The Impact of Heterogeneous Trading Rules on the Limit Order Book and Order Flows
The Impact of Heterogeneous Trading Rules on the Limit Order Book and Order Flows
This paper presents an agent-based model of an order-driven market where heterogeneous traders (fundamentalists, chartists, and noise traders) submit limit and market orders based on utility maximization. The model extends prior work by allowing agents to choose order sizes endogenously via a CARA utility function, rather than using fixed unit orders. Simulations reveal that chartist strategies are crucial for generating realistic statistical properties such as fat-tailed return distributions, volatility clustering, and long memory in volatility. The paper also finds that large price movements are primarily caused by large gaps in the limit order book, supporting empirical findings that market impact is not solely driven by order size. The model successfully reproduces key empirical regularities of order flows, including fat-tailed distributions of order sizes and placement distances, highlighting the importance of heterogeneous trading rules in explaining market microstructure phenomena.
Highlights
- 1Develops an order-driven market model where traders use heterogeneous strategies (fundamentalist, chartist, noise) and utility-maximizing order sizes.
- 2Shows that chartist strategies are primarily responsible for fat tails and volatility clustering in returns.
- 3Demonstrates that large price changes are associated with large gaps in the limit order book.
- 4Reproduces empirical regularities of order flows, including fat-tailed distributions of order sizes and placement distances.
Methods
- MAgent-based simulation with heterogeneous trading rules and utility maximization for order submission.
- MExpected utility maximization with CARA utility to determine optimal portfolio and order sizes.
- MStatistical analysis including Hill tail index, volatility autocorrelation, and modified R/S analysis.
Results
- RChartist strategies generate fat-tailed return distributions with tail exponents below 3, consistent with real markets.
- RVolatility clustering and long memory in volatility are observed only when chartist strategies are present.
- RLarge price changes are linked to large gaps in the limit order book, not necessarily large order volumes.
- ROrder flows exhibit fat-tailed distributions for market order sizes and limit order placement distances.
Analyze Paper
Generate insights from "The Impact of Heterogeneous Trading Rules on the Limit Order...".