{"paper_id":"2403.02500","title":"RVRAE: A Dynamic Factor Model Based on Variational Recurrent Autoencoder for Stock Returns Prediction","authors":["Yilun Wang","Shengjie Guo"],"paper_date":"2024-03-04","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2403.02500v1","page_url":"https://thequant.space/flowcharts/rvrae-a-dynamic-factor-model-based-on-variational-recurrent/","scores":{"empirical_rigor":8,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes RVRAE, a dynamic latent factor model that combines recurrent neural networks with a variational autoencoder to extract factors from noisy market data and predict cross‑sectional stock returns.","quote":"We introduce RVRAE as an innovative dynamic latent factor model tailored to extract key factors from noisy market data efficiently.","summary":"In recent years, the dynamic factor model has emerged as a dominant tool in economics and finance, particularly for investment strategies. This model offers improved handling of complex, nonlinear, and noisy market conditions compared to traditional static factor models. The advancement of machine…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The prior distribution of the latent variable z_t is a standard Gaussian.","type":"model","stated":"explicit","quote":"In vanilla VAE, pθ (zt) is assumed as a standard Gaussian distribution."},{"assumption":"Factor exposures β_t can vary over time (dynamic betas).","type":"model","stated":"explicit","quote":"However, challenges arise with dynamic betas, indicating that factor exposures may change over time as firms evolve."},{"assumption":"Monthly individual stock returns from CRSP are representative of the US equity market.","type":"data","stated":"implied","quote":"We collect monthly individual stock returns from CRSP The returns dataset starts from January 2000 to December 2020."},{"assumption":"The 46 firm characteristics capture relevant information for factor extraction.","type":"data","stated":"implied","quote":"According to [20] and [21], we choose 46 firm characteristics."},{"assumption":"Transaction costs are a flat 30 basis points per trade.","type":"execution","stated":"explicit","quote":"Following [9], we consider two situations: Sharpe Ratio without transaction costs and Sharpe Ratio with 30 bps transaction costs."}]},"datasets":{"status":"available","items":[{"name":"CRSP monthly individual stock returns","provider":"CRSP","kind":"market-prices","access":"commercial","period":"2000-01 to 2020-12","universe":"5682 US stocks","quote":"We collect monthly individual stock returns from CRSP The returns dataset starts from January 2000 to December 2020. The number of stocks is 5682.","known_id":"crsp"},{"name":"Firm characteristics","provider":null,"kind":"fundamentals","access":"unknown","period":null,"universe":null,"quote":"According to [20] and [21], we choose 46 firm characteristics."}],"known":[{"id":"crsp","name":"CRSP","provider":"Center for Research in Security Prices","kind":"market-prices","access":"commercial","quote":"We collect monthly individual stock returns from CRSP The returns dataset starts from January 2000 to December 2020.","mentions":1,"used_in":1},{"id":"french-library","name":"Kenneth French Data Library","provider":"Kenneth R. French","kind":"market-prices","access":"public","quote":"Sharpe Ratio = rt −rf standard deviation (rt) (34) Where rf is the risk-free rate and we use monthly Treasury bill rates from the Kenneth French Data Library as the risk-free rate.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"weekly-monthly"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"30 basis points per trade are subtracted when reporting Sharpe ratios.","quote":"Following [9], we consider two situations: Sharpe Ratio without transaction costs and Sharpe Ratio with 30 bps transaction costs.","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"30 basis points per trade are subtracted when reporting Sharpe ratios.","card_quote":"Following [9], we consider two situations: Sharpe Ratio without transaction costs and Sharpe Ratio with 30 bps transaction costs.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"All models’ Sharpe ratios decline when transaction costs are added, but RVRAE still outperforms the others.","kind":"performance","quote":"It is interesting to notice that all models perform worse when adding transaction costs, but models with Attention mechanism (ALSTM, Trans) have the smallest difference..."},{"finding":"RVRAE consistently outperforms baselines across robustness tests with omitted stocks (higher Rank IC and Rank ICIR).","kind":"statistical","quote":"It reveals that our RVRAE model outperforms other baseline methods for all values of ’m’."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":17,"citations":{"cited_by_in_corpus":2,"by_independent_authors":2,"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":"2026-11-14T18:13:41Z"},"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/"}