{"paper_id":"2611.20911","title":"A Second Order Transport Scheme for Dynamic Portfolio Subspaces under a Varying Risk Metric","authors":["Alejandro Rodriguez Dominguez"],"paper_date":"2026-11-16","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2611.20911","page_url":"https://thequant.space/flowcharts/a-second-order-transport-scheme-for-dynamic-portfolio-subspaces-under-a-varying/","scores":{"empirical_rigor":7,"math_complexity":9,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a geometric‑mean alignment and a symmetric Procrustes transport scheme that cancel the leading local defect of risk‑weighted Procrustes, yielding reversible endpoint updates with second‑order global accuracy on non‑uniform partitions.","quote":"Geometric mean alignment and a symmetric Procrustes construction cancel the term and yield reversible updates with second order global accuracy. It has local error O(h3) and global error O(h2 max) for the selected connection. Proof.","summary":"We analyze the transport of portfolio exposures between moving subspaces under a varying positive definite risk metric. The reference evolution is the induced connection obtained from the ambient connection of minimum norm in fixed asset coordinates. An explicit expansion of risk weighted…","strategy":"geometric mean transport"},"assumptions":{"status":"available","items":[{"assumption":"Smooth paths of fixed rank and uniformly positive risk metrics are assumed for the convergence analysis.","type":"model","stated":"explicit","quote":"The convergence analysis assumes smooth paths of fixed rank and uniformly positive metrics, with spectral separation required when an eigenspace perturbation bound is used.","support":"full"},{"assumption":"Spectral separation is required when an eigenspace perturbation bound is used.","type":"statistical","stated":"explicit","quote":"The convergence analysis assumes smooth paths of fixed rank and uniformly positive metrics, with spectral separation required when an eigenspace perturbation bound is used.","support":"full"},{"assumption":"The risk metric is positive definite and may commute with its derivative in certain configurations.","type":"model","stated":"explicit","quote":"A rank two construction with an exact continuous solution shows that this term can reduce global accuracy to first order even when the metric commutes with its derivative. The analysis also characterizes the constant metric, fixed subspace, and other configurations in which the defect vanishes and risk weighted Procrustes retains second order accuracy.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"48 industry portfolios","provider":"Kenneth R. French Data Library","kind":"market-prices","access":"public","period":null,"universe":"48 industry portfolios","quote":"We therefore add a historical geometry diagnostic using the 48 industry panel and the same Ledoit–Wolf metric and rank five representation used in the portfolio experiment. The first panel contains 48 industry portfolios and the second contains 100 portfolios formed at the intersections of ten size groups and ten book to market groups.","support":"full","known_id":"french-library"},{"name":"100 size‑value portfolios","provider":"Kenneth R. French Data Library","kind":"market-prices","access":"public","period":null,"universe":"100 portfolios formed at intersections of ten size groups and ten book‑to‑market groups","quote":"The first panel contains 48 industry portfolios and the second contains 100 portfolios formed at the intersections of ten size groups and ten book to market groups. The candidate sample is 2 January 1990 through 30 April 2026.","support":"full","known_id":"french-library"},{"name":"Manufactured examples","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"Manufactured examples confirm the predicted defect and convergence rates. Noncommuting metric path and convergence order The second manufactured problem uses six assets and a rank three representation.","support":"full"}],"known":[{"id":"french-library","name":"Kenneth French Data Library","provider":"Kenneth R. French","kind":"market-prices","access":"public","quote":"Historical portfolio implementation The financial exercise uses daily value weighted returns from the Kenneth R. French Data Library (Fama and French, 1993; French, 2026).","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":true,"detail":"Proportional transaction costs are applied; turnover is reported annualized in the primary results table.","quote":"Signals, constraints, costs, and persistence settings are common to all methods, as specified in Section C. Table 4: Primary results after proportional costs.","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Proportional transaction costs are applied; turnover is reported annualized in the primary results table.","card_quote":"Signals, constraints, costs, and persistence settings are common to all methods, as specified in Section C. Table 4: Primary results after proportional costs.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"The symmetric Procrustes update has local error O(h^3) and global error O(h^2) for the selected connection, confirming second‑order accuracy on non‑uniform partitions.","kind":"theoretical","quote":"It has local error O(h3) and global error O(h2 max) for the selected connection. Proof.","support":"full"},{"finding":"The historical geometry diagnostic predicts the subsequent separation between risk‑weighted Procrustes and both second‑order updates, while the geometric‑mean and symmetric constructions remain much closer to each other.","kind":"predictive-accuracy","quote":"A historical geometry stress test then estimates the defect from preceding observations and finds that it predicts the subsequent separation between risk weighted Procrustes and both second order updates, while the two second order schemes remain much closer to each other. The historical geometry diagnostic links the theoretical defect to estimated market geometry without using portfolio performance: the ex ante","support":"full"},{"finding":"Geometric‑mean transport reduces annualized turnover by 48.06% in the industry panel and 42.89% in the 100‑portfolio panel relative to independent local refitting.","kind":"performance","quote":"TO local TO geo. ∆CE, bp [95%] 48 industries 8.458% 8.364% 2.997 1.557 −9.37 [−95.93, 81.48] 100 size–value 6.781% 5.609% 4.654 2.658 −117.20 [−279.61, −1.64] At the primary persistence weight, geometric mean transport reduces annualized turnover by 48.06% in the industry panel and 42.89% in the 100 portfolio panel relative to independent local refitting, with paired intervals excluding zero in both cases.","support":"full"},{"finding":"Annual certainty‑equivalent differences relative to local refitting are –9.37 bp (industry) and –117.20 bp (100‑portfolio), with the industry interval spanning zero and the 100‑portfolio interval below zero.","kind":"performance","quote":"The corresponding annual certainty equivalent differences are −9.37 and −117.20 basis points.","support":"partial"},{"finding":"Both metric‑based second‑order updates (geometric‑mean and symmetric Procrustes) have smaller discrepancies from the composed reference than ambient projection, and remain substantially closer to each other than to risk‑weighted Procrustes.","kind":"theoretical","quote":"The two second order rules remain substantially closer to each other than either is to risk weighted Procrustes. 4.6. Both metric based updates have smaller discrepancies from the composed reference than ambient projection, with the geometric mean and risk weighted Procrustes results closer to each other.","support":"full"}],"reported_performance":{"strategy":"geometric mean transport","sharpe":null,"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":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":17,"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":3,"extracted_at":"2026-11-17T15:29:18Z"},"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/"}