{"paper_id":"2410.23536","title":"On Cost-Sensitive Distributionally Robust Log-Optimal Portfolio","authors":["Chung-Han Hsieh","Xiao-Rou Yu"],"paper_date":"2024-10-31","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2410.23536v1","page_url":"https://thequant.space/flowcharts/on-cost-sensitive-distributionally-robust-log-optimal-portfo/","scores":{"empirical_rigor":4,"math_complexity":8.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a novel cost-sensitive distributionally robust log‑optimal portfolio framework that incorporates ambiguous return distributions and a general convex transaction cost model.","quote":"cost-sensitive distributionally robust log-optimal portfolio problem, where the investor faces","summary":"This paper addresses a novel \\emph{“cost-sensitive”} distributionally robust log-optimal portfolio problem, where the investor faces \\emph{“ambiguous”} return distributions, and a general convex transaction cost model is incorporated. The uncertainty in the return distribution is quantified using…","strategy":"Cost-sensitive distributionally robust log-optimal portfolio"},"assumptions":{"status":"available","items":[{"assumption":"The transaction cost model is convex and follows a proportional cost structure.","type":"model","stated":"explicit","quote":"we assume that the convex cost model follows a proportional cost structure; i.e., T Ci(ui(0)) := ciwiV (0), where ci = c ∈[0, 1) for all i = 1, . . . , m."},{"assumption":"Return distributions are ambiguous and captured by a Wasserstein ball of radius ε ≥ 0.","type":"statistical","stated":"explicit","quote":"Let ε ≥0. Using the Wasserstein metric in Definition 2.2, we construct the ambiguity set consisting of"},{"assumption":"Asset returns are represented by iid random samples drawn from the empirical distribution.","type":"data","stated":"explicit","quote":"Consistent with Calaﬁore (2013), Mohajerin Esfahani and Kuhn (2018) and Li (2023), suppose that bX1, . . . , bXN represent N random samples of ﬁnite random vector Xn ∼F."},{"assumption":"Rebalancing occurs monthly with no trades between periods.","type":"execution","stated":"explicit","quote":"With the monthly rebalancing approach, account values only change at the beginning of each month, as the investor makes no trades between rebalancing periods."},{"assumption":"The risk‑free rate is taken from the U.S. Treasury one‑month constant maturity yield.","type":"market","stated":"explicit","quote":"For the risk‑free rate, we use the Market Yield on U.S. Treasury Securities at One‑Month Constant Maturity, Quoted on an Investment Basis."}]},"datasets":{"status":"available","items":[{"name":"S&P 500 top‑10 stocks (2022 year‑end)","provider":"Yahoo Finance","kind":"market-prices","access":"public","period":"2022‑01‑01 to 2024‑01‑01","universe":"Top 10 S&P 500 constituents by market cap","quote":"as shown in Table 1. The corresponding stock prices, obtained from Yahoo Finance for the two‑year period from January 1, 2022, to January 1, 2024, are shown in Figure 1.","known_id":"yahoo-finance"},{"name":"U.S. Treasury one‑month constant maturity yield","provider":"Federal Reserve Board","kind":"macro","access":"public","period":null,"universe":null,"quote":"For the risk‑free rate, we use the Market Yield on U.S. Treasury Securities at One‑Month Constant Maturity, Quoted on an Investment Basis."}],"known":[],"sample":{"start":"2022-01","end":"2024-01","market":"S&P 500 stocks","asset_class":"equities","frequency":"monthly"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":true,"detail":"Proportional cost structure with ci = c ∈[0,1) for all assets; examples include 0.1%, 0.5% and 1% per trade.","quote":"we assume that the convex cost model follows a proportional cost structure; i.e., T Ci(ui(0)) := ciwiV (0), where ci = c ∈[0, 1) for all i = 1, . . . , m.","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Proportional cost structure with ci = c ∈[0,1) for all assets; examples include 0.1%, 0.5% and 1% per trade.","card_quote":"we assume that the convex cost model follows a proportional cost structure; i.e., T Ci(ui(0)) := ciwiV (0), where ci = c ∈[0, 1) for all i = 1, . . . , m.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Without transaction costs the optimal portfolio converges to an equal‑weighted allocation.","kind":"theoretical","quote":"without transaction costs, the optimal portfolio converges to an equal‑weighted allocation"},{"finding":"With transaction costs the optimal portfolio shifts slightly towards the risk‑free asset.","kind":"theoretical","quote":"with transaction costs, the portfolio shifts slightly towards the risk‑free asset, reflecting the trade‑off between cost considerations and optimal allocation"},{"finding":"Larger Wasserstein radius ε yields lower standard deviation, lower maximum drawdown, higher cumulative return and Sharpe ratio compared to ε=0.","kind":"performance","quote":"when ε = 1, the portfolio demonstrates lower standard deviation (STD) and maximum drawdown (MDD)"},{"finding":"Across all ε>0 the robust strategy outperforms both the SPY benchmark and the ε=0 case.","kind":"performance","quote":"whenever ε̸ = 0, i.e., when ambiguity is considered, the portfolio consistently outperforms both SPY and the ε = 0 case"},{"finding":"Empirical results show that the robust strategy mitigates downside risk during market regime switches (2022 bearish to 2023 bullish).","kind":"performance","quote":"our proposed cost‑sensitive DRO approach provides a robust solution to handle such market regime switches"}],"reported_performance":{"strategy":"Cost-sensitive distributionally robust log-optimal portfolio","sharpe":0.374,"sharpe_benchmark":-0.006,"annual_return_pct":null,"total_return_pct":116,"max_drawdown_pct":12.79},"out_of_sample":{"method":"walk-forward","headline_out_of_sample":true}},"independent_evidence":{"status":"available","supports":[],"contradicts":[{"paper":{"paper_id":"2307.02178","title":"Non-Concave Utility Maximization with Transaction Costs","paper_date":"2023-07-05","authors":["Shuaijie Qian","Chen Yang"],"empirical_rigor":3,"page_url":"https://thequant.space/flowcharts/non-concave-utility-maximization-with-transaction-costs/"},"question":"How do transaction costs affect the optimal allocation between the risky asset and the risk‑free asset?","this_paper_finding":"With transaction costs the optimal portfolio shifts slightly towards the risk‑free asset.","other_paper_finding":"Introducing transaction costs can make it optimal for investors to hold a larger long position in the risky asset compared to the frictionless case, or a large short position despite a positive risk premium.","other_paper_quote":"Via extensive numerical analyses involving several types of utility functions, we find that the introduction of transaction costs into non-concave utility maximization problems can make it optimal for investors to hold on to a larger long position in the risky asset compared to the frictionless case, or hold on to a large short position in the risky asset despite a positive risk premium. For ex-ample, the investor","note":"Finding A finds transaction costs can lead to a larger long (or short) position in the risky asset, while Finding B finds the optimal portfolio shifts slightly toward the risk‑free asset, showing opposite effects.","confidence":0.651}],"qualifies":[],"compared_with":37,"citations":{"cited_by_in_corpus":1,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":1},"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":"2026-11-14T17:56:44Z"},"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/"}