What an execution model means for you

The same instruction, answered differently 

The previous lesson described a firm's arrangements from the firm's side: having written a contract with a client, it decides whether to carry the resulting exposure on its own book, net it against the opposite exposure of other clients, or pass it to an external counterparty. That is a question about the firm's risk, and it is settled inside the firm where nobody outside can watch it happen. This lesson is the same set of arrangements seen from the other end, from the order's side, which is the only side a client observes directly.

The arrangement is usually referred to by a label, and the label is worth less than the mechanism underneath it. Labels are marketed, applied inconsistently across the industry, and a single firm commonly runs more than one arrangement at once. The mechanism does not move: in every arrangement there is exactly one question, which is which party decides the price an instruction fills at. Everything a client can observe about execution follows from the answer.

What a client can actually see 

None of the arrangement is visible on a screen. What is visible is the record a platform writes for every instruction it handles, and that record supports four observations and no others.

• The price at which each instruction filled, set against the price displayed at the moment it was sent.
• Whether the difference between those two prices falls on both sides of the requested price across many orders, or only on one side.
• Whether instructions are ever declined outright, or returned with a different price attached instead of a fill.
• Whether any of the above changes with the size of an order, with the instrument, or in the minutes around a scheduled announcement.

Those four are measurements. Everything else that gets said about execution is either an inference drawn from them or a matter of disclosure, and the two are not the same kind of thing. How the measurements are properly constructed, what a sample has to hold constant before it means anything, and what an average of them conceals are the subject of a later lesson in this module. The rest of this one connects each observation to the arrangement that produces it.

Which party decides the fill price 

In the first arrangement the firm is itself the counterparty to the fill. The instruction is checked against the price the firm is currently showing, and if that price is still standing the contract is written at it. What the firm subsequently does with the exposure, whether it retains it, nets it against clients positioned the other way, or hedges it externally in its own time, happens after the fill and does not change the price the client received. The fill was the firm's own quote, honoured.

In the second arrangement the firm's quote is derived from external counterparties, and the instruction is worked against them. The fill is then the outcome of a transaction that took place somewhere else, at whatever price was standing there when the order arrived, and the firm passes that price back. The firm is still the counterparty to the client's contract, since a CFD is bilateral and that does not change, but it is not the party that determined the number.

The consequence is a difference in what constrains the fill, and it is the root of everything else in this lesson. Where the firm is quoting its own price, the constraint is the firm's willingness to stand behind that quote at that moment, and the firm can hold a price through an instant when no external price is available at all. Where the fill is the outcome of an external transaction, the constraint is what the external book actually held when the order arrived, including the possibility that the source showing the best price declines the order inside the brief window described earlier in this module. One arrangement concentrates the decision in the firm. The other distributes it across counterparties the client never sees.

Firms are rarely one thing. The same firm commonly quotes as principal on some instruments and works orders externally on others, and changes the mix as conditions change through the day. A firm described by a single label is almost always described inaccurately, which is why the arrangement belongs in a published order execution policy rather than in a headline.

Slippage, and why symmetry is the question 

Slippage is familiar from the costs module as an event: an instruction filled at a price other than the one requested. Stated arithmetically it is a single quantity, the fill price minus the requested price, and it carries a sign. Where the sign favours the position it is conventionally called price improvement, and where it does not it is conventionally called negative slippage. The industry uses two names for one subtraction, and the naming is worth noticing, because it is the point at which the two arrangements above stop behaving the same way.

Where the fill is the outcome of an external transaction, the sign is whatever the external book produced. Prices move in both directions in the interval between an instruction being sent and arriving, so fills land on both sides of the requested price, and the balance between the two sides is a property of the market and of how the order was worked rather than a decision anyone took about that order.

Where the firm decides the fill price itself, symmetry stops being an outcome and becomes a policy. A firm in that position can pass a favourable difference back in full, pass it back only up to a stated cap, or retain it while continuing to pass adverse differences on. All three practices exist in the industry and all three are lawful where they are disclosed. None of them is distinguishable from the others in a single record, because an order that filled at exactly the requested price leaves a record with no difference in it, and nothing in that record shows whether an improvement was available and was not passed on.

The two halves of that block differ in one respect only, which is what happened to the twenty instructions that could have filled better. The market was identical, the adverse cases were identical, and every individual fill in both halves is a real price at which a contract was genuinely written. The difference is invisible in any one record and only appears in the distribution across many of them, which is the whole reason symmetry is asked about as a policy question rather than measured from a handful of fills.

What a rejection or a requote indicates 

A rejection is a firm declining to fill an instruction at all. A requote is a firm declining at the requested price and returning a different price in its place, leaving the original instruction unexecuted. Mechanically they are the same event with two different responses, and the event is this: between the instruction being written and the instruction arriving, the price it was written against stopped being available, and the firm did not fill through the gap.

