{"paper_id":"2609.31345","title":"On the asymptotic shape of quantile surfaces","authors":["Florian Gach","Simon Hochgerner"],"paper_date":"2026-09-30","paper_type":"theoretical","paper_url":"https://arxiv.org/pdf/2609.31345","page_url":"https://thequant.space/flowcharts/on-the-asymptotic-shape-of-quantile-surfaces/","scores":{"empirical_rigor":2,"math_complexity":9,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"We provide a formula for the Hessian of the quantile function with respect to the control vector and show that for linear pay‑offs with log‑normal risk factors the quantile surface is concave in the left tail and convex in the right tail, with a universal separation of scale and shape in deep tails.","quote":"In this regard, our contributions are as follows: (1) For nonlinear h we provide a formula for the Hessian of qα with respect to u (Theorem 1.1). (2) For linear h and log-normal X1, . . . , Xn we show that qα(u) is concave in u for small α and convex in u for large α.","summary":"This article is concerned with the asymptotic shape of quantile surfaces, defined as the set of quantiles at a given level α generated by a controlled one-dimensional distribution. Specifically, when the distribution arises as a linear combination of log-normal random variables and the control is a…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The payoff function h is linear in the control vector.","type":"model","stated":"explicit","quote":"(2) For linear h and log-normal X1, . . . , Xn we show that qα(u) is concave in u for small α and convex in u for large α.","support":"full"},{"assumption":"Risk factors X1,…,Xn are log‑normal random variables.","type":"model","stated":"explicit","quote":"(2) For linear h and log-normal X1, . . . , Xn we show that qα(u) is concave in u for small α and convex in u for large α.","support":"full"},{"assumption":"Control coefficients ui are positive.","type":"model","stated":"explicit","quote":"This article is concerned with the asymptotic shape of quantile surfaces, defined as the set of quantiles at a given level $α$ generated by a controlled one-dimensional distribution.","support":"full"},{"assumption":"Correlation between log‑factors is not too high for the general concavity result.","type":"statistical","stated":"explicit","quote":"More precisely, we show that for all u with ui > 0 there exists a level α0 such that qα is strictly concave for all α ≤α0, when there is not too much positive correlation between log-factors, log(Xi).","support":"full"},{"assumption":"When correlation is high, results are restricted to the case n=2.","type":"model","stated":"explicit","quote":"On the other hand, when correlation between log-factors is high, we simplify the presentation by restricting to n = 2 and show that concavity still holds in the small α-regime (Theorem 3.6). Similarly, for high α we also restrict to n = 2 and show that convexity holds (Theorem 3.12).","support":"full"},{"assumption":"Moments of h(u,X) exist for all orders of partial differential operators D(k).","type":"statistical","stated":"explicit","quote":"Assume that E[D(k)h(u, X)] < ∞for all u ∈U for all k ∈N ∪{0}.","support":"full"}]},"datasets":{"status":"none_found","items":[],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"none"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"Quantile surfaces are globally concave in the left tail (α→0) and globally convex in the right tail (α→1) for linear pay‑offs with log‑normal risk factors.","kind":"theoretical","quote":"Specifically, when the distribution arises as a linear combination of log-normal random variables and the control is a vector of positive coefficients, we prove that quantile surfaces are globally concave in the left tail ($α\\to0$) and globally convex in the right tail ($α\\to1$). (2) For linear h and log-normal X1, . . . , Xn we show that qα(u) is concave in u for small α and convex in u for large α.","support":"full"},{"finding":"For all positive control vectors there exists a level α0 such that the quantile function is strictly concave for all α≤α0 when correlation between log‑factors is not too high.","kind":"theoretical","quote":"More precisely, we show that for all u with ui > 0 there exists a level α0 such that qα is strictly concave for all α ≤α0, when there is not too much positive correlation between log-factors, log(Xi).","support":"full"},{"finding":"Even with high correlation, concavity holds in the small‑α regime for n=2 (Theorem 3.6) and convexity holds in the large‑α regime for n=2 (Theorem 3.12).","kind":"theoretical","quote":"On the other hand, when correlation between log-factors is high, we simplify the presentation by restricting to n = 2 and show that concavity still holds in the small α-regime (Theorem 3.6). Similarly, for high α we also restrict to n = 2 and show that convexity holds (Theorem 3.12).","support":"full"},{"finding":"In deep tail regimes the quantile function separates into a scale function c(α) and a universal shape function s(u), independent of the log‑factor means.","kind":"theoretical","quote":"(3) In the deep tail regime, for linear pay-off and log-normal risk factors, we find that qα(u) develops a universal structure, where scale and shape are separated and the location parameter of the log-factors is forgotten: as α →0 and as α →1 we explicitly identify (respectively) level functions, c = c(α), and shape functions, s = s(u), such that qα(u) ∼c(α)s(u). See Section 4. The shape function q0 is universal in","support":"full"},{"finding":"Value‑at‑Risk is sub‑additive in the loss tail for linear portfolios of log‑normal factors, establishing a new instance of VaR sub‑additivity.","kind":"theoretical","quote":"Corollary 3.9 (Subadditivity of Value-at-Risk in loss tail). Hence VaRα[L1 + L2] ≤VaRα[L1] + VaRα[L2] i.e., sub-additivity holds. Proof.","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"none","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":2,"extracted_at":"2026-11-14T19:39:17Z"},"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/"}