Quant Letter
The week in quantitative finance research: 90 papers from arXiv, SSRN and RePEc, screened and ranked, 31 Oct – 6 Nov 2026.
This week spotlights credibility gaps in quantitative finance: LLMs lose predictive power in live trading despite strong backtests, knowledge cutoffs and future-look bias inflate strategy alphas, and reproducibility remains fragile. Standouts: LLM Stock-News Signals Under Scrutiny, LLM Knowledge Cutoffs and Backtests, and Bounded LLM Research Loop for Trading.
If you read five papers this week, read these.
A pre-registered blinded test finds that language-model readings of company news lose their predictive power in real time and that commercial models identify masked companies in two-thirds of cases.
The study estimates maximum Sharpe ratios in large cross-sections without constructing the efficient portfolio, finding annualized maximal Sharpe of 7.4 in U.S. equities versus benchmark estimates of 1.0 to 4.0.
Following AI adoption, financial analysts revise high-ESG forecasts upward and become more bullish, but forecast accuracy deteriorates and errors rise for top-ESG firms.
Topics drawing unusually many papers this week.
2.0% of this week's 441 new papers, 2.6x its share over the previous 6 weeks (0.8%).
2.0% of this week's 441 new papers, 1.8x its share over the previous 6 weeks (1.1%).
4.8% of this week's 441 new papers, 1.4x its share over the previous 6 weeks (3.3%).
What the quant community starred on GitHub this week, from a panel of 10,589 developers who follow financial machine learning (290 of the 858 we checked starred something). Counts only: we never publish who starred what.
Claude skills for day trading with backtesting and timing.
★ 5,704+2,004 this weekPythonNew repo
Real-time event intelligence and fraud detection for finance.
★ 231+233 this weekTypeScriptNew repo
QuantJourney backtester for testing trading strategies.
★ 597+426 this weekPython
QuantLib models with hidden Markov regime switching.
★ 11+3 this weekPythonNew repo
Local-first investment research system with 18 investor personas.
★ 681+32 this weekPython
Finance terminal for Bloomberg-style market data tracking.
★ 2,654+70 this weekTypeScript
Foundation model for the language of financial markets.
★ 47.0k+2,827 this weekPython
Multi-agent LLM framework for financial trading.
★ 121k+2,276 this weekPython
Framework for building semantic graph-based reasoning layers over data.
★ 8,496+7,839 this weekPython
CLI for AI agents to use computers.
★ 17.8k+15.2k this weekRust
Towards a more autonomous agent system.
★ 20.1k+20.1k this weekTypeScriptNew repo
Understand how any software works and craft your own.
★ 17.0k+17.0k this weekRustNew repo
Open-source bot with virtual machine for agent use.
★ 14.6k+5,650 this weekTypeScriptNew repo
AWS local emulator alternative.
★ 30.7k+3,363 this weekJava
Now published in Mathematical Finance · featured 20 Dec 2023, 6 days after release; 18 citations
Passed 25 citations · featured 24 Apr 2024, 5 days after release
Passed 10,000 citations · featured 5 Jun 2024
Filter by venue, topic or fanfare.
No papers match these filters.
Quantitative finance and ML-for-finance preprints
30 papers · 26 figures
A pre-registered blinded test finds that language-model readings of company news lose their predictive power in real time and that commercial models identify masked companies in two-thirds of cases.
Fanfare 4LLMs & TextAmir Dehkordi et al.PDF
A benchmark of 2,226 dated facts reveals that frontier language models underreport their knowledge and answer most facts they claim not to know, compromising backtest validity.
Fanfare 4LLMs & TextLingfei Qian et al.PDF
Question-based outlook scores extracted from analyst reports by an LLM improve prediction of target-price revisions beyond traditional sentiment measures.
Fanfare 2LLMs & TextKiyoun Kim and Sung-Yeon Nathan LimPDF
A code audit of the cryptocurrency trend factor reveals that after correcting for future-information use and technical-indicator issues, the factor's alpha becomes statistically insignificant and cannot replace momentum.
