{"paper_id":"2412.11575","title":"Cost-aware Portfolios in a Large Universe of Assets","authors":["Qingliang Fan","Marcelo C. Medeiros","Hanming Yang","Songshan Yang"],"paper_date":"2024-12-16","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2412.11575v2","page_url":"https://thequant.space/flowcharts/cost-aware-portfolios-in-a-large-universe-of-assets/","scores":{"empirical_rigor":6,"math_complexity":8,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes cost‑aware optimal portfolio estimators for construction and rebalancing in a high‑dimensional setting using a nonconvex penalty.","quote":"We propose cost‑aware optimal portfolio estimators for new construction and subsequent rebalancing in a high‑dimensional framework.","summary":"This paper considers the finite horizon portfolio rebalancing problem in terms of mean-variance optimization, where decisions are made based on current information on asset returns and transaction costs. The study’s novelty is that the transaction costs are integrated within the optimization…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Assumption (A1) bounds the infinity norm of the inverse covariance sub‑matrix to control ℓ1‑norms of the optimal weights.","type":"statistical","stated":"explicit","quote":"Assumption (A1) guarantees three key properties. The bound on ∥Σ^{‑1}_{At, Atr}∥_{∞} ensures that the ℓ1‑norms of w^{*}_{1} and δ^{*}_{t} are controlled"},{"assumption":"Assumption (A2) requires i.i.d. sub‑Gaussian return vectors with mean µ_t and covariance Σ_t and log p < n.","type":"statistical","stated":"explicit","quote":"Assumption (A2) holds when the sample mean vector and sample covariance matrix are used as estimators, provided that the returns"},{"assumption":"Transaction costs are modeled as proportional to the ℓ1‑norm of rebalancing amounts and quadratic as a function of the square of proportional costs.","type":"execution","stated":"explicit","quote":"Proportional transaction costs scale with the sum of the absolute values of rebalancing amounts"},{"assumption":"The number of assets p can exceed the sample size n (high‑dimensional setting).","type":"model","stated":"explicit","quote":"the number of assets exceeds the sample size"},{"assumption":"The SCAD penalty parameters satisfy a1 = a2 = 1 and a > 2.","type":"statistical","stated":"explicit","quote":"The SCAD penalty has a1 = a2 = 1 and its derivative is given by"}]},"datasets":{"status":"available","items":[{"name":"S&P 500 dataset","provider":null,"kind":"market-prices","access":"public","period":null,"universe":"S&P 500 constituents","quote":"Monte Carlo simulations and empirical studies using S&P 500 and Russell 2000 stocks show the satisfactory performance of the proposed portfolio"},{"name":"Russell 2000 dataset","provider":null,"kind":"market-prices","access":"public","period":null,"universe":"Russell 2000 constituents","quote":"Monte Carlo simulations and empirical studies using S&P 500 and Russell 2000 stocks show the satisfactory performance of the proposed portfolio"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"Proportional costs use bid‑ask spread estimates; quadratic costs are set to twice the square of the proportional estimate.","quote":"For quadratic transaction cost coefficients, we define the asset‑specific cost coefficient as twice the square of the corresponding proportional cost estimate","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Proportional costs use bid‑ask spread estimates; quadratic costs are set to twice the square of the proportional estimate.","card_quote":"For quadratic transaction cost coefficients, we define the asset‑specific cost coefficient as twice the square of the corresponding proportional cost estimate","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"CAPE‑S achieves the highest overall out‑of‑sample Sharpe ratio across all stages, 0.915 under quadratic costs and 0.909 under proportional costs, while maintaining low turnover.","kind":"performance","quote":"Its leading overall Sharpe ratio under both cost specifications (0.915 quadratic, 0.909 proportional) highlights its capability to balance returns, risk, and transaction costs"},{"finding":"Theoretical results show the LLA‑based estimator converges to the oracle estimator after two iterations with high probability.","kind":"theoretical","quote":"If ∥w^{+}_{t‑1}∥_{1} < C0, then bδ_{LLA,β}^{t} converges to bδ_{oracle,β}^{t} after two iterations with probability at least 1 −c_{1}p^{‑c_{2}}"}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":17,"citations":{"cited_by_in_corpus":1,"by_independent_authors":1,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":1,"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":"2026-11-14T17:49:29Z"},"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/"}