That single event has several ordinary causes, and a rejection message rarely distinguishes between them.

• The quoted price moved before the instruction arrived, by more than any tolerance the instruction carried, or the instruction carried no tolerance at all.
• An external counterparty declined the order inside its last look window, so the price the fill was to be worked against ceased to exist between the check and the attempt.
• The size requested exceeded the size standing behind the quoted price, so no single price covered the whole order.
• The margin held against the account was insufficient for the position the fill would have opened, measured at the price at which it would have been written.
• The instrument was closed, halted or in an auction period at the moment of arrival, so no continuous price existed to fill against.
• A pre-trade control at the firm or at the venue stopped the order before it reached a price at all.

A rejection therefore indicates a condition rather than a verdict. It reports that something which made the fill possible was no longer true, and on its own it is evidence of neither bad faith nor good faith on the part of the firm.

The more useful observation is that rejection and slippage are two answers to one question, and they trade directly against each other. A firm that fills through a moved price rejects fewer instructions and slips more of them. A firm that holds strictly to its quoted price rejects more and slips less. Neither behaviour is a fault, both are defensible, and a firm can move its own rejection figure substantially by changing a tolerance setting rather than by changing anything about the quality of what it delivers. That is why the two numbers carry information only when they are read together, and why neither of them ranks firms on its own.

Where the arrangement shows up in cost 

The costs module set out the structures in which a firm charges for a transaction: a single spread with no separate charge, a narrower spread with a commission billed alongside it, or a commission priced on the notional value of the contract. The arrangement behind the fill does not determine which structure a firm uses, but the two are related, because they describe where the firm's revenue sits and therefore what a figure quoted in isolation actually covers.

A firm quoting its own price earns from the width of that price. A firm working orders externally earns from the commission it bills and from any adjustment applied to the external price before it is shown, and an adjustment of that kind is a cost sitting inside the quote rather than on the statement. Neither structure is cheaper than the other by construction: one moves the charge into the price and the other states it separately. The only comparable figure is the total cost of opening and closing one position, in one instrument, at one size, with every component added together.

Questions with documentary answers 

The useful questions about a firm's execution are the ones whose answers exist in a document rather than in a claim. Five of them cover most of what this lesson has described.

1. Which legal entity is the counterparty to the contract, and under which regulator's rules does that entity execute orders?
2. Is the firm itself the counterparty to the fill, or is the fill the outcome of a transaction with an external counterparty, and does the answer differ by instrument, by account structure or by market condition?
3. Is a difference between the requested price and the fill price passed on in both directions, and is a favourable difference passed on in full or capped?
4. On what grounds is an instruction declined or returned with a new price, and is the proportion of instructions that are declined published anywhere?
5. How is cost structured, and what does the total cost of a round turn come to once the spread, any commission and any adjustment applied to the quote are added together?

What makes those questions useful is that each has an answer of a checkable kind, in an order execution policy, in a terms of business document, or in the contract specifications for an instrument. Those documents are published by the legal entity a client actually contracts with rather than by a brand or a website, and at YAL that entity is Yal Group Inc.. The distinction matters because regulatory obligations attach to entities, not to names on a page. A firm that answers any of the five in general terms, without reference to a document, has answered a different question from the one asked.

Where practitioners disagree 

Whether one arrangement is structurally better for clients than the other is the oldest argument in this industry, and it is genuinely unsettled. One tradition holds that passing every order to an external counterparty removes the conflict at its source, since a firm that retains no exposure has no position that benefits from a client's loss, and that no amount of supervision substitutes for removing the incentive. The opposing tradition holds that an externally worked order is exposed to whatever the external book contains at that instant, so declined orders and fills away from the requested price arrive more often precisely when conditions are worst, while a firm quoting its own price can hold that price through a moment when no external price exists at all. It adds that the conflict in the first model is addressed by supervision, by best execution obligations and by client money rules rather than being unaddressed.

Both descriptions are accurate about different circumstances, which is why the argument has not resolved and is unlikely to. This page takes no position on which arrangement suits any particular set of circumstances, because it knows nothing about any reader's circumstances and no general answer to that question exists. What can be said without taking a side is narrower: the arrangement determines which party decides a fill price, and that shapes the pattern of fills, declines and costs a client observes. A firm's disclosures are where the arrangement is stated, and a client's own records are where its effects appear.

A second and quieter disagreement is about measurement. One camp argues that firms should publish standardised execution statistics so that arrangements can be compared at all, on the reasoning that an imperfect common measure is better than none. The other argues that a single averaged figure conceals the distribution underneath it, that the moments which matter to a client are the tail cases rather than the typical ones, and that a published average invites competition on the number instead of on the handling. What execution quality actually measures, and what an average of it hides, is the subject of the next lesson but one.