The inside of the stock market

Traders call the best bid and the best offer the inside. The inside market, the inside spread, the inside quote — the tightest prices available, the top of the book, the front of the queue.

The same word does other work. Someone with the inside track. Someone trading on the inside.

That is not a coincidence of vocabulary. The width of the first inside is set, almost entirely, by fear of the second — and following that thread explains both why insider dealing is prohibited and why the prohibition is so difficult to draw a line around.

The fifth participant, returned to

Issue fifty-two set out the reasons people trade: hedging, investing, speculating, intermediating. Then it noted a fifth participant, and left them there.

They look exactly like everybody else. Their order is the same shape as a hedger’s. The difference only becomes apparent afterwards, in the direction the price moves once they are done.

Every quote in every market is written partly in fear of that person. A market maker showing a two-sided price does not know whether the next order is a pension fund rebalancing or somebody who has read tomorrow’s announcement, and cannot find out until after they have traded.

The formal version of this is now forty years old. Glosten and Milgrom showed in 1985 that a bid-offer spread arises from adverse selection alone — with no inventory cost, no processing cost, no capital charge, nothing but the possibility that the counterparty knows more. A quoter facing a mixture of informed and uninformed flow must charge everyone enough to survive the informed ones.

So the inside spread is a measurement of how afraid the market is of the inside information. Widen the second and the first widens with it, mechanically, whether or not anyone is actually trading on anything.

What counts

The prohibition sounds simple and is not, because “knowing something useful” is not the offence. Analysts are paid to know useful things.

Broadly, information is inside information when it is specific enough to act on, not public, and would move the price if it were. Vague pessimism about a sector fails the first test. A widely reported fact fails the second. A detail nobody would care about fails the third.

Where jurisdictions diverge is on what makes acting on it wrong.

The American approach is built on a breach of duty. Trading on such information is prohibited where the trader owed an obligation — to the company’s shareholders, or to whoever entrusted them with the information. A person who overhears a conversation on a train, owing nothing to anybody, sits in genuinely contested territory.

The European and British approach asks less about duty and more about possession. Dealing while holding inside information is prohibited in itself, with defined carve-outs for legitimate behaviour that would otherwise be caught.

Both are defensible and they produce different answers in real cases, which is worth knowing before assuming that any particular trade is obviously fine or obviously not.

The line nobody can draw

The genuinely hard problem is that research and inside information are made of the same material.

An analyst counts lorries leaving a factory. Another interviews a dozen customers about their spending. A third notices a supplier’s order book has thinned. No single observation is material or confidential. Assembled, they produce a conclusion the market does not have, and acting on that conclusion is legal — the mosaic principle, and the entire justification for the research industry.

Now change one detail. The analyst calls someone at the company, who confirms the conclusion. Same conclusion, same trade, and a line has been crossed, because the last piece was neither public nor theirs to have.

The distinction is real. It is also, in practice, a matter of degree that has to be judged after the fact, by people reconstructing what was known and when. This is why compliance departments are large and why firms build walls between the part of the business that talks to companies and the part that trades.

The prohibition does not draw a line around good information. It draws one around a source. What you worked out is yours. What you were told is not.

Why it is prohibited

The fairness argument is the one usually made and the weaker one. Markets are not otherwise organised around equalising advantage — a faster connection, a better model and a larger research budget are all permitted asymmetries.

The mechanical argument is stronger. If uninformed participants come to believe they are systematically trading against people with private information, they do not merely lose money. They widen their quotes, reduce their size, or leave. The spread paid by everybody rises, and the cost falls on participants who never met the insider and have no idea why trading became more expensive.

That is the real damage: not the insider’s profit, which is a transfer, but the withdrawal of the liquidity that made the market work — the thing selective disclosure rules were written to protect when they required companies to tell everyone at once or nobody at all.

The uncomfortable symmetry

And yet a market with no information asymmetry at all would be a market in which nobody was paid to find anything out.

Issue fifty-four made the point: prices are accurate in proportion to how much it is worth somebody’s while to check, and checking is only worth doing if the checker can profit from what they learn. Perfect symmetry would remove the incentive that produces the accuracy.

So the rule is not “no advantages”. It is: advantages you built are yours, and advantages you were given are not. That distinction has no clean edge, which is why this area of law is so heavily litigated — and why the inside spread, the one number that measures the whole problem, will never go to zero.

Three things worth keeping

1. The inside spread prices the fear of inside information. Adverse selection alone is sufficient to produce a bid-offer spread.
2. The offence is about the source, not the quality. Worked out is legal; handed over is not, and the boundary is judged afterwards.
3. The cost is paid in the spread by everyone, not in the insider’s profit, which is only a transfer.

Next week: what a central bank is actually doing when it says it is providing liquidity — and why it is not the same word.