{"paper_id":"2405.10654","title":"Microstructure Modes -- Disentangling the Joint Dynamics of Prices & Order Flow","authors":["Salma Elomari-Kessab","Guillaume Maitrier","Julius Bonart","Jean-Philippe Bouchaud"],"paper_date":"2024-05-17","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2405.10654v1","page_url":"https://thequant.space/flowcharts/microstructure-modes-disentangling-the-joint-dynamics/","scores":{"empirical_rigor":8,"math_complexity":6,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a double coarse‑graining procedure to extract minute‑scale information and defines a calibrated Vector Auto‑Regressive (VAR) model for the microstructure modes.","quote":"we propose a double coarse-graining procedure that allows us to extract meaningful information at the minute time scale. We","summary":"Understanding the micro-dynamics of asset prices in modern electronic order books is crucial for investors and regulators. In this paper, we use an order by order Eurostoxx database spanning over 3 years to analyze the joint dynamics of prices and order flow. In order to alleviate various problems…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"EuroStoxx is a large‑tick, liquid asset with a spread almost always equal to one tick.","type":"market","stated":"explicit","quote":"It is noteworthy that EuroStoxx is a liquid large-tick asset, with a spread almost always equal to one tick."},{"assumption":"After binning, null flows disappear, making the VAR model applicable.","type":"data","stated":"explicit","quote":"... an effect that completely disappears when the data is binned."},{"assumption":"The linear VAR framework is suitable for modeling the dynamics of the transformed variables.","type":"model","stated":"explicit","quote":"We adopt a Vector Autoregression (VAR) model to capture the dynamic relationships among the variables associated with each price change."},{"assumption":"For short time scales, market orders issued by slow traders have little short‑term predictability, allowing the second term in the impact definition to be neglected.","type":"execution","stated":"explicit","quote":"for short enough time scales, market orders issued by slow traders have little short term predictability such that the second term in Eq. (11) is negligible."},{"assumption":"Box‑Cox transformations render the binned variables approximately Gaussian.","type":"statistical","stated":"explicit","quote":"These parameters are chosen to maximise the likelihood of the Gaussian distribution of the transformed variables, which yields λv = 0.20 and λt = 0.14."},{"assumption":"The impact curves from prior literature can be used to compute instantaneous returns within the VAR framework.","type":"model","stated":"implied","quote":"due to the use of the impact curves from [22] for computing the instantaneous return, the return produced by our model is diluted in scale."}]},"datasets":{"status":"available","items":[{"name":"EuroStoxx futures order‑by‑order database","provider":null,"kind":"order-book","access":"unknown","period":"September 2016 to August 2019 (545 days)","universe":"EuroStoxx futures contract","quote":"The dataset used in this study consists of 545 days of the futures contract on EuroStoxx from September 2016 to August 2019."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R1","candidates":{"front_matter":"n/a","evidence_card":"unclear","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"The VAR model parameters are extremely stable over time and achieve relatively high R² prediction scores, especially for symmetric liquidity modes (≈0.28–0.32).","kind":"predictive-accuracy","quote":"R2 scores are much higher (∼0.28 −0.32) for symmetric modes, which carry no information on returns, than for anti-symmetric modes (∼0.01 −0.03)."},{"finding":"In‑sample and out‑of‑sample R² scores are close, indicating statistically significant predictive value of the VAR model.","kind":"statistical","quote":"the in-sample R2 scores and out-of-sample scores are close, highlighting the fact that the predictive value of the VAR model is statistically significant."},{"finding":"Adding more lags (up to 10) increases out‑of‑sample R² for all modes by about 25% but makes the system marginally less stable.","kind":"predictive-accuracy","quote":"adding more lags reduces auto‑correlation of residuals and increases the out of sample R2 score of all the modes, by ∼25% both for the symmetric and anti‑symmetric ones when p increases from 1 to 10."},{"finding":"The VAR framework does not reproduce the well‑known square‑root law of price impact.","kind":"negative","quote":"our VAR framework does not account for the well known square‑root law of price impact."},{"finding":"Flow variables improve return prediction: VAR yields a 1.6% R² for returns, higher than the 0.49% obtained using past returns alone.","kind":"predictive-accuracy","quote":"... significantly higher than the score of 0.49% obtained when predicting returns using the past return as the only feature."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[],"contradicts":[{"paper":{"paper_id":"2311.18283","title":"The two square root laws of market impact and the role of sophisticated market participants","paper_date":"2023-11-30","authors":["Bruno Durin","Mathieu Rosenbaum","Grégoire Szymanski"],"empirical_rigor":4,"page_url":"https://thequant.space/flowcharts/the-two-square-root-laws-of-market-impact-and-the-role-of-so/"},"question":"Does price impact follow a square‑root law with respect to traded volume/participation rate?","this_paper_finding":"The VAR framework does not reproduce the well‑known square‑root law of price impact.","other_paper_finding":"During execution, market impact follows a square‑root law in the cumulated traded volume for a given participation rate.","other_paper_quote":"Our results lead to the recovery of two square root laws: (i) For a given participation rate, during the execution of a metaorder, the market impact evolves in a square root manner with respect to the cumulated traded volume. More precisely, we show that during the execution given metaorder, the market impact evolves following a square root pattern in relation to the cumu-lative traded volume, exhibiting then a","note":"A confirms the square‑root law; B’s VAR framework fails to reproduce it, contradicting A.","confidence":0.412}],"qualifies":[],"compared_with":42,"citations":{"cited_by_in_corpus":5,"by_independent_authors":1,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":1,"mentions":0,"unclassified":4},"independent_papers":[{"paper":{"paper_id":"2501.17096","title":"Why is the estimation of metaorder impact with public market data so challenging?","paper_date":"2025-01-28","authors":["Manuel Naviglio","Giacomo Bormetti","Francesco Campigli","German Rodikov","Fabrizio Lillo"],"empirical_rigor":4,"page_url":"https://thequant.space/flowcharts/why-is-the-estimation-of-metaorder-impact-with-public-market/"},"how":"discusses_findings","quote":"Specifically, in the first part of this paper we confirm these observations by considering the linear Transient Impact Model"}],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:09:26Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}