{"paper_id":"2409.09684","title":"Return Prediction for Mean-Variance Portfolio Selection: How Decision-Focused Learning Shapes Forecasting Models","authors":["Junhyeong Lee","Haeun Jeon","Hyunglip Bae","Yongjae Lee"],"paper_date":"2024-09-15","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2409.09684v4","page_url":"https://thequant.space/flowcharts/return-prediction-for-mean-variance-portfolio-selection-how/","scores":{"empirical_rigor":4,"math_complexity":8.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a Decision-Focused Learning (DFL) approach that adjusts stock return prediction models to improve mean-variance portfolio decisions.","quote":"This study investigates how DFL adjusts stock return prediction models to optimize decisions in MVO.","summary":"Markowitz laid the foundation of portfolio theory through the mean-variance optimization (MVO) framework. However, the effectiveness of MVO is contingent on the precise estimation of expected returns, variances, and covariances of asset returns, which are typically uncertain. Machine learning…","strategy":"Decision-Focused Learning (DFL) for Mean-Variance Optimization"},"assumptions":{"status":"available","items":[{"assumption":"The analysis focuses on the unconstrained mean-variance optimization problem.","type":"model","stated":"explicit","quote":"Our theoretical analysis focused on the unconstrained MVO problem, which allowed us to derive closed-form solutions and clear insights."},{"assumption":"Sample covariance matrices computed from historical daily returns are used as the true covariance for MVO.","type":"data","stated":"explicit","quote":"For covariances, we use the sample covariance matrix ˆΣ calculated using historical data."},{"assumption":"Expected returns are predicted from a 60‑day look‑back window of daily returns.","type":"data","stated":"explicit","quote":"Prediction models were trained to predict the returns of"},{"assumption":"Daily rolling forecasts are used for both expected returns and covariances.","type":"execution","stated":"explicit","quote":"Both the prediction of expected returns and the covariances are calculated on a daily rolling basis."},{"assumption":"No transaction costs or market impact are incorporated in the reported portfolio performance.","type":"execution","stated":"implied","quote":"Our results demonstrate that intermediate α values consistently outperform both extremes."}]},"datasets":{"status":"available","items":[{"name":"DOW 30","provider":null,"kind":"market-prices","access":"public","period":"2010-01-01 to 2024-12-31","universe":"DOW30 constituents","quote":"For the dataset, we used the DOW30 and S&P100 asset universes, covering the period from 01-Jan-2010 to 31-Dec-2024."},{"name":"S&P 100","provider":null,"kind":"market-prices","access":"public","period":"2010-01-01 to 2024-12-31","universe":"S&P100 constituents","quote":"For the dataset, we used the DOW30 and S&P100 asset universes, covering the period from 01-Jan-2010 to 31-Dec-2024."}],"known":[],"sample":{"start":"2010-01","end":"2024-12","market":"US equities (DOW 30)","asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":false,"detail":null,"quote":null,"source":"front_matter","rule":"R1","candidates":{"front_matter":"no","evidence_card":"no","evidence_card_quoted":false},"meaning":"a trading or execution result is reported without costs"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Models trained with DFL (α>0) consistently achieve higher returns, Sharpe ratios and lower maximum drawdowns than pure MSE models (α=0).","kind":"performance","quote":"Models with α> 0 consistently outperform pure MSE (α= 0), with optimal performance typically at intermediate values (α∈[0.25, 0.75])."},{"finding":"DFL induces systematic prediction biases: it overestimates returns for assets included in the portfolio and underestimates those excluded.","kind":"predictive-accuracy","quote":"This tilting mechanism leads to systematic prediction biases where DFL overestimates returns for assets included in portfolios while underestimating excluded assets."},{"finding":"The DFL gradient can be interpreted as tilting the MSE error vector by the inverse covariance matrix, incorporating inter‑asset correlations into learning.","kind":"theoretical","quote":"DFL's gradient can be interpreted as tilting the MSE-based prediction errors by the inverse covariance matrix, effectively incorporating inter-asset correlations into the learning process."},{"finding":"Increasing the strength of the MVO loss (higher α) raises the MSE loss but improves decision quality, confirming the trade‑off between prediction accuracy and portfolio regret.","kind":"statistical","quote":"Figure 2 shows that increasing the strength of the MVO loss leads to a consistent rise in the MSE loss. This trend is particularly"},{"finding":"The risk‑aversion parameter λ behaves as expected: higher λ reduces portfolio drawdowns and volatility, with the most pronounced effects at λ=5.0.","kind":"performance","quote":"The risk aversion parameter λ exhibits expected behavior across all models. Increasing λ reduces portfolio drawdowns and volatility, with the most pronounced effects at λ= 5.0."}],"reported_performance":{"strategy":"Decision-Focused Learning (DFL) for Mean-Variance Optimization","sharpe":1.302,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2605.01176","title":"Decision-Induced Ranking Explains Prediction Inflation and Excessive Turnover in SPO-Based Portfolio Optimization","paper_date":"2026-05-02","authors":["Yi Wang","Takashi Hasuike"],"empirical_rigor":5.5,"page_url":"https://thequant.space/flowcharts/decision-induced-ranking-explains-prediction-inflation-and-excessive-turnover/"},"question":"How does decision‑focused learning (SPO/DFL) bias predicted asset returns in portfolio selection?","this_paper_finding":"DFL induces systematic prediction biases: it overestimates returns for assets included in the portfolio and underestimates those excluded.","other_paper_finding":"SPO‑trained portfolios produce highly inflated return predictions and excessive turnover across all datasets.","other_paper_quote":"Table 1 and Figure 1 show that the SPO+-based DFL model produces highly inflated return predictions and excessive portfolio turnover across the three datasets.","note":"Both find that SPO/DFL inflates return forecasts; A notes overestimation for included assets and underestimation for excluded ones, while B reports inflation across all datasets and excessive turnover.","confidence":0.645}],"contradicts":[],"qualifies":[],"compared_with":35,"citations":{"cited_by_in_corpus":4,"by_independent_authors":2,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":2,"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":"2026-11-14T18:02:14Z"},"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/"}