In simple terms — An aggressive order consumes liquidity and may move price in its own direction: a buy pushes upward and a sell pushes downward. Calling the entire fill deviation “impact” is incorrect, however, because the market may also move for reasons independent of the order during the same interval.
Market impact is the component of a price change caused by an order, sequence of orders, or execution program. It is an endogenous cost: it rises because a participant is trying to trade. The concept is central when quantity is material relative to available liquidity, but it can also arise from a small order in a thin book.
Impact is a causal and therefore counterfactual quantity. Measuring it perfectly would require knowing the price that would have formed in the same market at the same time without that order. This scenario cannot be observed. Empirical measures are estimates conditional on a benchmark, horizon, dataset, and model.
Market impact, spread, and slippage
| Term | What it measures |
|---|---|
| distance between the best bid and best ask at a given time | |
| difference between execution price and a declared benchmark | |
| Market impact | portion of the movement caused by the order itself |
| Market drift | movement that would have occurred without the order |
| Fee | explicit charge imposed by a broker, venue, or service |
Slippage can be observed once the benchmark is chosen; impact requires attribution. If a buy executes above the arrival midpoint, the deviation may include spread crossing, consumption of several levels, delay, independent quote changes, and impact. Adding those components without a methodology creates double counting.
A price change immediately after a fill does not prove causality by itself. News, another order, or flow shared by several participants may move it in the same direction. Attribution becomes harder as the market grows more volatile and fragmented.
Mechanism in a limit order book
A marketable order meets resting passive orders. If its quantity exceeds the amount available at the best price, it crosses subsequent levels. This book walking worsens the immediate average price. The execution may also prompt other participants to cancel, reprice, or add liquidity.
Example — The initial midpoint is 100.00. The ask side shows 100 units at 100.01, 300 at 100.02, and 600 at 100.04. If the book does not change, an aggressive buy of 1,000 units has an average price of 100.031. Its deviation from the midpoint is 0.031, or 3.1 bps. The number describes execution against that snapshot; by itself, it cannot separate spread, impact, and independent market movement.
Displayed depth is incomplete: reserve orders, other venues, cancellations, and new liquidity can alter the outcome. Research by Cont, Kukanov, and Stoikov documents, for a specific sample of U.S. equities and short horizons, a relationship among price changes, order-flow imbalance, and depth. It is sample-dependent empirical evidence, not a universal law with coefficients that can be transferred unchanged to every market.
Immediate, temporary, and persistent impact
Execution analyses often distinguish:
- immediate impact, observed while the order consumes or changes quotes;
- a temporary component, which declines as the book replenishes;
- a persistent or permanent component, associated with a more durable revision of price.
The distinction is useful in models but is not directly observable trade by trade. A persistent move may reflect information incorporated into price, correlated orders from other participants, or general drift in addition to the order under review. Results also depend on the horizon: what appears permanent after one minute may reverse after one hour, or vice versa.
The realized spread and the movement of the midpoint after a trade are possible proxies for adverse selection and impact, provided that horizon and convention are stated. The SEC's 2022 Order Competition Rule proposal, for example, discusses price impact through midpoint changes at later intervals. It is a reference for U.S. equity-market analysis, not a universal regulatory definition.
Factors that affect impact
Expected impact depends on the interaction between the order and market state:
- size relative to depth and available volume;
- urgency and price aggressiveness;
- participation rate relative to concurrent flow;
- volatility, spread, and book resiliency;
- time of day, information events, and market regime;
- fragmentation and routing quality;
- predictability of the order sequence;
- priority rules, tick size, and supported order types.
Percentage of daily volume and ADV are useful normalizations, but they may hide intraday liquidity. Executing 2% of average daily volume in one minute is not equivalent to distributing it across the session. Two orders of the same size may also have different impacts if they arrive under different depth and volatility regimes.
The execution trade-off
Slowing down and using order splitting may reduce immediate book consumption, but it increases exposure to market movement and the risk of non-completion. A passive limit order avoids accepting a price worse than its limit, but may remain unfilled and may face adverse selection. An iceberg order reduces displayed size without concealing the entire sequence of fills and refreshes.
TWAP and liquidity-seeking algorithms organize time, venues, and aggressiveness. They do not eliminate impact; they shift the trade-off among:
- immediate cost and price risk while waiting;
- completion probability and limit control;
- order visibility and speed;
- mechanical impact and potential information leakage.
The Almgren–Chriss model formalizes a trade-off between expected impact cost and volatility risk during execution. Its parameters must be estimated from market data and updated. They are not constants that can be transferred automatically between instruments.
Pre-trade and post-trade measurement
A verifiable estimate records at least size, side, decision and submission timestamps, quotes and depth, venue, every fill, unexecuted quantity, and the midpoint at predetermined horizons.
Before the trade, a model can estimate expected cost as a function of size, urgency, and conditions. After the trade, transaction cost analysis compares average price and the market path against declared benchmarks. Results should be segmented by instrument, session, volatility, and participation rate.
For an execution program with several fills, it is useful to report:
- total cost relative to arrival price;
- spread crossing under a consistent convention;
- midpoint movement during and after execution;
- explicit fees;
- residual quantity and its opportunity cost.
The decomposition remains an estimate. Reporting intervals, distributions, and sensitivity to the benchmark is more defensible than attributing every basis point to the order alone.
Common mistake — Calling any negative slippage market impact. Slippage is an observed deviation; impact is only the component caused by one's own flow and requires an attribution model.
Sources
- BIS, Committee on the Global Financial System — Structural Aspects of Market Liquidity from a Financial Stability Perspective
- SEC — Proposed Rule, Order Competition Rule, Release No. 34-96495
- Rama Cont, Arseniy Kukanov, and Sasha Stoikov, The Price Impact of Order Book Events, Journal of Financial Econometrics, 12(1), 2014, pp. 47–88; preprint
- Robert Almgren and Neil Chriss, Optimal Execution of Portfolio Transactions, Journal of Risk, 3(2), 2001, pp. 5–39