{"paper_id":"2410.21858","title":"Joint Estimation of Conditional Mean and Covariance for Unbalanced Panels","authors":["Damir Filipovic","Paul Schneider"],"paper_date":"2024-10-29","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2410.21858v5","page_url":"https://thequant.space/flowcharts/joint-estimation-of-conditional-mean-and-covariance-for-unba/","scores":{"empirical_rigor":6,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a nonparametric, kernel‑based joint estimator for conditional mean and covariance matrices in large, unbalanced panels (the COCO estimator).","quote":"We develop a nonparametric, kernel-based joint estimator for conditional mean and covariance matrices in large and unbalanced panels.","summary":"We develop a nonparametric, kernel-based joint estimator for conditional mean and covariance matrices in large and unbalanced panels. The estimator is supported by rigorous consistency results and finite-sample guarantees, ensuring its reliability for empirical applications. We apply it to an…","strategy":"COCO conditional mean‑covariance estimator (cMVE portfolio)"},"assumptions":{"status":"available","items":[{"assumption":"The conditional covariance can be decomposed into systematic and idiosyncratic components.","type":"model","stated":"explicit","quote":"Specifically, we adopt the common assumption that the conditional covariance can be decomposed into a systematic and an idiosyncratic component."},{"assumption":"Data points ξt are i.i.d. draws from a distribution with support in Ξ.","type":"statistical","stated":"explicit","quote":"We assume that the data points ξt, t = 0, . . . , T −1, are i.i.d. drawn from a distribution P with support in Ξ."},{"assumption":"Regularization parameters λsy and λid are set to zero, relying on low‑rank approximation for implicit regularization.","type":"model","stated":"explicit","quote":"For simplicity, we set both regularization parameters to the boundary values, λsy = λid = 0, in the implementation."},{"assumption":"The conditional mean is assumed to be zero in the baseline model used for comparison.","type":"model","stated":"explicit","quote":"model in (28) predicts it to be zero, as the conditional mean is identically zero in that case."}]},"datasets":{"status":"available","items":[{"name":"Unbalanced monthly US stock returns (Gu et al. 2020a)","provider":"Gu et al.","kind":"market-prices","access":"public","period":"1962-2021","universe":"approximately 30,000 US common stocks, average ~6,200 per month","quote":"We use unbalanced monthly stock data compiled by Gu et al. (2020a), covering March 1957 to December 2021. This dataset includes approximately 30,000 stocks, with an average of 6,200 stocks per month."}],"known":[],"sample":{"start":"1962","end":"2021","market":"US common stock excess returns","asset_class":"equities","frequency":"monthly"},"data_resolution":"weekly-monthly"},"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":"Idiosyncratic risk explains, on average, more than 75% of the cross‑sectional variance.","kind":"statistical","quote":"we find that idiosyncratic risk explains, on average, more than 75% of the cross‑sectional variance."},{"finding":"The COCO estimator significantly outperforms a baseline model that accounts only for idiosyncratic risk.","kind":"performance","quote":"By jointly assessing both moments, the COCO estimator significantly out‑performs a baseline model that accounts only for idiosyncratic risk."},{"finding":"The conditional mean‑variance efficient (cMVE) portfolio built from COCO estimates achieves substantial annualized out‑of‑sample Sharpe ratios, markedly outperforming equal‑weighted portfolios.","kind":"performance","quote":"The conditional mean‑variance efficient (cMVE) portfolio constructed from the COCO estimates achieves substantial annualized out‑of‑sample Sharpe ratios, markedly outperforming equal‑weighted portfolios over the entire sample period."},{"finding":"Predicted maximum Sharpe ratios increase with the number of systematic factors, reaching up to eight for the Gaussian kernel with m=40.","kind":"predictive-accuracy","quote":"The Gauss kernel achieving peaks of a predicted maximum Sharpe ratio of up to eight for m = 40."},{"finding":"Realized out‑of‑sample Sharpe ratios from cMVE portfolios exceed five for both kernels, and are largely unrelated to the Fama‑French five factors.","kind":"performance","quote":"Higher values of m tend to generate higher realized Sharpe ratios, with peaks exceeding five for both kernels"}],"reported_performance":{"strategy":"COCO conditional mean‑covariance estimator (cMVE portfolio)","sharpe":5,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"walk-forward","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":21,"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":"2026-11-14T17:57:00Z"},"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/"}