Paper: arXiv 2408.03594
Authors: Aditya Nittur Anantha, Shashi Jain
Abstract
Market information events are generated intermittently and disseminated at high speeds in real-time. Market participants consume this high-frequency data to build limit order books, representing the current bids and offers for a given asset. The arrival processes, or the order flow of bid and offer events, are asymmetric and possibly dependent on each other. The quantum and direction of this asymmetry are often associated with the direction of the traded price movement. The Order Flow Imbalance (OFI) is an indicator commonly used to estimate this asymmetry. This paper uses Hawkes processes to estimate the OFI while accounting for the lagged dependence in the order flow between bids and offers. Secondly, we develop a method to forecast the near-term distribution of the OFI, which can then be used to compare models for forecasting OFI. Thirdly, we propose a method to compare the forecasts of OFI for an arbitrarily large number of models. We apply the approach developed to tick data from the National Stock Exchange and observe that the Hawkes process modeled with a Sum of Exponential’s kernel gives the best forecast among all competing models.
Complexity vs Empirical Score
- Math Complexity: 7.5/10
- Empirical Rigor: 8.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced mathematical modeling with Hawkes processes and detailed kernel specifications, while demonstrating high empirical rigor through application to high-frequency tick data, model comparison, and specific forecasting results on the National Stock Exchange.
Research Flowchart
flowchart TD A["Research Goal:<br>Forecast High-Frequency OFI<br>using Hawkes Processes"] --> B["Data Input:<br>Tick Data from<br>National Stock Exchange"] B --> C["Methodology 1:<br>Estimate OFI using<br>Hawkes Process with<br>Sum of Exponentials Kernel"] C --> D["Methodology 2:<br>Forecast Near-Term<br>OFI Distribution"] D --> E["Methodology 3:<br>Compare Forecasts<br>across Multiple Models"] E --> F["Key Findings:<br>Hawkes Process with<br>Sum of Exponentials Kernel<br>provides best forecast"]
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 using Hawkes processes with a Sum of Exponentials kernel to forecast the near‑term distribution of Order Flow Imbalance and a framework to compare many forecasting models.
Source sentence
observe that the Hawkes process modeled with a Sum of Exponential’s kernel gives the best forecast among all competing models.
Assumes 6
- Trades can be deterministically classified as BUY or SELL using the two exchange order numbers in the NSE tick data. data
Source sentence
This allows for a deterministic classification of every trade.
- The cross‑effect between BUY and SELL market order flows is essential and must be captured by the model. model
Source sentence
Note that accurately capturing the cross effect of the buy and the sell signed trades on each other is essential for the latter.
- Order flow arrivals can be modeled as a Hawkes point process. model
Source sentence
Bacry et al. [2015] and Fosset et al. [2022] model the market order point process using a Hawkes process.
- The loss function used for model comparison is correctly specified and can be aggregated across models. statistical
Source sentence
We develop a general framework for computing the loss function.
Show 2 more
- The tick data from a single day (September 19th 2018) is representative for evaluating the forecasting methods. data
Source sentence
We use processed tick data recorded on September 19th 2018 from the National Stock Exchange of Nifty futures expiring on September 27th 2018.
- Inter‑trade arrival times contain information about the dependency structure of the limit order book. model
Source sentence
Our results show that adequately modeling the inter-arrival time of trades and the dependency structure of the limit order book, as detailed previously, leads to a significant improvement in forecast quality.
Data
- NSE Nifty futures tick data (Sept 2018) order-book unknown
Source sentence
We use processed tick data recorded on September 19th 2018 from the National Stock Exchange of Nifty futures expiring on September 27th 2018 (current month futures contract).
Costs
Not applicable
Finds 5
- The Hawkes process with a Sum of Exponential kernel provides the best OFI forecast among all competing models. predictive-accuracy
Source sentence
observe that the Hawkes process modeled with a Sum of Exponential’s kernel gives the best forecast among all competing models.
- Incorporating inter‑trade arrival times and the dependency structure of the limit order book significantly improves forecast quality. predictive-accuracy
Source sentence
We observe a significant enhancement in forecast quality when considering the dependency structure of the limit order book through inter-trade arrival times.
- A method to forecast the near‑term distribution of OFI is presented. theoretical
Source sentence
In this paper, we present a method to forecast a near-term distribution of the OFI.
- A general framework for computing the loss function enables comparison of multiple forecasting models. theoretical
Source sentence
We develop a general framework for computing the loss function.
- Modeling the time distribution of market order events and the temporal dependence between BUY and SELL flows enhances confidence in selecting a benchmark model in high‑frequency settings. statistical
Source sentence
By modeling the time distribution of market order events and the temporal dependence between ‘BUY’ and ‘SELL’ market order flows, this paper demonstrates that incorporating the inter-arrival time of trades significantly enhances the confidence with which a benchmark model can be chosen in high-frequency settings.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 20 related papers). Coverage is partial, so this is neither replication nor refutation.
2 papers in the archive : Hybrid Vector Auto Regression and Neural Network Model for … ; TABL-ABM: A Hybrid Framework for Synthetic LOB Generation
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.
Cited References (34)
- GRASP with path-relinking for the three-index assignment problem (2000)
- Probability distribution of solution time in GRASP: an experimental investigation (2003)
- Parallel GRASP with path-relinking for job shop scheduling (1984)
- Title of a Book in Another Language (Quoted in that Language) (1988)
- Multi-start and strategic oscillation methods – principles to exploit adaptive memory (2000)
- Bubbles in Japan’s stock markets: a macroeconomic analysis (1995)
- The resurgent Japanese economy and a Japan–United States free trade agreement (1997)
- Efficient algorithms for layer assignment problems (1996)
and 26 more
- Gravitation (1973)
- Parallel GRASP with path-relinking for job shop scheduling (1983)
- Title of Book (2002)
- A study on Algorithmic Trading / High Frequency Trading In The Indian Capital Market (2017)
- Does Algorithmic Trading Improve Liquidity? (2011)
- The cost of algorithmic trading: A first look at comparative performance (2005)
- The price impact of order book events (2014)
- Flow toxicity and liquidity in a high-frequency world (2012)
- Point spectra of some mutually exciting point processes (1971)
- Price formation and liquidity in the U.S. Treasury market: The response to public information (1999)
- Order imbalances and stock price movements on October 19 and 20, 1987 (1993)
- Stock market crashes and the performance of circuit breakers: Empirical evidence (1993)
- Order imbalance, liquidity, and market returns (2002)
- Futures volatility forecasting based on big data analytics with incorporating an order imbalance effect (2022)
- High-frequency quoting: Short-term volatility in bids and offers (2018)
- Hawkes processes in finance (2015)
- Non-parametric estimation of quadratic Hawkes processes for order book events (2022)
- A reality check for data snooping (2000)
- A test for superior predictive ability (2005)
- Second order statistics characterization of Hawkes processes and non-parametric estimation (2014) — link
- Inferring trade direction from intraday data (1991)
- On Lewis’ simulation method for point processes (1981)
- A non-parametric EM algorithm for multiscale Hawkes processes (2011)
- The diversity of high-frequency traders (2013)
- Liquidity, Information, and Infrequently Traded Stocks (1996)
- The Economic Role of the CBOE’s Short-Term Volatility Index Futures (1992)
0 of 34 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.