Quant Letter: May 2025, Week-2

Weekly (91st Edition)

ArXiv

Finance

FlowHFT: High-Frequency Trading: FlowHFT, a new imitation learning framework, outperforms traditional high-frequency trading models by learning from multiple expert models. (2025-05-09, shares: 26)

Loss-Rebalancing in Blockchains: A study reveals that constant block intervals in blockchain settings provide the best protection against arbitrage for Automated Market Makers' liquidity providers, using random walk theory. (2025-05-08, shares: 47)

LLM Investing Strategies Evaluation: FINSABER, a backtesting framework, shows that Large Language Models' effectiveness in stock trading decreases over longer periods and larger symbol universes, emphasizing the need for trend detection and risk controls. (2025-05-11, shares: 19)

Bias of Resampled Backtests: A study finds that resampling techniques in backtests can cause bias in Sharpe Ratio estimates, suggesting a need for structure-preserving resampling methods. (2025-05-09, shares: 19)

NewsNet-SDF: Discount Factor Estimation: NewsNet-SDF, a deep learning framework, successfully integrates pretrained language model embeddings with financial time series for asset pricing and risk assessment, outperforming traditional models. (2025-05-11, shares: 18)

Deep PDE Solvers: The study compares the effectiveness of two deep learning algorithms, the Deep Galerkin Method and the Time Deep Gradient Flow method, in solving PDEs for option pricing. (2025-05-08, shares: 16)

Excess Volatility Revisited: The Chiarella model is expanded to handle long-term value drifts, revealing excess volatility and bimodal mispricings in various assets. (2025-05-12, shares: 15)

Multi-Asset Portfolio Selection: The use of the soft actor-critic (SAC) algorithm in multi-asset portfolio selection is explored, showing superior performance in both simulated and real markets. (2025-05-12, shares: 15)

Grey Extension: The rough Bergomi model is extended by replacing the fractional Brownian motion with a generalised grey Brownian motion, improving SPX/VIX options calibration. (2025-05-13, shares: 14)

Tariff Wars Impact: The paper introduces an algorithm for creating fair trade relations based on a world trade model, aiming to prevent tariff wars and discussing the impact of tariff restrictions on trade prices and recession. (2025-05-08, shares: 13)

Economics

LLMs in Market Experiments: Large Language Models (LLMs) have potential in mimicking human behavior in economic markets, but need more research for improved diversity and accuracy. (2025-05-12, shares: 21)

Signature Decomposition for Trading: A new pairs trading strategy using path signature techniques enhances futures trading by providing better interpretability, robustness, and returns. (2025-05-08, shares: 17)

Inflation's Impact on Unemployment: Inflation driven by supply increases unemployment risks for vulnerable workers, emphasizing the need for inflation control for fair labor market growth. (2025-05-09, shares: 16)

GenAI Effects on Entrepreneurship: A literature review identifies five key themes in the impact of Generative AI on entrepreneurship, calling for more broad-scale research and effective regulations. (2025-05-08, shares: 16)

Coffee Value Chain in Davao del Sur: Research on Davao del Sur's coffee value chain dynamics reveals that trust, risk, and transaction costs influence key interactions, and a balanced system benefiting all can be achieved with adjustments to demand, pricing, trust, and risk attitudes. (2025-05-09, shares: 15)

Optimal Market Choices for Coffee Farmers: A study on Sultan Kudarat's coffee farmers' profits shows positive annual profit from selling all dried cherries, but negative profit from producing and selling all green coffee beans due to extra costs and minimal price differences. (2025-05-09, shares: 15)

Smart Meter Adoption for Energy Efficiency: A study on smart meter adoption in the US shows that federal funding and state legislative actions positively impact adoption and energy efficiency savings, emphasizing the need for multi-level governance in policy integration. (2025-05-08, shares: 14)

Big Data in Entrepreneurship: The chapter highlights the potential of using large-scale data in entrepreneurship and innovation research, suggesting that machine-learning models and big data can create precision measurements and 'digital doubles' for virtual experimentation. (2025-05-13, shares: 13)

Miscellaneous

Efficient Leverage Estimator: The article introduces a new method for estimating the leverage effect using high-frequency data, offering two new estimators that simplify the process and increase noise resistance, as confirmed by simulations and empirical analyses. (2025-05-13, shares: 16)

Transfer Learning for Curves: The paper proposes a framework for transferring learning of discount curves across different fixed-income product classes, enhancing kernel ridge regression and introducing a term that promotes curve smoothness, resulting in improved extrapolation performance and tighter confidence intervals. (2025-05-12, shares: 15)

Economic Facts by LLMs: The research investigates the use of hidden states in large language models to estimate and fill in economic and financial statistics, showing that a simple linear model trained on these hidden states performs better than the models' text outputs and requires minimal labelled examples for training. (2025-05-13, shares: 11)

