Decomposing Execution Cost: The Five Components Institutional Desks Actually Track

Most desks know they pay spread. Fewer can tell you, on a Friday, what their week cost them in market impact, what it cost them in adverse selection, and what it cost them in queueing. That asymmetry is not an accident. The number that is easiest to report is the spread, so the spread is what gets reported; the numbers that are harder to attribute are the ones that quietly dominate the P&L.

The cheapest thing a counterparty can do for an institutional desk is hand back the full decomposition. Drovix does. The point of this post is to walk through the five components we surface, why each one is real, and what to do operationally when one of them moves.

1. Spread cost

This is the obvious one: the distance between the side you crossed and the mid at the moment of the cross. It is also, on aggregate, the smallest component for any serious volume. A desk paying half a pip on $50m of EUR/USD spends $2,500. A desk that gets read on the same order and slips a quarter pip on the second clip pays $1,250 a clip on every clip after the first — and there are usually a lot of clips.

Spread alone is a misleading benchmark. The full pricing context — why spreads compress in calm conditions and widen synchronously in stress — is covered in The Architecture of a Fair Spread. Read that one if you want to understand what you are paying for; read this one if you want to understand what you are not.

2. Market impact

Market impact is the move that your own order creates in the price between the first child and the last. It is real, it is measurable, and on any order that takes more than a few hundred milliseconds to complete it is usually larger than the spread.

The honest way to measure impact is to compare the volume-weighted average price you achieved against the arrival mid — the mid that was on the screen at the moment you decided to trade. Anything else (interval VWAP, TWAP-since-decision) is either flattering or noisy.

Drovix reports arrival-mid impact on every parent order, broken out by child clip, so you can see where the impact actually accreted. In our experience, 70% of total impact lives in the first 30% of the order — which is also why routing matters more than people who haven't measured it think it does.

3. Signalling cost

Signalling is the cost of being read. It is the move the market makes against you AFTER your last fill, in the seconds and minutes during which a counterparty who internalised your flow can act on what they just learned about you.

This component is invisible on conventional TCA reports because conventional TCA stops measuring when the parent order completes. It is also, for any desk that trades the same direction repeatedly during a session, the largest single contributor to long-run drag.

We discuss the mechanics of why signalling happens, and what the routing architecture has to look like to suppress it, in Routing Beyond the Inside Quote. The short version: a counterparty that cannot route your residual outside its own book is structurally going to leak your intent.

3a. The 30-minute post-trade window

Drovix reports the move in the mid for 30 minutes after your final fill, in your direction, on the same instrument. It is one number. If that number is large and consistently in the same sign as your direction, your counterparty is leaking — full stop. The recourse is not to argue about TCA methodology; it is to route differently.

4. Adverse selection

Adverse selection is the spread component you pay because, on any given fill, you might be the better-informed side. The counterparty quoting to you does not know whether you are about to be right or wrong about the next tick — so it charges a small premium on every quote to insure against the times you turn out to be right.

For desks running directional strategies on macro data, that premium can be substantial. For desks running mean-reversion or market-making strategies, the SIGN of adverse selection reverses: you are the one collecting the premium, and the relevant question is whether your counterparty is paying you enough of it.

5. Queueing cost

Queueing is the opportunity cost of resting orders that don't fill. A passive order that sits in the book and watches the price move away has paid for its place in the queue with the slippage you took on the order you then had to send aggressively to keep up.

It rarely shows up in PnL attribution as a discrete line item, because it usually appears disguised as larger spread cost on the aggressive follow-up. Drovix's TCA infers it by tagging every fill with whether it was passive or aggressive, and reporting the opportunity cost of the unfilled passive children separately.

Putting it together

On a typical week of institutional flow we see the components roughly in the proportions: spread 15-20%, market impact 25-35%, signalling 20-30%, adverse selection 10-20%, queueing 10-15%. These are not universal — a desk that trades only on macro releases will see signalling collapse and adverse selection explode; a desk running a passive-only style will see queueing dominate.

What is universal is that the dominant component is almost never the one a retail-grade report would surface. If your current broker's TCA tells you that you 'paid 0.4 pips average spread on the week' and nothing else, you are reading roughly 18% of your real cost. The other 82% is in the report you are not getting.

What to do operationally

• Demand arrival-mid TCA on every parent order. Anything that anchors to a within-execution benchmark is grading itself.
• Demand a post-trade window report — Drovix uses 30 minutes; some desks prefer 5 or 60. The number matters less than that you have one.
• Tag each strategy with its expected dominant cost component, then check weekly that the dominant component is the expected one. If it isn't, something has changed about how your flow is being handled.
• Treat any counterparty that won't break the report down at this granularity as a counterparty that is not confident in its own numbers.

Where to go next

→ Last Look in 2026 — what last-look symmetry actually means for your fill rate, and how to negotiate it.

→ Signed Flow and Information Half-Life — the mathematics of why some flow ages faster than others, and why your counterparty should price that explicitly.

Drovix institutional desks get the full TCA report by default. To request an evaluation file or a sample week of reporting on your historical flow, write to institutional@drovix.com.

Methodology and data

The five-component framework is calibrated against an evaluation sample of 1.8 million parent orders executed across Drovix's institutional desks during the trailing twelve months, decomposed into: spread cost, market impact, timing risk, opportunity cost, and operational friction. Decomposition follows the implementation-shortfall convention with the arrival mid as anchor and 5-minute post-trade mid as reversion benchmark. Per-strategy decompositions are produced weekly for every counterparty.

Limitations and scope

Implementation shortfall is a model — not a measurement. The decomposition is sensitive to the choice of arrival benchmark, reversion window, and treatment of intra-order child-clip refresh quotes. Two TCAs run by two desks on the same fills will disagree at the second-decimal level. The framework is most useful as a longitudinal tool within a single desk, less so as a cross-desk benchmark.

Further reading

→ The Anatomy of an Effective Spread — The microstructural counterpart of the TCA decomposition. See /blog/effective-spread-2026-microstructure-survey.

→ Capacity Planning for Execution Strategies — How desks budget per-component cost against expected strategy capacity. See /blog/capacity-planning-for-execution-strategies.

Drovix Research is the in-house institutional desk of Drovix MU Ltd, regulated by the Financial Services Commission of Mauritius. All notes are informational only and do not constitute investment advice, a solicitation, or a recommendation to transact in any financial instrument.