{"paper_id":"2507.09196","title":"Functionally Generated Portfolios Under Stochastic Transaction Costs: Theory and Empirical Evidence","authors":["Nader Karimi","Erfan Salavati"],"paper_date":"2025-07-12","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2507.09196v1","page_url":"https://thequant.space/flowcharts/functionally-generated-portfolios-under-stochastic-transacti/","scores":{"empirical_rigor":6,"math_complexity":9.2,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper extends stochastic portfolio theory to include proportional stochastic transaction costs and derives a cost‑adjusted master inequality providing lower bounds for functionally generated portfolios.","quote":"we extend SPT to a continuous-time equity market with proportional, stochastic transaction costs. We derive closed-form lower bounds on cost‑adjusted relative wealth for a large class of functionally generated portfolios; these bounds provide sufficient conditions for relative arbitrage to survive random costs.","summary":"Assuming frictionless trading, classical stochastic portfolio theory (SPT) provides relative arbitrage strategies. However, the costs associated with real-world execution are state-dependent, volatile, and under increasing stress during liquidity shocks. Using an Ito diffusion that may be connected…","strategy":"Diversity-weighted functionally generated portfolio (p = 0.7)"},"assumptions":{"status":"available","items":[{"assumption":"Asset prices follow an Itô diffusion with drift b_t and nonsingular volatility matrix σ_t.","type":"market","stated":"explicit","quote":"with drift b_t and nonsingular volatility matrix σ_t."},{"assumption":"Trading incurs a proportional cost κ_t per unit turnover, where κ_t follows a mean‑reverting stochastic process.","type":"execution","stated":"explicit","quote":"Trading incurs proportional cost κ_t per unit turnover, where dκ_t = α (¯κ −κ_t) dt + η dB_t, κ_0 > 0, and B may be correlated with the market Brownian motion W."},{"assumption":"Portfolios are rebalanced piecewise‑constant on a mesh Δ (e.g., weekly).","type":"model","stated":"explicit","quote":"Exploiting piecewise‑constant rebalancing, we derive a cost‑adjusted master inequality (Theorem 3.1) that is valid pathwise and isolates the tug‑of‑war between excess growth and stochastic cost erosion."},{"assumption":"The excess‑growth integral ∫_0^T γ_t dt diverges almost surely (market is ε‑diverse).","type":"statistical","stated":"explicit","quote":"Step 1: Divergence of the drift term. By assumption (i) DT :="},{"assumption":"Cost erosion grows sub‑linearly, of order √T.","type":"other","stated":"explicit","quote":"A Milstein Monte‑Carlo experiment confirms the theoretical O(√T) cost growth."},{"assumption":"The CRSP small‑cap sample excludes stocks with zero or extreme (>500 bps) spreads and days with missing quotes.","type":"data","stated":"explicit","quote":"The panel is filtered for zero and extreme (>500 bps) spreads and for days with missing quotes, leaving a balanced sample of 981 permnos."}]},"datasets":{"status":"available","items":[{"name":"CRSP small‑cap daily database","provider":"CRSP","kind":"market-prices","access":"commercial","period":"1994‑2024","universe":"one thousand smallest U.S. common stocks","quote":"We use the CRSP daily database to extract the one thousand smallest U.S. common stocks","known_id":"crsp"},{"name":"Synthetic Monte‑Carlo simulation","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":"50‑stock small‑cap universe","quote":"Table 1 summarises the baseline parameters. Stock volatilities are drawn i.i.d. from Unif(15%, 35%) and kept constant across paths..."