that pays them to gather the information the price needs. One cannot have the price discovery without the predation, for they are the same act seen from two sides — which is why informed trading is at once condemned and indispensable, a tax on the unlucky that is also the wage of the diligent, a theft that is also a public service.
The law, faced with this, draws a line the economics will not honour. It permits — even celebrates — the trader whose edge comes from honest labour, from research and analysis and the patient assembly of public facts into private insight, and it counts his profit the fair reward of price discovery. It forbids — and jails — the trader whose edge comes from stolen secrets, from information he was duty-bound to keep, and it counts his profit a theft. The distinction is a real one, and morally weighty, for it turns on whether the advantage was come by fairly; but it is invisible in the spread itself, for the dealer loses identically to both, and prices identically against both, and the uninformed trader is fleeced identically by both. The economics knows only that some trade with an edge and the rest are taxed to pay for it; whether the edge was earned by study or stolen from a desk is a question the order flow cannot put and the spread cannot answer. So the volume's verdict on this darkest face of the gap must be a double one. The spread itself is no toll of the dealer's; he is the blameless conduit of a tax. But the tax is real, and falls on the uninformed, and is paid to the informed; and whether that payment is a wage for the genuine service of making prices true or a plunder of the ignorant by the privileged depends on a distinction — fairly or foully come by — that lives entirely outside the spread, in the murky law of who was entitled to know.
The informed trader, then, leaves the dealer strangely exonerated and the volume's question strangely displaced. Of the three things folded into the gap, this is the one for which the middleman is least to blame, for he keeps none of it and would gladly be rid of the whole; and it is the one that carries the question of the earned and the extracted clean out of his hands and into the crowd he serves, where it becomes a question about the traders themselves — about who knows, and how they came to know, and whether the knowing is a service to all or a theft from some. The three faces of the dealer's gap are now turned: the wage of immediacy, the premium of the holding, and the tax of the informed. What remains beneath them is a residue the later plates must weigh — the part of any real spread that no service, no risk, and no informed trader can explain, and that is the dealer's toll at last. But the gap is not the only place a middleman stands between two parties, and the next plate leaves the trading floor for a different threshold: the bank that stands between the saver and the borrower, watching the one on behalf of the other, and earning its keep as the delegate of those who cannot watch their money themselves. As the table says of every price, first to last: the argument remains open.