One screen, two entirely different arrangements
Every filled order has two names on it. On one kind of venue you will never learn the second. On another you knew it before you opened the app.
Two arrangements wearing the same screen
Two people open apps on the same phone. One buys two hundred shares in a FTSE 100 company. The other opens a position on that same company as a contract for difference. The confirmations land in the same second, the running profit updates in the same font, and underneath the two arrangements have almost nothing in common.
The difference is who is on the other side. The share buyer has dealt with a stranger, picked by a matching engine on price and time, whose name nobody will ever disclose. The other has dealt with the firm whose logo sits at the top of the screen. Both are ordinary and lawful. Only one is widely understood by the person inside it.
An exchange runs a queue. Buying and selling interest arrive carrying a price and a size, wait in a public book in the order received, and when a bid meets an offer the pair are matched and removed. Your counterparty is whoever reached the front of the opposite queue: a fund rebalancing, a market maker earning the spread, somebody in Falkirk who decided to sell.
The venue itself takes no view. It charges a fee for matching and holds no position in what follows, which is why nobody loses sleep over who operates an exchange. Anonymity of that kind differs from opacity: you never learn the name, but the venue’s income does not move when your position does.
Not every market runs a queue
Britain’s largest financial market barely uses one. In the Bank of England’s October 2025 survey of the UK market, foreign exchange swaps made up 48% of turnover at about USD 1.84tn a day and spot 28% at about USD 1.06tn a day. Both are transacted bilaterally, one party asking another for a price and dealing on it. So a customer facing a named firm rather than an anonymous queue has stumbled on nothing unusual. Dealing against a price somebody quotes you is the ordinary structure of wholesale finance, scaled down and put behind an app. Asking the question in the register of a scandal produces a wrong answer before any evidence.
The instrument is an agreement to settle the change in a price between opening and closing. Agreements need two parties. No share is bought, no listing touched, no register amended, nobody else introduced. The customer signs, the firm signs, and what the customer holds afterwards is an obligation owed by that firm.
This is where the vocabulary of routing misleads people arriving from share dealing. The firm is not standing between you and a market. It is the market, for as long as the contract runs. The price quoted is its price, taken from an underlying it watches but need not match, and the position booked against yours is its own.
Internalisation, and why the matched book is the cheap outcome
Having written the contract, the firm decides what to do about the risk it now carries. The cheapest answer is usually very little, because a large retail book holds customers on both sides of the same instrument at once. Longs cancel shorts, and only the residual, the amount by which one side exceeds the other, needs acting on.
That is internalisation, and it explains why the question so rarely produces a villain. Netting costs nothing, happens automatically inside a risk system, and leaves the firm collecting the spread from two customers while holding no view on price. The residual is then hedged wholesale, kept deliberately, or hedged in part, and which met your ticket was settled after you clicked.
| What became of the position | Who carries the price risk | Where the firm earns | What it would take for you to find out |
| Netted against another customer’s opposite position | Nobody; the clients offset | Spread and financing from both | A snapshot of the firm’s client book at the second you dealt |
| Hedged with a wholesale provider | The provider, at a price the firm pays | The gap between the two prices | A confirmation between two firms, addressed to neither of you |
| Kept on the firm’s own book | The firm | The other side of your outcome | The internal exposure report, which is nobody’s to hand out |
| Part netted, part hedged, rest kept | Split three ways, re-split hourly | All three at once | All of the above, timestamped, and the split still moves |
| Routed to an exchange, as in share dealing | Whoever’s resting order you met | Commission and conversion | The contract note, which names the venue |
Five things that can become of a position once the contract exists, and the document that would settle which of them did.
The charges do not read your profit and loss
There is a folk version of this in which the broker simply wants you to lose. It survives because it is nearly right about the mechanism and careless about where the money comes in. The lines that pay for the building are the spread crossed going in and coming out, financing on every night a leveraged position survives, and conversion on anything priced abroad.
Look at what each of those is charged against. The spread against the transaction, financing against the exposure and the nights, conversion against the amount converted. Not one consults the result. A winning position and a losing one of the same size, held the same length of time, generate the same bill either way.
Your outcome reaches the firm only through whatever slice of the residual it happened to be holding, a decision taken after your ticket by somebody not thinking about you.
The firm writes the contract, so it decides what exists
Because the contract belongs to the firm, so does the range of contracts. Published comparison testing of three FCA-regulated brokers recorded 17,000 tradable markets at one, 12,000 at another and 1,350 at the third. A count of that kind measures nothing about access to venues. It counts what each firm will write, describing three appetites rather than three grades.
The same authorship runs through the pricing. The firm sets the quote you deal on, sets the tolerance inside which an order is filled rather than rejected, and chooses how much leverage to offer beneath the regulatory ceiling. Each is a judgement by a party that may hold the opposite side, which is a description rather than an allegation.
The hedging book is the page nobody publishes
Which of those five outcomes met your ticket last Tuesday is settled inside a risk system no outsider can see.
No regulated provider publishes its net position by instrument, and none is likely to, since aggregate exposure is among the most commercially sensitive figures a dealing firm holds. So this is an observation about a structure, and no charge against any firm.
Funding an account does not close the gap either, which is worth saying from inside a method built on funded accounts. Deposit money, deal, hold overnight, time the withdrawal, and you obtain cost, fill quality and operational conduct, all real findings. None reaches the disposition. The Investors Centre opens and funds live accounts with its own money to test UK trading platforms rather than compiling rankings from providers’ published fee schedules, and that method covers everything on the customer’s side of the contract and nothing whatever on the other.
Which documents name the other party
The hedging book is unobtainable. The arrangement is published. A client agreement states whether the firm deals as principal, taking the other side, or as agent, passing the instruction on and charging for it. An order execution policy describes where orders are executed, including the case where the venue is the firm you were already dealing with.
Reading them alters nothing about the market and a good deal about what can surprise you later.
It also reframes a choice larger than the one between two firms. Moving to something exchange traded lowers neither risk nor cost; it swaps a counterparty with a stake in your position for one with a stake only in the price. Whether that is worth its cost is a question for you, not for whoever quoted you.