Fanfare 3Asset Pricing & FactorsYifei Wang and Zhenling PangPDF
A Gaussian-process factor model with nonlinear loadings estimated from firm characteristics improves stock ranking and portfolio Sharpe ratios over linear alternatives on S&P 500 data.
Fanfare 3Asset Pricing & FactorsAndrea Ruglioni and Damir FilipovicPDF
Machine learning models predict when cross-factor predictability will erode, achieving high C-index on held-out pairs and finding that stronger signals decay faster.
Fanfare 2Asset Pricing & FactorsChorok LeePDF
A two-step LASSO-based method selects variables and model specification jointly in high dimensions, with oracle properties; applied to characteristics-based pricing, it reveals linear and nonlinear effects and interactions with firm size.
Fanfare 2Asset Pricing & FactorsChaohua Dong et al.PDF
An open-source research loop constrains language-model proposals to strict JSON and enforces immutable data splits, discovery gates, and hash-chained logs to prevent false discoveries in backtest-driven strategy development.
Fanfare 3Trading, Microstructure & ExecutionHenry ZhangPDF
A protocol using e-processes and Ville arguments lets researchers monitor candidates born at data-dependent times and adaptively chosen, bounding false promotion rates without a fixed sample size.
Fanfare 3Trading, Microstructure & ExecutionEric Benhamou et al.PDF
The paper models learning traders on automated market makers, showing herds and arbitrage-dominated markets have identical statistics yet stress episodes differing unboundedly in length, and that maker-taker pricing restores stability.
Fanfare 3Trading, Microstructure & ExecutionIrene AldridgePDF
Decomposing cross-account order-flow persistence on decentralized exchanges, the study isolates two components and finds that on-chain herding mostly reflects arbitrageurs racing to centralized prices, not independent traders.
Fanfare 3Trading, Microstructure & ExecutionYichong Bai and Dingyue KanPDF
Using pseudonymous wallet addresses from on-chain order books as features, the study finds persistent top-trader markouts reaching 3.11 basis points at ten seconds and improving one-second prediction.
Fanfare 3Trading, Microstructure & ExecutionJingqiu Ding and Daojing ZhaiPDF
A tensor-native model for limit-order-book forecasting preserves the multi-way structure of market side, price level, and order attributes, outperforming flattened matrix approaches on benchmarks.
Fanfare 3Trading, Microstructure & ExecutionYuxuan Gu and Danilo MandicPDF
A deterministic market-making policy achieves logarithmic regret against fixed quotes and square-root regret against dynamic quotes when orders persist until execution.
Fanfare 2Trading, Microstructure & ExecutionNicolò Cesa-Bianchi et al.PDF
A graph-structured framework discovers alpha factors using an LLM, connecting factors by generation, similarity, and correlation while applying Pareto-guided search to balance quality, fertility and diversity.
Fanfare 3ML & AI MethodsZhan-He Wang et al.PDF
A graph framework that weights extracted relations by reliability improves stock ranking on the S&P 500 by 16.4% in average recall and 22.6% in success rate.
Fanfare 3ML & AI MethodsHao Zhou et al.PDF
A rough-memory VIX model with Mittag-Leffler decay improves term-structure fit and out-of-sample option pricing, with larger gains during elevated volatility and shorter maturities.
Fanfare 3Derivatives & VolatilityTianfeng Yang et al.PDF
Learning forecast-supervised weights for intraday volatility aggregation reduces forecasting error while preserving the structure of classical realized measures like raw realized variance.
Fanfare 2Derivatives & VolatilityBrian Sing Fan Chan and Jennifer So Kuen ChanPDF
A score-driven volatility model using neural networks to represent innovation density flexibly outperforms Gaussian and Student-t benchmarks, especially during high-volatility periods like COVID-19.
Fanfare 2Derivatives & VolatilityQianli Zhao et al.PDF
An iterated neural estimator for GARCH processes alternates between learning the recursion and updating conditional variances, achieving minimax-optimal rates for in-sample and forecasting tasks.
Fanfare 2Derivatives & VolatilityRuizhi Deng et al.PDF
Measuring minute-by-minute volatility response to preceding trends, the research finds asymmetry and evidence of one-sided temporal response in cryptocurrencies but not equities, decomposed via time-reversal symmetry.
