Gatheral (2010) Transient Market Impact & Resiliency Lab

Model optimal institutional liquidation, decaying order book resilience kernels \(G(\tau)\), post-trade price relaxation, and Jim Gatheral's fundamental No-Dynamic-Arbitrage theorems.

Expected Cost (IS) -- -- bps
Execution Risk -- Variance of Capture
Savings vs. TWAP -- Optimized schedule
Resilience Half-Life -- --

Portfolio & Resiliency Parameters

6.5 hrs
20%
6.0 h⁻¹
$2.5 / 100k
$0.5 / 100k
1.0 × 10⁻⁶
Gatheral (2010)
Peak Price Impact: --
Post-Trade Rebound: --
Dynamic Arbitrage: Absence Guaranteed

Theoretical Foundations: Transient Impact & Dynamic Arbitrage

1. Transient Resiliency Kernel Formulation

In standard Almgren-Chriss (2000) execution models, temporary impact vanishes instantly upon the completion of each trade slice. Jim Gatheral (2010) generalized this by recognizing that order books possess finite resilience: limit orders take time to replenish. The asset price \(S(t)\) at time \(t\) follows:

\[ S(t) = S_0 - \int_0^t \dot{x}(s) G(t - s) ds + \sigma W(t) \]

where \(\dot{x}(s)\) is the selling rate and \(G(\tau)\) is the impact decay kernel. For exponential order book replenishment with decay rate \(\rho\) and permanent impact \(\gamma_{\text{perm}}\):

\[ G(\tau) = \lambda e^{-\rho \tau} + \gamma_{\text{perm}} \]

The characteristic resilience relaxation time is \(t_{\text{rel}} = 1/\rho\), giving an impact half-life of \(t_{1/2} = \frac{\ln 2}{\rho}\).

2. Gatheral's Theorem & No-Dynamic-Arbitrage

A central contribution of Gatheral (2010) is establishing the conditions under which transient impact models prevent price manipulation (profitable round-trip trading strategies):

\[ \mathbb{E}[\text{PnL}_{\text{round-trip}}] = -\int_0^T \dot{x}(t) \left( \int_0^t \dot{x}(s) G(t - s) ds \right) dt \le 0 \]

Gatheral proved that if temporary impact is non-linear (e.g. \(h(v) = v^\delta\) with \(\delta < 1\)), decaying impact kernels will inevitably permit dynamic arbitrage—traders can manipulate prices by buying slowly and dumping rapidly to extract free cash. Absence of price manipulation strictly requires the impact kernel to be positive definite, establishing the linear impact model as the unique arbitrage-free formulation for decaying kernels.

Dual Institutional Sensitivity Analysis

Sensitivity 1: Resiliency Decay (ρ) vs Trade Size (X₀) on Shortfall (bps)

Analyzes how order book recovery speed \(\rho\) and block trade volume \(X_0\) influence the expected percentage implementation shortfall cost.

Decay \ Size 50k shs 100k shs 150k shs 200k shs 250k shs

Sensitivity 2: Volatility (σ) vs Horizon (T) on Initial Velocity v₀ (shs/hr)

Measures how asset volatility and time constraints shift execution urgency, altering the optimal initial liquidation velocity \(v_0^*\).

Vol \ Horizon 1.0 hr 2.0 hrs 4.0 hrs 6.5 hrs 8.0 hrs

Institutional Mastery Quiz: Gatheral Transient Impact

1. What fundamental limitation of the Almgren-Chriss (2000) framework does Gatheral's (2010) transient model address?

2. What did Jim Gatheral prove regarding non-linear temporary impact (e.g. \(h(v) = v^\delta\) with \(\delta < 1\)) paired with decaying resilience kernels?

3. For an exponential decay kernel \(G(\tau) = \lambda e^{-\rho \tau}\), what is the characteristic relaxation time scale of the limit order book?

4. When risk aversion is very low (\(\phi \to 0\)), what characteristic geometric profile does the optimal trade rate \(v^*(t)\) display under transient impact?

Frequently Asked Questions

Transient market impact represents price displacement caused by an execution that neither vanishes instantaneously nor persists indefinitely. Instead, it builds up during active trading and decays exponentially or via power law over a characteristic resilience timescale as limit order queues replenish.

The Almgren-Chriss (2000) framework assumes that temporary market impact dissipates immediately after each trade interval (infinite resiliency). Gatheral (2010) incorporates finite order book resiliency via a time-decay kernel G(tau), capturing how past trades continue to push prices against subsequent execution slices.

Jim Gatheral proved that a market impact model is economically viable only if no trading strategy can generate positive expected profits on a closed round-trip trade (buying and subsequently selling the same number of shares). He proved that non-linear temporary impact with decaying kernels generates price manipulation unless impact is strictly linear.

For an exponential resilience kernel G(tau) = lambda * exp(-rho * tau), the characteristic relaxation time is t_rel = 1 / rho, and the impact half-life is ln(2) / rho. In liquid US equity markets, resilience half-lives typically range between 5 and 45 minutes.

In the absence of high risk aversion, minimizing total transient impact cost leads to higher trading velocities at the beginning and end of the execution horizon. Trading more aggressively at the start allows the order book time to relax during the middle lull, while trading at the very end avoids suffering self-impact during post-trade periods.

Related Quantitative Execution & Microstructure Simulators

Almgren-Chriss Optimal Execution Lab

Model optimal portfolio liquidation frontiers and execution risk under instantaneous impact.

Order Flow Imbalance (OFI) Lab

Decompose limit order book event transitions and quantify contemporaneous high-frequency price impact.

Avellaneda-Stoikov Market Making Lab

Optimize high-frequency limit order quoting schedules and inventory risk indifference pricing.