Boundary correction yields unbiased Greeks for barrier options
The formula removes smoothing bias and reduces variance by orders of magnitude, tested on Heston and Hull-White models.
q-fin
Quantitative Finance
sort pith recommended most recent
EVaR risk parity beats equal weight on Sharpe ratio in three markets
Backtests on ETFs, momentum deciles, and sector funds show consistent but often insignificant improvements; ICA-based ERC has higher…
· “Entropic Value-at-Risk parity for tempered stable returns”
Deep RL market maker survives persistent buy/sell pressure
A Bayesian change-point filter and queue-aware exposure signal prevent inventory saturation under regime-switching order flow.
· “Deep Learning of Robust Market Making under Regime-Switching Order Flow”
Skew stickiness ratio tends to 2 at short expiry
First rigorous proof under mild conditions settles Bergomi's conjecture using Watanabe expansion.
· “Short-maturity skew stickiness ratio under local volatility”
Market-aware weighting lifts extreme-event forecast scores 11-15%
Joint-extreme weighting wins for both tails across three S&P 100 sectors.
· “Market-Informed Networks for Modeling and Forecast Evaluation of Financial Extremes”
Parse Uniswap events with a single Python library
dexamine converts transaction logs into token amounts, pool state, and execution metadata for empirical research.
· “dexamine: A Python package for Uniswap event data on Ethereum”
Surgical hubs boost physician productivity 14.5%
Dedicated elective facilities in England increase output per physician pound, with larger gains from standalone and multiple hubs.
· “High Volume Low Complexity Surgical Hubs in England: Can They Improve Physician Productivity?”
Italian B2B network is scale-free
Analysis of 1.6M firms reveals a few hubs dominate; targeted shocks could spread fast through the economy.
· “Italian Business-to-Business Invoicing Data: A Network Analysis”
Known bias warps both the story and the listener's trust
Lab experiment confirms: senders exaggerate more under conflict of interest, receivers discount more — net persuasion only when bias is low.
17% of wind, 4% of solar curtailed in Germany's negative-price hours
Almost all wind had an incentive to stop at deeply negative prices; only half did. The public bill: about €2.5 billion.
Mixed adversarial training beats single-attack defences in credit scoring
Four-attack benchmark on 400,000 loans shows robustness transfers within gradient methods but not beyond; mix your threats.
人口下降如何重塑城市：小城市消失的行业为东京创造工资溢价
基于日本面板数据的证据表明，收缩导致行业从小城市消失并集中到东京，驱动迁移溢价，并且在变得太小之前预报了哪些地点应该被保留。
Drug alerts doctors can ignore cut errors by 8.6%
Real-time, non-mandatory warnings also prompt doctors to proactively avoid harmful drug pairs.
One local loss rule yields two different diffusion consensus models
Expert-first inflates volatility with drift dispersion; candidate-first picks a Hessian-corrected barycenter covariance.
· “Geometric and Arithmetic Likelihood Aggregation for Diffusions with Heterogeneous Volatility”
Gaussian test splits rough from multifractal volatility
Under small intermittency, the Log S-fBM model yields a normal test statistic whose power against nonzero Hurst exponent is expressed in…
Welfare economics can rescue climate policy from dueling certitudes
A respected economist argues that both mainstream and contrarian reports lack formal tradeoff analysis needed for credible policy.
· “Credible Discourse on Climate Policy: Beyond Dueling Certitudes”
BNPL users are financially stretched
A literature review finds mixed welfare effects and one giant blind spot: BNPL loans rarely reach credit bureaus.
· “Buy Now, Pay Later: Academic Insights and Open Policy Questions”
Middle groups get neutral treatment only with two shared traits
Two shared markers made a group get rule-following treatment; one marker left it lumped with outsiders.
· “Going Beyond the In-/Out-Group Dichotomy: Investigating Altruism towards Middle-Groups”
Variance Swap Delta Falls to Zero for Standard Smile Models
A paper proves that when the volatility smile depends only on moneyness, the variance swap price is insensitive to underlying moves, and…
High operating costs kill SMR profitability
Even with zero construction cost, most SMRs cannot pay back within their lifetime due to hourly losses from high marginal costs.