Fanfare 2Derivatives & VolatilityGulsah Bicki and Taylan AkdoganPDF
The paper solves regime-switching models for affine and exponential-Lévy securities by reducing to one linear system, expanding the solution in mean holding time and pricing bonds, options, and credit instruments.
Fanfare 2Derivatives & VolatilityPeter CottonPDF
Option-surface repairs that satisfy no-arbitrage constraints admit multiple admissible solutions, and finite-difference amplification causes downstream Greeks and risk measures to diverge substantially.
Fanfare 2Derivatives & VolatilityTenghan ZhongPDF
A new Stata command implements panel local projections with split-panel jackknife correction, removing the Nickell bias that invalidates standard inference in dynamic models across regions or countries.
Fanfare 3Econometrics & ForecastingShu Shen et al.PDF
Panel VARs converge at the faster rate with N and T combined, while panel LPs converge only in T; the paper characterizes this efficiency-robustness trade-off and provides a test to choose between them.
Fanfare 3Econometrics & ForecastingMahmut Ablay et al.PDF
The study unifies identification results for structural shocks when variance paths differ, solving a generalized eigenvalue problem that nests prior work and clarifies what rank tests on second moments actually identify.
Fanfare 2Econometrics & ForecastingGeorge MilunovichPDF
A diagnostic backtesting framework evaluates deep research systems on four financial forecasting tasks across three markets, with controlled information cutoffs to prevent information leakage and assess whether historical backtests predict live performance.
Fanfare 2Econometrics & ForecastingFengyuan Liu et al.PDF
Aggregate forecast-interval coverage can hide state-specific failures when economic regimes return after long absence; recent-only calibration undercovers during high volatility despite meeting overall targets.
Fanfare 2Econometrics & ForecastingRuihang Han and Li-Hsiang LinPDF
When estimating expected shortfall from overlapping multi-day returns, the mismatch with standard independent-sampling theory causes systematic downward bias that weakens capital coverage, especially under heavy tails.
Fanfare 3Risk, Credit & BankingMartin Aichele et al.PDF
A post-model layer that reduces portfolio concentration by penalizing shared exposures cuts maximum drawdown from 23.6% to 14.1% on a cross-asset test.
Fanfare 3Portfolio & AllocationEric Benhamou et al.PDF
New working papers in finance, economics and ML
30 papers · 1 figures
The study estimates maximum Sharpe ratios in large cross-sections without constructing the efficient portfolio, finding annualized maximal Sharpe of 7.4 in U.S. equities versus benchmark estimates of 1.0 to 4.0.
Fanfare 4Portfolio & AllocationMikhail Chernov et al.
A convex optimization method combines risk budgeting with expected returns, delivering higher average returns than risk parity with lower turnover and concentration than traditional mean-variance portfolios.
Fanfare 3Portfolio & AllocationRaul Riva et al.
Testing 459 quantitative-finance machine-learning papers' released code, the study finds that execution success rates fall steeply due to environment and dependency issues, not link rot alone.
Fanfare 4ML & AI MethodsAkram Khan
Large language models recall Indian firm financials from within their training data with accuracy that declines sharply for smaller firms, creating serious look-ahead bias in LLM-based backtests.
Fanfare 3ML & AI MethodsSethu Madhavan Chandavarapu
Using language models as synthetic investors reveals institutional portfolios tilt greener despite weaker green tastes, and beliefs about green returns turn pessimistic after 2020.
Fanfare 3ML & AI MethodsXindi He and Lubos Pastor
Headlines scored by LLM within minutes predict cross-sectional cryptocurrency returns with a daily long-short Sharpe of 2.21 until the signal decays.
Fanfare 3Crypto & DeFiALEKSANDR SENKEVICH
NFT airdrops on Ethereum raise block fees and failure rates, shifting price discovery off-chain and impeding native-token information incorporation critical for proof-of-stake security.
Fanfare 3Crypto & DeFiAgostino Capponi et al.
