Paper: arXiv 2505.17388
Authors: Chen Hu, Kouxiao Zhang
Abstract
We conduct modeling of the price dynamics following order flow imbalance in market microstructure and apply the model to the analysis of Chinese CSI 300 Index Futures. There are three findings. The first is that the order flow imbalance is analogous to a shock to the market. Unlike the common practice of using Hawkes processes, we model the impact of order flow imbalance as an Ornstein-Uhlenbeck process with memory and mean-reverting characteristics driven by a jump-type Lévy process. Motivated by the empirically stable correlation between order flow imbalance and contemporaneous price changes, we propose a modified asset price model where the drift term of canonical geometric Brownian motion is replaced by an Ornstein-Uhlenbeck process. We establish stochastic differential equations and derive the logarithmic return process along with its mean and variance processes under initial boundary conditions, and evolution of cost-effectiveness ratio with order flow imbalance as the trading trigger point, termed as the quasi-Sharpe ratio or response ratio. Secondly, our results demonstrate horizon-dependent heterogeneity in how conventional metrics interact with order flow imbalance. This underscores the critical role of forecast horizon selection for strategies. Thirdly, we identify regime-dependent dynamics in the memory and forecasting power of order flow imbalance. This taxonomy provides both a screening protocol for existing indicators and an ex-ante evaluation paradigm for novel metrics.
Complexity vs Empirical Score
- Math Complexity: 7.5/10
- Empirical Rigor: 6.0/10 (capped from 6.5: transaction costs not modelled; see penalty flags)
- Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 6.5)
Why this score: The paper employs advanced stochastic calculus (SDEs, OU processes driven by Lévy jumps) with extensive analytical derivations, indicating high mathematical complexity. It is substantiated with empirical analysis on high-frequency tick data, regime detection, and practical metric evaluation, demonstrating solid empirical rigor.
Research Flowchart
flowchart TD A["Research Goal:<br>Model Price Dynamics from Order Flow Imbalance"] --> B["Data & Inputs:<br>CSI 300 Index Futures Order Flow Data"] B --> C["Methodology:<br>Order Flow Imbalance as Lévy Process"] C --> D["Computational Model:<br>OU Process replaces Drift in Price Model"] D --> E["Stochastic Differential Equations<br>Derive Log Returns & Quasi-Sharpe Ratio"] E --> F["Key Findings:<br>1. Horizon-Dependent Heterogeneity"] E --> G["Key Findings:<br>2. Regime-Dependent Dynamics & Forecasting Power"] E --> H["Key Findings:<br>3. Novel Protocol for Indicator Evaluation"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a modified asset price model that replaces the drift term of geometric Brownian motion with an Ornstein‑Uhlenbeck process driven by a jump‑type Lévy process.
Source sentence
we propose a modified asset price model where the drift term of canonical geometric Brownian motion is replaced by an Ornstein-Uhlenbeck process.
Assumes 5
- µs is a Lévy‑driven Ornstein‑Uhlenbeck process with mean 0, symmetric distribution, finite variance, and increments σ²L. model
Source sentence
We assume µs is a L´evy-driven O-U process with mean 0, symmetric distribution, finite variance, and increments σ2 L.
- The drift’s Lévy process Lt and the GBM’s Wiener process Wt are independent. statistical
Source sentence
The drift’s L´evy process Lt and GBM’s Wiener process Wt are independent.
- The OFI metric quantifies persistent market impacts and exhibits stationary fluctuations around a long‑run equilibrium. market
Source sentence
We assume the OFI metric quantifies persistent market impacts, exhibiting stationary fluctuations around a long-run equilibrium.
- Tick data are snapshot updates every 500 ms, each representing a trading event. data
Source sentence
snapshot updates every 500 milliseconds. A market tick is generated and transmitted whenever a trade occurs or there are changes in price or quantity of the limit order book within this 500-millisecond window.
Show 1 more
- When focusing on the expected value of the aggregate price impact, stochastic terms σ√Δtϵt and ηΔJt have zero expectation and can be disregarded. model
Source sentence
When focusing on the expected value of Sn, we may disregard the stochastic terms σ√Δtϵt and ηΔJt, as established previously, these represent symmetric Wiener and L´evy processes with zero expectation.
Data
- CSI 300 index futures tick data order-book unknown
Source sentence
All the following analysis is based on one year CSI 300 futures tick data, with total of 6 million ticks
Costs
Not modeled
Finds 5
- Order flow imbalance acts like a market shock and can be modeled by a mean‑reverting Ornstein‑Uhlenbeck process. theoretical
Source sentence
the order flow imbalance is analogous to a shock to the market.
- Results show horizon‑dependent heterogeneity in how conventional metrics interact with order flow imbalance. statistical
Source sentence
our results demonstrate horizon-dependent heterogeneity in how conventional metrics interact with order flow imbalance.
- Regime‑dependent dynamics are identified in the memory and forecasting power of order flow imbalance. theoretical
Source sentence
we identify regime-dependent dynamics in the memory and forecasting power of order flow imbalance.
- Aggregate profits rise from about 64,000 points at a 0.5‑second horizon to roughly 324,537 points at 30 seconds and 352,975 points at a 2‑minute horizon. performance
Source sentence
aggregate profits increase significantly from 64,000 points at the 0.5-second horizon to approximately 324,537 points at 30 seconds and 352,975 points at the 2-minute horizon.
- Regression coefficients grow from 0.027 for 0.5‑second predictions to 0.066 at 30 seconds and stabilize at 0.067 for 2‑minute forecasts, indicating improved predictive efficacy over longer horizons. predictive-accuracy
Source sentence
regression coefficients exhibit a parallel enhancement, rising from 0.027 for the 0.5-second prediction to 0.066 at 30 seconds and stabilizing at 0.067 for the 2-minute forecast
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 32 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.