{"paper_id":"2510.26165","title":"Learning to Manage Investment Portfolios beyond Simple Utility Functions","authors":["Maarten P. Scholl","Mahmoud Mahfouz","Anisoara Calinescu","J. Doyne Farmer"],"paper_date":"2025-10-30","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2510.26165v1","page_url":"https://thequant.space/flowcharts/learning-to-manage-investment-portfolios-beyond-simple-utili/","scores":{"empirical_rigor":7,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a GAN-based generative framework that learns latent representations of fund manager strategies without explicit utility specification.","quote":"We propose a generative framework that learns latent representations of fund manager strategies without requiring explicit utility specification.","summary":"While investment funds publicly disclose their objectives in broad terms, their managers optimize for complex combinations of competing goals that go beyond simple risk-return trade-offs. Traditional approaches attempt to model this through multi-objective utility functions, but face fundamental…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Fund manager strategy can be modeled as a conditional probability distribution over portfolio weights.","type":"model","stated":"explicit","quote":"We model a manager’s strategy as a conditional probability distribution over portfolio weights, without specifying a utility function."},{"assumption":"Temporal split of data prevents look-ahead bias.","type":"data","stated":"explicit","quote":"training on 2010-2018, validation on 2019, and testing on 2020-2024, ensuring no look-ahead bias."},{"assumption":"Funds must have at least 12 months of holdings data, report at least 75% of holdings by weight, and allocate at least 75% within the largest 500 U.S. stocks.","type":"data","stated":"explicit","quote":"funds must have at least 12 months of holdings data, report at least 75% of holdings by weight, and allocate at least 75% within our stock universe (the largest 500 U.S. stocks by market cap)."},{"assumption":"An 8‑dimensional latent space is sufficient to capture the essential variation in investment strategies.","type":"model","stated":"implied","quote":"The learned 8-dimensional latent space meaningfully organizes investment strategies."},{"assumption":"Adversarial training is necessary for realistic portfolio generation.","type":"model","stated":"implied","quote":"adversarial training proves essential—the discriminator forces the generator beyond simple moment matching to capture the full distributional properties of real portfolios."},{"assumption":"Factor exposures and other asset characteristics are adequately represented by the selected market, fundamental, and technical features.","type":"model","stated":"implied","quote":"The model provides good explanatory power while remaining parsimonious. It has features from three important categories: market, fundamental, and technical characteristics."}]},"datasets":{"status":"available","items":[{"name":"CRSP Survivor-Bias-Free US Mutual Fund Database","provider":"CRSP","kind":"fundamentals","access":"commercial","period":"2010-2024","universe":"largest 500 U.S. stocks by market cap","quote":"Our sample covers actively managed U.S. equity funds from 2010 to 2024","known_id":"crsp"},{"name":"Lipper classifications","provider":"Lipper","kind":"benchmark","access":"commercial","period":null,"universe":null,"quote":"We use Lipper classifications (e.g., \"Large-Cap Growth\", \"Small-Cap Value\") as ground-truth labels."}],"known":[{"id":"crsp","name":"CRSP","provider":"Center for Research in Security Prices","kind":"market-prices","access":"commercial","quote":"4.1 Data We use the CRSP Survivor-Bias-Free US Mutual Fund Database, a standard academic finance dataset.","mentions":2,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"weekly-monthly"},"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":"The full model achieves the lowest errors on key hold-out metrics: count error of 15 stocks and concentration matching error of 0.0047.","kind":"statistical","quote":"the full model achieves the lowest errors on key hold-out metrics: count error (15 stocks), concentration matching (0.0047)."},{"finding":"The Zero-Trade baseline performs surprisingly well on concentration and replication metrics, reflecting the slow turnover in fund holdings.","kind":"performance","quote":"The Zero-Trade baseline performs surprisingly well on concentration, replication metrics, reflecting the slow turnover in fund holdings."},{"finding":"Linear SVM trained on learned strategy encodings recovers Lipper categories with a macro‑averaged score of 77%.","kind":"predictive-accuracy","quote":"A macro-averaged score of 77% shows the Lipper scheme is contained in the latent space in a simple linear form."},{"finding":"Over four‑fifths of active funds exhibit Markowitz‑like efficiency seeking.","kind":"statistical","quote":"over four-fifths of active funds exhibit Markowitz-like efficiency seeking."},{"finding":"Generator‑Only ablation degrades performance, with count error increasing to 84 stocks, showing adversarial feedback is essential for realism.","kind":"performance","quote":"The Generator-Only ablation test shows degraded performance across many metrics except the performance on the synthetic data, with count error increasing to 84 stocks."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":31,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"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":"2026-11-14T17:23:21Z"},"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/"}