· “The consequences of high SMR operating costs in electricity markets”
Short-selling bans slice tail risk but blunt liquidity
New evidence from 2020 Europe: bans reduce max drawdowns by 3 percentage points but widen spreads by 13 basis points, with the trade-off…
One entropy rule survives numeraire changes
Switching units shifts the forward-entropy objective by a constant; the reverse rule can split the price of one claim.
· “Numeraire Invariance of Entropy-Projected Martingale Measures”
Persistent memory and stochastic returns improve stock signal discovery
AlphaRJM wins 10 of 12 metric comparisons by retaining evaluation history across formula episodes.
· “AlphaRJM: Reward-Jump Memory for Stochastic Return-Guided Alpha Discovery”
Semi-closed hedges converge in rough Hawkes–Heston
Kernel regularization yields strategies whose gains and errors approach the variance-optimal limit in the original market.
· “Variance-Optimal Hedging in the Rough Hawkes--Heston Model”
AI collaboration boosts device innovation and productivity in medtech
Small firms benefit most from external AI research partnerships, raising labour productivity by up to 1.9% per percentage point of device…
· “AI Innovation and Firm Performance in the Medical Device Industry”
Time averaging hid true inequality in Bouchaud-Farmer model
Correcting the estimand raises the Gini from 0.7 to 0.88 and shows the Pareto exponent argument fails at the model's calibration.
· “Self-Fulfilling Prophecies, Quasi Nonergodicity, and Wealth Inequality: A Comment”
Wholesale prices hit $46–49/MWh even with optimal build-out
A model of the Texas grid shows why new power plants cannibalize their own revenue, so investment stalls as AI load grows.
Nyström attention equals full attention at O(mN) cost
Nyström with 32 landmarks replicates ranking accuracy while cutting computation from O(N²) to O(N)
· “Nystr\"om Attention Matches Full Attention for Cross-Sectional Stock Prediction”
A simple model ties a complex rival at forecasting pro LoL maps
On 5,135 pro maps, a minimal logistic forecaster matches a heavy rival and stays within noise of Polymarket.
· “Pre-game paired-comparison modeling of professional League of Legends map outcomes”
A log-normal duration law wins for order-flow regimes
Duration-aware changepoint detection improves NASDAQ order-flow forecasts; multivariate BOCPDMS falls short.
Yield curves are modern inventions
Five institutional conditions had to converge before a term structure of interest rates could be observed.
· “Historical Reflections on Interest Rates and the Emergence of the Yield Curve”
Simple public-data rules nearly double 60/40 Sharpe ratio
Volatility targeting plus a small Markowitz optimizer lifts Sharpe from 0.56 to 1.08 over 2006–2026, net of costs.
Two public-data ratios reveal when DeFi pools turn fragile
The ratios flag the coin-price crash that would trigger mass liquidations or defaulted deposits, day by day.
Noisy information overturns standard disclosure predictions
Even small processing noise unravels full disclosure and creates two distinct equilibria with opposite comparative statics
A profit mandate blunts LLM safety warnings by 6.8 points
Eight models, 3,600 trials: models spot the hazard, then talk themselves out of escalating — with no instruction to ignore risk.
· “The Profit Alignment Problem: How Profit Mandates Induce Alignment Failures in LLMs”
A $200B cyber catastrophe once per 206 years
Sub-$1B cyber kittens fit a power law, giving insurers a way to estimate events 100 times larger than MOVEit.
Live AI advice leaves risk preferences unchanged
An incentivized lottery experiment finds no evidence that a live AI advisor shifts risk-taking beyond its advice.
· “Access to Live AI Advice and Behavior Under Risk: An Incentivized Experiment”
Fixed-point solutions for q-Bass martingales exist for all marginals
Existence, uniqueness, stability and W∞-convergence of the fixed-point iteration for general marginals under the absolute continuity of q.
· “Fixed Points for the q-Bass Martingale: Existence, Stability, and Convergence”
One wave function fits 89% of stock trades
A probability-wave model from quantum mechanics explains collective decision patterns better than rational-agent theory.
· “Adaptive Entangled Game Modules in Artificial General Intelligence”
Simulation filter flattens volatility signature to one second
A non-recursive robust filter handles infinite-variance microstructure noise and yields reliable 1-second volatility estimates.