}],"known":[{"id":"crsp","name":"CRSP","provider":"Center for Research in Security Prices","kind":"market-prices","access":"commercial","quote":"Finally, we use intraday bid-ask spreads as a stand-in for cost volatility in a back-test of CRSP small-cap data (1994--2024).","mentions":8,"used_in":3},{"id":"taq","name":"NYSE TAQ","provider":"NYSE","kind":"market-prices","access":"commercial","quote":"Daily NBBO bid and ask quotes are obtained from TAQ and matched to CRSP permnos; the mid-price is 1 2(bid + ask) and the half–spread defines the proportional cost κi,t = 1 2 aski,t −bidi,t midi,t .","mentions":1,"used_in":1}],"sample":{"start":"1994","end":"2024","market":"CRSP small-cap US equities","asset_class":"equities","frequency":"monthly"},"data_resolution":"unclear"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":true,"detail":"Proportional cost κ_t per unit turnover with stochastic dynamics dκ_t = α(¯κ−κ_t)dt + η dB_t; spreads are measured from intraday NBBO bid‑ask quotes.","quote":"Trading incurs proportional cost κ_t per unit turnover, where dκ_t = α (¯κ −κ_t) dt + η dB_t, κ_0 > 0, and B may be correlated with the market Brownian motion W.","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Proportional cost κ_t per unit turnover with stochastic dynamics dκ_t = α(¯κ−κ_t)dt + η dB_t; spreads are measured from intraday NBBO bid‑ask quotes.","card_quote":"Trading incurs proportional cost κ_t per unit turnover, where dκ_t = α (¯κ −κ_t) dt + η dB_t, κ_0 > 0, and B may be correlated with the market Brownian motion W.","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"Diversity‑ and entropy‑weighted functionally generated portfolios beat the value‑weighted benchmark by 2.9–3.6 percentage points annually after cost deduction.","kind":"performance","quote":"diversity- and entropy-weighted portfolios continue to beat the value-weighted benchmark by 3.6 and 2.9 percentage points annually, respectively, after cost deduction."},{"finding":"Monte‑Carlo simulations confirm that cumulative trading cost grows at the rate O(√T), matching the theoretical bound.","kind":"theoretical","quote":"A Milstein Monte‑Carlo experiment confirms the theoretical O(√T) cost growth."},{"finding":"Coarser rebalancing meshes (weekly or bi‑weekly) preserve relative arbitrage even under temporary spread spikes up to 80 bps.","kind":"performance","quote":"A weekly or bi‑weekly schedule preserves relative arbitrage even under temporary spikes."},{"finding":"Cost erosion is sub‑linear (∝√t), and the excess‑growth term eventually dominates, allowing the diversity‑weighted portfolio to recover costs after about 1.5 years.","kind":"theoretical","quote":"Cost erosion indeed grows sub‑linearly (∝√t), in line with the theoretical bound."}],"reported_performance":{"strategy":"Diversity-weighted functionally generated portfolio (p = 0.7)","sharpe":null,"sharpe_benchmark":null,"annual_return_pct":10.7,"total_return_pct":null,"max_drawdown_pct":65},"out_of_sample":{"method":"none","headline_out_of_sample":false}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2303.00858","title":"Quantifying dimensional change in stochastic portfolio theory","paper_date":"2023-03-01","authors":["Erhan Bayraktar","Donghan Kim","Abhishek Tilva"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/quantifying-dimensional-change-in-stochastic-portfolio-theory/"},"question":"Does the excess‑growth term dominate portfolio growth and offset costs in diversity‑weighted portfolios?","this_paper_finding":"Cost erosion is sub‑linear (∝√t), and the excess‑growth term eventually dominates, allowing the diversity‑weighted portfolio to recover costs after about 1.5 years.","other_paper_finding":"Excess growth (EG) term contributes significantly to the growth in log relative wealth for diversity‑weighted portfolios, especially for smaller p values.","other_paper_quote":"Especially, in graphs (b) and (c), the excess growth significantly contributes to the growth in the log relative wealth Uψ.","note":"Both findings show EG is significant; B quantifies cost erosion ∝√t and recovery after ~1.5 years, confirming A’s EG impact for small p values.","confidence":0.435}],"contradicts":[],"qualifies":[],"compared_with":27,"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":"2026-11-14T17:32:58Z"},"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/"}