Memecoin traders get back 0.754 per unit staked, worse than any casino game, while perpetual futures and prediction markets match casino returns.
Fanfare 2Crypto & DeFiDaniel Gatto
A gradient-boosted classifier trained on 219 USD-pegged stablecoins predicts which coins will lose their peg within one week, achieving 0.59 average precision on test data.
Fanfare 2Crypto & DeFiJude Enajero
Fine-tuned LLM analysis of hedge fund letters reveals policy expectations that predict yield-curve moves beyond market rates and correlate with fund returns.
Fanfare 3Macro-Finance & RatesPatrick Coggi et al.
Using equity and risk-free rate term structures, monetary policy shocks affect both expected dividend growth and equity risk premium, with effects beyond duration matching.
Fanfare 3Macro-Finance & RatesJens Carsten Jackwerth and Benjamin Golez
Models how wealth redistribution among arbitrageurs with different beliefs amplifies exchange rate volatility and generates interest-parity wedges, accounting for 62 percent of observed variance.
Fanfare 3Macro-Finance & RatesRikuto Onishi
Research shows trade exposure determines whether commodity-price rises reduce or increase sovereign risk, with exporters gaining from higher prices and importers facing reinforced costs.
Fanfare 3Macro-Finance & RatesNaomi Cohen et al.
Models how central banks must retain more reserves for rate control when interbank trading cannot redistribute reserves across heterogeneous banks, even if aggregate demand is sufficient.
Fanfare 2Macro-Finance & RatesJunko Oguri and Cristoforo Pizzimenti
The classic TSMOM long-short premium has decayed to a zero net Sharpe, but a long-only volatility-targeted version across multiple horizons earns 0.58 with strong crisis protection.
Fanfare 3Asset Pricing & FactorsXianghui Fang
Measuring data investment from earnings calls, the research shows data capital's share of investment reached about one-eighth in 2025 and commands a market premium of 12 dollars per dollar invested.
Fanfare 3Asset Pricing & FactorsZhenkai Ran et al.
Shows that stock returns respond more to the same percentage surprise when reported earnings are larger in cents, consistent with anchoring on nominal rather than percentage earnings beats.
Fanfare 3Asset Pricing & FactorsPetrus H. Ferreira et al.
A 2021 life-insurer capital rule change shows that when incumbent creditors' capacity shrinks, corporate debt issuance falls and firms struggle to refinance maturing debt.
Fanfare 3Risk, Credit & BankingZiHao Ji et al.
Analysis of 19.8 million securitized auto loans reveals that reported data systematically understate risk through selection bias, carry-over reporting gaps, and priced-at-origination hidden credit differences.
Fanfare 3Risk, Credit & BankingPekka Honkanen et al.
Private credit funds that borrow from banks transmit funding shocks to firm lending; following SVB's collapse, exposed banks increased credit-line commitments while cutting direct lending to nonfinancial firms.
Fanfare 3Risk, Credit & BankingManasa Gopal et al.
The study identifies two types of squeezes in corporate bonds that degrade price discovery and returns, with lender squeezes predicting negative short-horizon returns and market squeezes predicting positive ones.
Fanfare 3Risk, Credit & BankingAngel Tengulov et al.
Finds that authorized participants create corporate bond ETF shares at discounts to offload overvalued inventory, causing delivered bonds to underperform by 5.3 basis points.
Fanfare 3Risk, Credit & BankingYuan Zhou
Analyzing 1.68 million stock-day events from 4.57 million articles, the study finds that directional price moves concentrate before and at publication day, with little incremental return later.
Fanfare 3Trading, Microstructure & ExecutionAlireza Kargarzadeh et al.
Using machine learning to map firm disclosures to congressional committee jurisdictions, finds no evidence that members earn abnormal returns from legislative access or committee hearings.
Fanfare 3Trading, Microstructure & ExecutionYuan Zhou
A likelihood-ratio framework ranks wallets by their information content in binary prediction markets, separating ranking from multiple testing and validating the method through simulation.