· “Filtering without recursion and some of its uses in financial economics”
Discrete Asian options priced in Volterra–Heston model
Semi-closed formulas for geometric Asian options and variance-optimal hedges via a measure-valued affine transform.
· “Pricing and Hedging of Discretely Monitored Asian Options in the Volterra-Heston Model”
Profit–bias identity: bookmaker profit as public prediction error
A new identity decomposes betting margins into hold, shading, and outcome covariance; tested on 1,139 MLB games with surprising null…
· “The profit-bias identity in sports betting: bookmaker profit as the public's prediction error”
This paper introduces a model for stock market volatility that treats positive and…
A new asymmetric GARCH model decomposes volatility asymmetry into level and memory channels, finding the level channel dominates…
· “Asymmetric Long-Memory GARCH: Sign-Dependent Kernel Injection in a Two-Dimensional Markov Chain”
Housing boom didn't boost Toronto car commuting
Null correlation between dwelling-value growth and commuting change suggests remote work decoupled relocation from daily travel.
One linear program now unifies matching
The same primal-dual structure prices scarce resources, couples distributions, and recovers preferences from data.
One curve shows every trailing return at once
Fixing the evaluation date and ranging over past purchase dates puts crossings and entry-date sensitivity on one graph
· “From Discrete Trailing Returns to a Continuous Graphical Profile: Return-to-Present Curves”
Annuity beats compound interest for slow-growing fish at low rates
Optimal harvesting model reveals a threshold growth rate where compound interest always wins for faster stocks.
Perron-Frobenius eigenvector cuts tail-risk error in lognormal sums
It outperforms Taylor-based and maximal-variance approximations on extreme-tail risk measures when volatility is high.
· “A Perron-Frobenius comonotonic approximation for sums of lognormals”
Neural network explanations yield a simple model that beats the standard benchmark
A parsimonious one-lag nonlinear model, derived from SHAP, captures the dynamics of returns and order flow.
China's 4,000-year rise reduces to one institutional logic
Flood-born collectivism, exam-tested elites, and state capability: one story that challenges development orthodoxy.
Beta-times-quantity model predicts stock returns with 1% R^2
A factor model using noise-trading flows explains expected stock returns as well as machine learning.
Creators overvalue their own ideas
Six studies show people submit more of their own ideas than others', and the extras are worse—even when told about the bias.
· “Creators Have Difficulty Abandoning Ideas They Generated”
Anti-correlated Heston undercuts local vol on every variance call
The standard Heston model itself refutes the claim that local volatility minimizes convex variance payoffs.
No public proof yet that AI earns durable net alpha
A review of equity and crypto evidence through 2026 finds technical progress but no sustainable, cost-aware returns.
Impact cost with power-law decay bans arbitrage round trips
Coupling trades across contracts and days keeps cost convex as liquidity shifts, so spread legs net their impact.
One qubit learns Bitcoin volatility’s rough fractal structure
A three-parameter quantum circuit reproduces anti-persistence and multifractality without a pre-specified model.
This paper studies whether diversifying a portfolio still lowers risk when the dependence…
Under complete dependence uncertainty, robust portfolio selection for VaR, ES, RVaR, and often standard deviation is optimized by…
· “Portfolio Diversification and Concentration under Dependence Uncertainty: A Majorization Approach”
Optimal rare-event stratification ignores class imbalance under weighted loss
When the loss already reweights the minority class, the Neyman-optimal split between positive and negative strata depends only on…
· “Optimal Stratified Allocation for Rare-Event Onset Forecasting in Dependent Sequences”
Non-Marovian volatility: stitching hides cross-period dependence
While local and global calibrations are equivalent, stitched laws miss multi-period risk; a new algorithm exposes the trade-off.
· “Global Multi-Maturity SPX-VIX Calibration Beyond Markovian Stitching”
Purely causal validation can't create a feedback singularity
Same-cycle settlement times optimized curvature decides when the multiplier hits unit gain; latency sets the deadline.
New financial QA benchmark achieves 79% accuracy with optimized RAG
FinRAG-QA tests 999 cross-bank questions on 10 indicators; optimized pipeline beats baselines by over 30 percentage points.
