{"paper_id":"2501.10564","title":"Crossing penalised CAViaR","authors":["Tibor Szendrei"],"paper_date":"2025-01-17","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2501.10564v1","page_url":"https://thequant.space/flowcharts/crossing-penalised-caviar/","scores":{"empirical_rigor":6,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a crossing‑penalised CAViaR estimator that jointly estimates multiple quantiles while penalising quantile crossing.","quote":"we penalise quantile crossing, and thereby propose a non-crossing CAViaR estimator.","summary":"Dynamic quantiles, or Conditional Autoregressive Value at Risk (CAViaR) models, have been extensively studied at the individual level. However, efforts to estimate multiple dynamic quantiles jointly have been limited. Existing approaches either sequentially estimate fitted quantiles or impose…","strategy":"DynQR"},"assumptions":{"status":"available","items":[{"assumption":"Stationarity and Ergodicity of the (Yt, Xt) process.","type":"statistical","stated":"explicit","quote":"Assumption 1. Stationarity and Ergodicity: The sequence (Yt,XT t ) is a stationary and ergodic stochastic process."},{"assumption":"Continuity of the conditional density of Yt.","type":"statistical","stated":"explicit","quote":"Assumption 2. Continuity of Density: Yt is continuously distributed for each t given the parameters ω ∈Ω, such that the CDF (Ft) and the PDF (ft) are continuous on R."},{"assumption":"Continuity of the coefficient vector over a compact set.","type":"statistical","stated":"explicit","quote":"Assumption 3. Continuity of coefficient vector: Let D be a compact subset of Rℓ. The sequence of Qt(ω, δ): (1) is measurable for each t and δ ∈D, and; (2) is continuous for each t and ω ∈Ω on D."},{"assumption":"Correct specification and identification of the true parameter δ*.","type":"statistical","stated":"explicit","quote":"Assumption 4. Correct Specification and Identification: (1) There exists δ∗∈D such that P[Yt ≤Qτq,t(·, δ∗)] = τq ∀t, q."}]},"datasets":{"status":"available","items":[{"name":"FTSE100 2008","provider":null,"kind":"market-prices","access":"unknown","period":"2008","universe":"FTSE100 index","quote":"The method is applied to FTSE100 during 2008."},{"name":"Monte Carlo synthetic datasets","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"In total, we simulate 50 Monte Carlo datasets generated with two sample sizes T = {50, 200}."}],"known":[],"sample":{"start":"2008-01","end":"2008-12","market":"FTSE100 index","asset_class":"equities","frequency":"daily"},"data_resolution":"daily"},"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":"forecast-accuracy","findings":[{"finding":"Quantile models with lagged quantiles significantly outperform traditional quantile regression in capturing FTSE100 volatility during the 2008 crisis.","kind":"performance","quote":"Our results demonstrate that quantile models that incorporate lagged quantiles significantly outperform traditional quantile regression methods in capturing the evolving volatility of the FTSE100 index during the financial crisis of 2008."},{"finding":"The DynQR estimator with λ > 0 yields superior forecast accuracy under heightened market uncertainty.","kind":"predictive-accuracy","quote":"The proposed DynQR estimator, with λ > 0, showed superior forecast accuracy highlighting its robustness in forecasting under conditions of increased market uncertainty."},{"finding":"DynQR provides the largest improvement over traditional quantile estimators at the tails, reflecting inertia in FTSE100 volatility and skewness.","kind":"performance","quote":"the DynQR estimator has the biggest improvement over traditional quantile estimators at the tails, highlighting how volatility (and potentially skewness) in the FTSE100 had inertia."},{"finding":"In Monte Carlo experiments, the penalised estimator matches traditional QR when λ = 0 and reduces coefficient bias as λ increases.","kind":"statistical","quote":"Through a Monte Carlo study, we find that the proposed estimator is capable of getting solutions extremely close to the traditional QR in QAR(1) settings when λ = 0, and even reduce coefficient bias by increasing the penalty on crossing."},{"finding":"The CMA‑ES optimisation is not influenced by the starting point for the iterations.","kind":"theoretical","quote":"algorithm is not influenced by the starting point for the iterations."}],"reported_performance":{"strategy":"DynQR","sharpe":null,"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":48,"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":"2026-11-14T17:47:33Z"},"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/"}