Fanfare 2Trading, Microstructure & ExecutionAhura Azadi
Studying 244 resolved Kalshi markets, the research finds that efficiency is measurable ex post but not predictable ex ante, and naive trading strategies fail against market benchmarks.
Fanfare 2Trading, Microstructure & ExecutionCristina Bixhili
A neural network that analytically enforces the Gibson-Schwartz PDE recovers commodity convenience yield from futures and avoided implausible estimates during the April 2020 WTI dislocation.
Fanfare 2Derivatives & VolatilityEvan Lynch
Tests term-structure and skew trading signals in equity volatility from free public data, showing term-structure carry lifts Sharpe ratio to 2.94 but skew lacks sufficient historical data.
Fanfare 2Derivatives & VolatilityYasanji Ratnaike
Following AI adoption, financial analysts revise high-ESG forecasts upward and become more bullish, but forecast accuracy deteriorates and errors rise for top-ESG firms.
Fanfare 4LLMs & TextTong Li et al.
Cross-sectional IC volatility has a floor from factor risk that does not shrink with breadth, capping effective sample size at roughly fifty observations.
Fanfare 4Econometrics & ForecastingHimal Naidu
Working papers curated by RePEc's NEP field reports
30 papers · 23 figures
Wall Street Journal market summaries explain 30-40% of daily and monthly returns out of sample when filtered for stated reasons, compared to under 10% for macro variables.
Fanfare 4ML & AI MethodsBryan T. Kelly et al.PDF
Facial and textual emotions from Fed Chair press conferences generate immediate effects on foreign exchange and equity returns, with non-additive cross-modal interactions.
Fanfare 3ML & AI MethodsYiru Chen et al.PDF
A hybrid FinBERT-HAR-X framework using multilingual news reduces S&P 500 volatility forecast error by 21% and improves realized Sharpe ratios out of sample.
Fanfare 3ML & AI MethodsSara Boughanou and Adil El MarhoumPDF
Using LLM personas, disagreement over news headlines predicts trading volume, with income and education disagreement most predictive and fundamentals generating least disagreement.
Fanfare 3ML & AI MethodsVineet Bhagwat et al.PDF
AI-scenario contributions to financial-sector tail risk vary substantially by country, with downside asymmetry in Europe and the UK but limited impact in North America.
Fanfare 2ML & AI MethodsJavier Ojea Ferreiro et al.PDF
HiDVAR, a data-rich distributional framework, improves inflation tail-risk assessment by 28% in CRPS relative to standard benchmarks and performs well in real-time evaluation.
Fanfare 4Econometrics & ForecastingLuca SchmitzPDF
Survey of 170,000 consumers shows that attention to monetary policy is driven by incentives; those planning major purchases are significantly more attentive to interest rate news.
Fanfare 3Econometrics & ForecastingIna Hajdini et al.PDF
A model allowing expectations to shift to a low-inflation equilibrium cuts forecast errors roughly 50 percent versus central bank projections for euro area core inflation in the 2010s.
Fanfare 3Econometrics & ForecastingAlicia Pita Marcet et al.PDF
The study develops nonparametric local projections to identify how economic shocks affect outcomes differently across states, uncovering a hump-shaped investment response to monetary policy that linear methods understate.
Fanfare 3Econometrics & ForecastingJoel M. David et al.PDF
Large oil discoveries with production lags of five years show that anticipated future output raises core inflation a decade ahead, with effects concentrated in nontradables under fixed exchange rates.
Fanfare 3Econometrics & ForecastingBohou Cai et al.PDF
Lower-income households form inflation beliefs two and a half times more sensitive to their emotional state than higher-income households, affecting expectations substantially independent of fundamentals.
Fanfare 3Econometrics & ForecastingShahdah Hanif and Wali UllahPDF
A news-based stress index for Canada reveals that perceived financial stress operates through household precautionary behavior, independently of market-based financial shocks.
Fanfare 2Econometrics & ForecastingBrett Barnes et al.PDF
Model shows that stablecoin design cannot remove peripheral currency risk but redeploys it; a new Triffin dilemma emerges where the issuer's capital must grow with its open position.
