Price Impact¶
Overview¶
Price impact is the change in asset price caused by trading activity. Large orders move the market against the trader, increasing execution costs. Understanding and modeling price impact is essential for institutional trading.
Difficulty advanced
Components of Price Impact¶
Temporary Impact¶
Temporary: Price moves during execution but reverts after
Caused by: Liquidity providers demanding compensation for risk
Persists: Minutes to hours
Example:
Buy 100,000 shares → price rises $0.10 during execution
→ price falls back $0.05 within 30 minutes
Temporary impact: $0.05
Permanent Impact¶
Permanent: Price moves and stays at new level
Caused by: New information conveyed by trading
Persists: Indefinitely
Example:
Large buy order signals informed trading → market reprices
→ new equilibrium price is permanently higher
Price Impact Models¶
Square Root Model (Industry Standard)¶
Impact = σ × √(Q / V) × Y
Where:
σ = Volatility (daily)
Q = Order size (shares)
V = Average daily volume
Y = Constant (typically 0.5-1.5)
This model captures:
- Impact increases with order size
- Impact scales with square root (not linearly)
- Impact decreases with liquidity (volume)
- Impact increases with volatility
where:
σdaily volatility ·Qorder size ·Vaverage daily volume ·Ymarket-specific impact constant. does: the empirical regularity holding across asset classes — impact grows with √(participation), not linearly. Calibrate per-name from execution history; used in pre-trade cost estimation and to choose between execution algos.
Linear Impact Model¶
Impact = η × Q + ε
Where:
η = Price impact coefficient
Q = Order size
ε = Noise
Simpler but less accurate for large orders.
where:
ηper-share impact coefficient ·Qorder size ·εresidual noise. does: first-order approximation suitable for small orders well within a few percent of ADV. Use it for back-of-envelope cost checks; switch to square-root once orders cross liquidity thresholds.
Almgren-Chriss Model¶
Minimize: Expected Cost + λ × Variance
Expected Cost = Σ (τ_n × f(v_n) + g(S_n))
Where:
τ_n = Time interval
v_n = Trading rate
f(v_n) = Temporary impact function
g(S_n) = Permanent impact function
λ = Risk aversion parameter
Optimal trajectory balances:
- Fast execution → high impact, low timing risk
- Slow execution → low impact, high timing risk
where:
τ_nslice time interval ·v_ntrading rate in slice n ·f(v_n)temporary impact function ·g(S_n)permanent impact function ·S_nshares remaining ·λtrader risk-aversion parameter. does: the canonical mean-variance optimal-execution objective. Increasingλaccelerates the schedule (front-loaded, more impact, less timing risk); decreasingλflattens it. Foundation for Implementation Shortfall execution algos.
Reducing Price Impact¶
1. Algorithmic Execution¶
Split large orders into smaller child orders using VWAP, TWAP, or implementation shortfall algorithms.
2. Dark Pools¶
Execute in venues without pre-trade transparency to minimize signaling.
3. Iceberg Orders¶
Display only a portion of the order, refresh as filled.
Measuring Price Impact¶
Implementation Shortfall¶
IS = (Execution Price - Decision Price) × Shares + Commissions
Components:
- Delay cost: Price moved before execution started
- Market impact: Price moved during execution
- Timing cost: Adverse price movement over execution period
where:
Execution Priceaverage fill price ·Decision Pricearrival/benchmark price at the moment the order was placed ·Sharesorder size ·Commissionsexplicit fees. does: the industry-standard execution-cost benchmark. Decomposes total cost into delay, impact, and timing — used in post-trade TCA and as the loss function that Implementation Shortfall algos minimize.
Checklist¶
- [ ] Order size relative to ADV assessed
- [ ] Price impact model calibrated to instrument
- [ ] Execution algorithm selected based on urgency
- [ ] Dark pool availability checked
- [ ] Spread vs. impact trade-off analyzed
- [ ] Market conditions considered (vol, volume)
- [ ] Post-trade impact analysis performed
- [ ] Implementation shortfall calculated
- [ ] Benchmark selection appropriate
References¶
- Almgren, R. & Chriss, N. (2000). "Optimal Execution of Portfolio Transactions." Journal of Risk, 3, 5-39.
- Hasbrouck, J. (2007). Empirical Market Microstructure. Oxford University Press.
- Kissell, R. (2013). The Science of Algorithmic Trading and Portfolio Management. Academic Press.