TL;DR
Market impact is the price movement caused by your own trading. Temporary impact is the part that decays after you stop, reflecting the cost of demanding immediate liquidity. Permanent impact is the part that persists, reflecting information your trading revealed.
Transaction cost analysis, or TCA, is the measurement discipline around this: comparing executed prices to benchmarks to understand what trading actually cost. For most strategies, costs and capacity decide profitability more than prediction accuracy does.
Key points
- Impact grows with size relative to available liquidity, which caps capacity.
- Temporary and permanent impact have different causes and different remedies.
- TCA benchmarks each answer a different question; choose deliberately.
- Cost assumptions must be fixed before testing, not tuned afterwards.
Components of trading cost
| Component | What it is | How to reduce it |
|---|---|---|
| Spread cost | Paying the bid-ask spread to trade immediately | Trade passively, accept fill uncertainty |
| Fees and rebates | Exchange, clearing and broker charges | Venue and order type selection |
| Temporary impact | Price moves against you while you trade | Slow down, split the order, use more venues |
| Permanent impact | Price shifts because trading revealed information | Reduce footprint, randomise, trade less |
| Delay or opportunity cost | Price moves while you wait | Trade faster, accept more impact |
| Adverse selection | Being filled by better-informed traders | Manage queue position, widen quotes |
Execution is the trade-off between these: trading faster reduces delay cost and increases impact, and trading slower does the reverse.
TCA benchmarks
| Benchmark | Question it answers | Weakness |
|---|---|---|
| Arrival price | What did the decision cost from the moment we decided? | Sensitive to short-term volatility |
| VWAP | Did we trade near the day's volume-weighted average? | Gameable, ignores decision timing |
| TWAP | Did we spread trading evenly through time? | Ignores volume patterns |
| Close price | Did we beat the closing mark? | Only relevant to specific mandates |
| Implementation shortfall | Total cost from decision to completion, including unfilled quantity | Requires careful bookkeeping |
Implementation shortfall is generally the most honest measure for alpha-driven trading, because it includes both what you paid and what you failed to capture.
Why capacity follows from impact
As order size grows relative to typical volume, impact grows faster than linearly in most empirical models. A signal earning a given edge per unit traded eventually pays all of that edge away in impact. That point is the strategy's capacity, and it explains why very high Sharpe ratio strategies often run small books. See Sharpe and Sortino.
Doing cost analysis honestly in research
- Fix the cost model in advance, agreed at team level rather than per experiment.
- Test cost sensitivity: results at base, double and triple assumed costs.
- Model fills realistically for passive strategies, including queue behaviour.
- Include delay, not only executed cost.
- Report capacity alongside returns, not as an afterthought.
Who works on this
Execution research and TCA sit between trading, research and technology. Firms hire execution-focused quants, TCA analysts and algorithm developers for this work, especially at brokers, banks and funds trading larger sizes. See algorithmic trading recruitment.
Frequently asked questions
What is market impact?
The price movement caused by your own orders, split into temporary impact that decays and permanent impact that persists.
What is transaction cost analysis?
The measurement of execution quality against benchmarks such as arrival price, VWAP or implementation shortfall, used to understand and reduce trading costs.
Why does impact limit strategy capacity?
Because impact grows with size relative to liquidity, eventually consuming the strategy's edge. The size at which that happens defines capacity.
Which TCA benchmark is best?
Implementation shortfall is usually the most complete for alpha-driven trading, because it captures both execution cost and opportunity cost.