· “Enhancing Financial Question Answering: A Novel Benchmark Dataset of Banks' financial statements”
Entropic tracker cuts broken-asset weight 20-fold in a crash
A Fermi-Dirac entropy rule beat least squares through the COVID crash, with far lower turnover.
· “An Entropic Factor Model for Robust Portfolio Replication”
Market impact feedback cancels common predictable signals
In a mean-field model, when agents fully account for collective impact, the observed price reduces to a martingale, reconciling…
· “Mean-field equilibrium of heterogeneous agents under market impact”
Scaling laws hold for insurance pricing models
On a 4.5-million-row motor portfolio, TabM converts data into lower loss; transformer size alone adds little.
· “Scaling Laws, Tabular Data and Actuarial Ratemaking Models”
Dutch-book test reveals systematic incoherence in LLM forecasts
Arbitrage profit from betting against model probabilities grows with logical complexity and irrelevant context
A better reputation can make an AI agent likelier to cheat
When identities are cheap to reset, the value of betrayal can outgrow the value of honesty.
· “Tempting the Agent: The Economics of Reputation without Persistent Identity in AI Agent Markets”
Pairwise portfolio comparisons infer region-specific ESG preferences
GP-based preference elicitation combined with multi-objective RL captures how European and Texas investors trade off returns against…
· “Eliciting ESG Preferences for Reinforcement Learning-Based Portfolio Optimization”
Mixture model fixes integer-second bias in forex trade durations
The GA-ACD model outperforms standard models across seven major currency pairs by explicitly modeling heaping.
· “Modeling Trade Durations under Temporal Granularity Effects in Forex Markets”
LLM markets converge slower than human markets in double auctions
Three LLM populations fail to reach equilibrium; efficiency peaks at 91%, well below human benchmarks
Uniform coverage at spectral boundaries via projection
Projected intervals beat local approximation when eigenvalue collisions make delta methods unreliable
· “Uniform Inference and Certified Capacity at a Reflexive Stability Boundary”
Half-index mapping ties portfolio spells to order-flow memory
The core discovery is that under Assumption 3 (regularly varying tail of the execution-weighted residence scale R = κ/σ with index θ), the…
· “Switching Frictions, Heterogeneous Trading Horizons, and Long-Memory Order Flow”
Nonlinear impact before memory forces linearity
Applying a concave law after a power-law kernel is always safe; before it, only linear laws avoid price manipulation.
When jumps dominate drift, bubble solutions blow up in finite time
Global viscosity supersolutions for the Lévy-jump obstacle problem exist only when local drift exceeds a jump-intensity threshold…
· “Viscosity Supersolution Barriers to a Non-local Free Boundary Problem”
Dynamic CVaR hedging beats static on short maturities
A conditionally elicitable actor-critic RL agent reduces tail risk at 2–6 months for basket options, while static risk increases.
· “Insights on Time-consistent Deep Hedging under Elicitable Dynamic Risk Measures”
Ending extreme poverty: $304B a year
Targeted cash transfers using machine learning could reduce world poverty to 1% for a fraction of universal basic income cost.
Avoid full grid: adaptive method prices equity-linked contracts with surrender
Deterministic recursion with piecewise-cubic representation converges to Heston–CIR++ model, handles non-recombining fund
· “Adaptive singular-point method for pricing and hedging surrenderable equity-linked contracts”
One framework: AI is six economic forces at once
A unified analysis positions AI as capital, synthetic labor, human capital, infrastructure, and more, explaining its broad impact on…
· “AI and the Economy: An Economic Examination of Production, Distribution, Firms, Labor, and Welfare”
This paper tests whether selling put options on nuclear and energy stocks consistently…
A 25-year backtest of short-put option writing on 39 nuclear-related stocks reports a Sharpe ratio of 7.8 and zero losing months, but the…
Million-runner probit calibration drops to about 20 minutes
Pricing all n alternatives in one shared-field pass beats the standard GHK simulator's per-alternative cost.
Jumps are necessary to forecast extreme liquidity dry-ups
A new Illiquidity-at-Risk metric shows continuous models underestimate extreme dry-ups, and jump models correct the bias.
