{"paper_id":"2507.04208","title":"Behavioral Probability Weighting and Portfolio Optimization under Semi-Heavy Tails","authors":["Ayush Jha","Abootaleb Shirvani","Ali M. Jaffri","Svetlozar T. Rachev","Frank J. Fabozzi"],"paper_date":"2025-07-06","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2507.04208v1","page_url":"https://thequant.space/flowcharts/behavioral-probability-weighting-and-portfolio-optimization/","scores":{"empirical_rigor":6,"math_complexity":8.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a unified framework that integrates behavioral probability weighting into rational portfolio optimization using Gaussian and Normal-Inverse-Gaussian return models.","quote":"This paper develops a unified framework that integrates behavioral distortions into rational portfolio optimization by extracting implied probability weighting functions (PWFs) from optimal portfolios modeled under Gaussian and Normal-Inverse-Gaussian (NIG) return distributions.","summary":"This paper develops a unified framework that integrates behavioral distortions into rational portfolio optimization by extracting implied probability weighting functions (PWFs) from optimal portfolios modeled under Gaussian and Normal-Inverse-Gaussian (NIG) return distributions. Using DJIA…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Asset returns follow either a Gaussian or a Normal-Inverse-Gaussian (NIG) distribution.","type":"model","stated":"explicit","quote":"Using the thirty constituents of the Dow Jones Industrial Average, we construct mean–variance and CVaR99 efficient frontiers and extending the analysis to the Sharpe-maximizing and CVaR-maximizing tangent portfolios. Under two comparative distributions, Gaussian and Normal Inverse Gaussian (NIG), we derive the PWFs imp"},{"assumption":"Daily arithmetic returns are used to preserve aggregation properties and enable convex optimization.","type":"data","stated":"explicit","quote":"Daily arithmetic returns are computed from raw prices, consistent with discrete-time portfolio analysis and convex optimization requirements."},{"assumption":"Unadjusted closing prices exclude dividend effects, reflecting only capital appreciation.","type":"data","stated":"explicit","quote":"These unadjusted series exclude dividend effects and reflect capital appreciation alone, ensuring return calculations isolate price dynamics."},{"assumption":"Two risk‑free rate proxies (3‑month and 10‑year U.S. Treasury yields) are used as continuously‑compounded rates.","type":"market","stated":"explicit","quote":"We employ two proxies for the risk‑free rate: the three‑month and ten‑year U.S. Treasury yields, reported as annualized, continuously compounded rates."},{"assumption":"Out‑of‑sample performance is evaluated over a 100‑day rolling window.","type":"statistical","stated":"explicit","quote":"Rolling 4‑year window Generate 100‑day out‑of‑sample returns vs. DJIA benchmark"}]},"datasets":{"status":"available","items":[{"name":"DJIA constituent price series","provider":"Bloomberg","kind":"market-prices","access":"commercial","period":"2020‑11‑19 to 2025‑05‑03","universe":"30 Dow Jones Industrial Average constituents","quote":"The dataset includes unadjusted closing prices obtained from Bloomberg as of June 14, 2025, covering the period from November 19, 2020 to May 3, 2025, yielding five years of daily data.","known_id":"bloomberg"},{"name":"U.S. Treasury yields","provider":"U.S. Treasury","kind":"macro","access":"public","period":null,"universe":"3‑month and 10‑year yields","quote":"We employ two proxies for the risk‑free rate: the three‑month and ten‑year U.S. Treasury yields, reported as annualized, continuously compounded rates."}],"known":[{"id":"bloomberg","name":"Bloomberg","provider":"Bloomberg L.P.","kind":"market-prices","access":"commercial","quote":"The dataset includes unadjusted closing prices obtained from Bloomberg as of June 14, 2025, covering the period from November 19, 2020 to May 3, 2025, yielding five years of daily data.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R1","candidates":{"front_matter":"no","evidence_card":"unclear","evidence_card_quoted":false},"meaning":"a trading or execution result is reported without costs"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Heavier‑tailed NIG laws produce probability‑weighting functions with sharper inverse‑S curvature, reflecting heightened fear and greed.","kind":"theoretical","quote":"Our results reveal that heavier‑tailed laws produce PWFs with sharper inverse‑S curvature, capturing heightened fear (overweighting of extreme losses) and greed (overweighting of extreme gains)."},{"finding":"CVaR99 optimization amplifies behavioral distortions more than mean‑variance optimization.","kind":"theoretical","quote":"These effects are magnified under CVaR99 optimization relative to mean‑variance, indicating that tail‑risk minimization intensifies behavioral distortions."},{"finding":"The NIG distribution provides a superior fit to out‑of‑sample returns compared to the Normal distribution across AIC, BIC, and KS statistics.","kind":"statistical","quote":"In all cases, the NIG distribution dominates on both likelihood‑based and non‑parametric metrics, reflecting its superior ability to accommodate skewness and excess kurtosis observed in realized returns."},{"finding":"For the Sharpe‑optimized portfolio, the estimated NIG tail index (α≈0.39) captures jump behavior that the Normal distribution cannot replicate.","kind":"statistical","quote":"In the Sharpe‑optimized portfolio, the estimated NIG parameters imply a pronounced tail index (α ≈0.39), capturing discrete jump behavior that standard Normal distributions cannot replicate."},{"finding":"Shifts in the term structure of risk‑free rates alter the curvature of the implied probability‑weighting function.","kind":"theoretical","quote":"We show that increasing tail fatness amplifies these distortions and that shifts in the term structure of risk‑free rates alter their curvature."}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2505.13019","title":"Characterizing asymmetric and bimodal long-term financial return distributions through quantum walks","paper_date":"2025-05-19","authors":["Stijn De Backer","Luis E. C. Rocha","Jan Ryckebusch","Koen Schoors"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/characterizing-asymmetric-and-bimodal-long-term-financial-re/"},"question":"How can long‑term financial return distributions be modeled beyond the Gaussian assumption?","this_paper_finding":"For the Sharpe‑optimized portfolio, the estimated NIG tail index (α≈0.39) captures jump behavior that the Normal distribution cannot replicate.","other_paper_finding":"Return distributions for ∆t ≳100 trading days deviate significantly from a Gaussian profile.","other_paper_quote":"Figures 1, 3, 4 and 5 show that return distributions for ∆t ≳100 trading days can deviate significantly from a Gaussian profile.","note":"Both find Gaussian fails for Δt ≳ 100 days; B shows NIG tail index α≈0.39 captures jumps that A’s deviation highlights.","confidence":0.516}],"contradicts":[],"qualifies":[],"compared_with":55,"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":"2026-11-14T17:33:42Z"},"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/"}