{"paper_id":"2308.01419","title":"Graph Neural Networks for Forecasting Multivariate Realized Volatility with Spillover Effects","authors":["Chao Zhang","Xingyue Pu","Mihai Cucuringu","Xiaowen Dong"],"paper_date":"2023-08-01","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2308.01419v1","page_url":"https://thequant.space/flowcharts/graph-neural-networks-for-forecasting-multivariate-realized/","scores":{"empirical_rigor":8,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel GNN-based methodology (GNNHAR) for modeling and forecasting multivariate realized volatility with spillover effects.","quote":"In this article, we propose a novel methodology GNNHAR for modeling and forecasting RV, while taking into account volatility spillover effects in the U.S. equity market.","summary":"We present a novel methodology for modeling and forecasting multivariate realized volatilities using customized graph neural networks to incorporate spillover effects across stocks. The proposed model offers the benefits of incorporating spillover effects from multi-hop neighbors, capturing…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The input vector of N nodes is drawn from a multivariate Gaussian distribution.","type":"statistical","stated":"explicit","quote":"It assumes the input vector of N nodes is drawn from a multivariate Gaussian distribution N(0, Σ)."},{"assumption":"Stocks among the DJIA components traded continuously throughout the entire sample period.","type":"data","stated":"explicit","quote":"we include only those stocks among the DJIA components that traded continuously throughout the entire period."},{"assumption":"Models are re‑calibrated monthly using a rolling window of the past 1000 days.","type":"model","stated":"explicit","quote":"All models are re-calibrated every month based on a rolling sample window of the past 1000 days"},{"assumption":"Training uses 36‑month data and validation uses the most recent 12‑month data.","type":"model","stated":"explicit","quote":"Specifically, we use 36-month data for model training, and the recent 12-month data as the validation set to tune the hyperparameters"},{"assumption":"Adjacency matrix is built via Graphical LASSO to capture conditional independence of assets.","type":"model","stated":"explicit","quote":"Based on these compelling results, we adopt Graphical LASSO to establish the volatility graph in our study."},{"assumption":"The normalized adjacency matrix W = O^{-1/2} A O^{-1/2} is used to avoid numerical instabilities.","type":"model","stated":"explicit","quote":"where W = O^{-1/2} A O^{-1/2} is the normalized adjacency matrix, used to avoid numerical instabilities and exploding/vanishing gradients during the training phrase."}]},"datasets":{"status":"available","items":[{"name":"LOBSTER intraday data","provider":"LOBSTER","kind":"order-book","access":"commercial","period":"July 1, 2007 to Jun 30, 2021","universe":"DJIA components (27 stocks)","quote":"The intraday data of Dow Jones Industrial Average (DJIA) components are obtained from the LOBSTER database.","known_id":"lobster"},{"name":"S&P100 components","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":"S&P100 components","quote":"Additionally, for robustness checks, we consider a larger universe of S&P100 components."}],"known":[{"id":"lobster","name":"LOBSTER","provider":"LOBSTER","kind":"order-book","access":"commercial","quote":"f Dow Jones Industrial Average (DJIA) components are obtained from the LOBSTER database.11 The time period under consideration is from July 1, 2007 to Jun 30, 2021.12 Following Bollerslev et al. [2016], we include only those stocks among the DJIA components that traded continuously throughout the entire period.","mentions":2,"used_in":2}],"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":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"Incorporating spillover effects from multi‑hop neighbors alone does not provide a clear advantage in predictive accuracy.","kind":"negative","quote":"Our empirical findings provide compelling evidence that incorporating spillover effects from multi‑hop neighbors alone does not yield a clear advantage in terms of predictive accuracy."},{"finding":"Modeling nonlinear spillover effects improves forecasting accuracy of realized volatilities, especially for short‑term horizons up to one week.","kind":"predictive-accuracy","quote":"However, modeling nonlinear spillover effects enhances the forecasting accuracy of realized volatilities, particularly for short‑term horizons of up to one week."},{"finding":"Training with the QLIKE loss leads to substantial improvements in model performance compared to the commonly‑used MSE.","kind":"performance","quote":"our results consistently indicate that training with the Quasi‑likelihood loss leads to substantial improvements in model performance compared to the commonly‑used mean squared error."},{"finding":"The nonlinear GNNHAR1L model outperforms the linear baseline and avoids over‑fitting.","kind":"performance","quote":"GNNHAR1L shows significant promise as a model for capturing nonlinearity, while avoiding the overfitting problem."},{"finding":"Increasing the number of GNN layers generally degrades forecasting performance, though QLIKE‑trained models remain beneficial.","kind":"performance","quote":"Generally, as we increase the number of layers in the GNNHAR models, their forecasting performance tends to decline. Nevertheless, we still observe the benefits of training models with the QLIKE loss function."}],"reported_performance":null,"out_of_sample":{"method":"walk-forward","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":44,"citations":{"cited_by_in_corpus":2,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":2},"independent_papers":[],"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":"2026-11-14T18:23:56Z"},"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/"}