Fanfare 4Crypto & DeFiSilvia SgherriPDF
New U.S. legislation grants stablecoins superpriority in bankruptcy, allowing distressed issuers to subordinate existing creditors and shift losses to depositors before any filing occurs.
Fanfare 2Crypto & DeFiJeffery Y. Zhang and Peter ZimmermanPDF
U.S. pension funds facing negative cash flows sell equities as primary liquidity source rather than rebalancing, absorbing $0.67 per dollar of shock through equity sales.
Fanfare 3Portfolio & AllocationAleksandar Andonov et al.PDF
Literature review finds genuine skill in private equity performance persistence, especially in venture capital, but shows estimates often weaken under stricter methods and richer data.
Fanfare 2Portfolio & AllocationOylum Sehvez ErguzelPDF
Using policy-rule discontinuities, the study finds BOJ stock purchases raise long-term rates but lift prices only after yield curve control pegged rates at zero.
Fanfare 3Asset Pricing & FactorsMasao Fukui and Masayuki YagasakiPDF
A new uncertainty index for commercial real estate predicts prices and construction spending and accounts for 46 percent of spending forecast errors, outperforming existing uncertainty measures.
Fanfare 2Asset Pricing & FactorsVanessa Dietl and Gabriel LeePDF
Tighter statutory constraints on central bank lending to government reduce real borrowing costs in developing economies while raising debt-to-GDP, with effects strongest where credibility gains matter most.
Fanfare 3Risk, Credit & BankingAngelos Athanasopoulos et al.PDF
Analysis of 300,000 monthly bond prices since 1822 shows wars lower external bond returns by five percentage points and raise default risk, while military threats depress threatened states' borrowing costs.
Fanfare 3Risk, Credit & BankingJonathan Federle et al.PDF
A sovereign default model shows that high borrowing costs and debt overhang depress public adaptation spending, yet adaptation yields the largest welfare gains when financing is tightest near default.
Fanfare 3Risk, Credit & BankingCarlos Bolivar et al.PDF
Analysis of 17 domestic debt restructurings reveals that monetary transmission often fails to recover post-restructuring due to balance-sheet constraints across sovereigns, central banks, and financial institutions.
Fanfare 2Risk, Credit & BankingUdaibir S. Das and David A. GrigorianPDF
The liquidity coverage ratio significantly boosts private credit growth in counties with more LCR-subject banks, an effect comparable in magnitude to 2012 capital requirement increases.
Fanfare 2Risk, Credit & BankingSebastian Doerr et al.PDF
Analysis of COPOM announcements reveals target and path surprises move the yield curve differently, with nonlinear effects and sign asymmetries in inflation expectation responses.
Fanfare 2Macro-Finance & RatesFrederico Bastone and Mauro Sayar FerreiraPDF
Adding financial conditions to macro forecasts raises recession tail probabilities and lowers growth quantiles, though out-of-sample forecast improvement is inconsistent.
Fanfare 2Macro-Finance & RatesLarissa NawoPDF
The study estimates the full distribution of firms' investment responses to monetary policy using clustering regression, finding only 5 percent of firm-time observations show strong sensitivity.
Fanfare 2Macro-Finance & RatesThomas Drechsel et al.PDF
Analytical framework shows how debt size, composition, and fiscal rules drive three channels—debt service, debt composition, and fiscal cyclicality—that amplify or dampen monetary policy effects.
Fanfare 2Macro-Finance & RatesS. LavenderPDF
Rate increases substantially reduce house prices and mortgage purchases, with housing starts and permits falling slowly, suggesting demand responds more sharply than supply.
Fanfare 2Macro-Finance & RatesStéphane SurprenantPDF
The research finds that euro-area banks transmit interest rate changes more slowly to lending rates in the fourth quarter during tightening, consistent with year-end volume targets.
Fanfare 2Macro-Finance & RatesZuzana Fungáčová et al.PDF
Analysis of quoted prices reveals that banks charge firms 2-3.5% for cross-border payments, with foreign exchange margins dominating costs and substantial price dispersion across banks for identical corridors.
Fanfare 2Trading, Microstructure & ExecutionJiaqi LiPDF