Crypto & Blockchain

Unified Asymptotic Theory for ACD Models: The research applies a unified asymptotic theory to high-frequency cryptocurrency ETF trading data, revealing infinite-mean durations for all five cryptocurrencies studied. (2025-05-09, shares: 18)

DeFi Liquidation Risk Modeling with Reflection Principle: The article introduces an analytical method for calculating the collateral liquidation probability in DeFi stablecoin single-collateral lending, improving risk assessment efficiency. (2025-05-12, shares: 14)

Efficient Risk Assessment in DeFi Platforms: The paper suggests a lightweight, exact solution for computing the collateral liquidation probability in DeFi stablecoin single-collateral lending, offering a more efficient alternative to complex simulations. (2025-05-12, shares: 14)

Historical Trending

Risksensitive RL with Convex Scoring Functions: The article presents a reinforcement learning framework for managing risk objectives. This is achieved through a specialized Actor-Critic algorithm and an auxiliary variable sampling method. The effectiveness of this approach is confirmed through simulation experiments in statistical arbitrage trading. (2025-05-07, shares: 19)

SSRN

Recently Published

Quantitative

Machine Learning for GDP Forecasting: The study finds that machine learning models can effectively forecast U.S. GDP, especially during economic volatility. (2025-05-12, shares: 2.0)

Offshore Wind Power Prediction: The Informer, a new deep learning algorithm, is used for ultra-short-term offshore wind power prediction, effectively extracting features and capturing sequence dependency from long time-series data. (2025-05-14, shares: 2.0)

Incremental Category Discovery: A Decoupled Likelihood Modeling framework is proposed to address class imbalance in Incremental Generalized Category Discovery, showing strong scalability in large-scale class settings. (2025-05-10, shares: 2.0)

Multi-Market Coupling Model: The paper presents a model that predicts day-ahead electricity prices in Central Western Europe by incorporating Flow-Based Market Coupling into a residual demand framework. (2025-05-11, shares: 3.0)

Hierarchical Risk Clustering vs Portfolios: The paper warns that hierarchical risk clustering strategies in portfolio allocation can be affected by inaccuracies in the covariance matrix. (2025-05-08, shares: 2.0)

Financial

Mutual Fund Model: A new model for assessing global mutual funds' financial performance has been validated using data from 35 countries over 34 years, considering factors like risk size, diversification, and liquidity. (2025-05-08, shares: 4.0)

Interest Rates vs Stock Returns: A study reveals a strong negative correlation between expected inflation sensitivity and firm growth, with low inflation sensitivity firms experiencing high, sustained growth. (2025-05-12, shares: 3.0)

Risk Sharing with Recursive Utility: A proposed model allows optimal risk sharing in dynamic settings with diverse preferences, introducing a new traded security as an endogenous variable. (2025-05-09, shares: 3.0)

Ambiguity in Insurance: Price movements in catastrophe bonds can be predicted by ambiguity preference in economic outlook and natural disasters, especially during crises and geopolitical conflicts. (2025-05-11, shares: 3.0)

Recently Updated

Quantitative

Economic Crises Prediction with ML: The research uses machine learning to predict national financial crises, with the Balanced Random Forest model proving most effective. (2025-04-30, shares: 7.0)

ML for Option Exercise Prediction: The paper uses machine learning to predict the exercise of American call options, outperforming traditional methods. (2025-04-25, shares: 3.0)

ML in Wealth Management: The study discusses the transformation of the wealth management industry through machine learning and cloud computing, improving competitiveness and efficiency. (2021-12-17, shares: 2.0)

SSR Calculation in Quadratic Rough Heston: The quadratic rough Heston model, improved with Gatheral 2022's hybrid scheme and finite difference methodology, offers reliable skew-stickiness ratio values and fits to SPX and VIX volatility smiles. (2025-05-02, shares: 3.0)

RealTime PV Power Forecasting: The XGBoost model, using historical weather and PV output data, offers more precise ultrashort-term PV power predictions than the SVR model, contributing to grid stability. (2025-03-01, shares: 3.0)

AI for Specialty Insurance Analytics: The creation of AI solutions for specific insurance predictions is vital due to industry consolidation and the rise of InsurTech startups, with gaps in domain knowledge and machine learning guidance. (2023-12-25, shares: 2.0)

Short-Term Rental Price Forecasting: A custom artificial neural network model using data from Airbnb, AirDNA, and the Federal Reserve accurately predicts short-term rental prices, highlighting key features like guest capacity and review count. (2025-05-06, shares: 3.0)

Financial

Model Complexity in Asset Pricing: Complex algorithms in global models outperform regional models in cross-sectional asset pricing, contradicting previous studies favoring regional methods. (2025-05-05, shares: 5.0)

Volatility in Continuous Trading: Cryptocurrencies, unlike stock markets, show a flat volatility profile due to the lack of restricted trading hours. (2025-05-05, shares: 3.0)

Numerical Finance Methods: The article analyzes advanced numerical techniques in finance, discussing their uses, limitations, and challenges like computational complexity and model risk. (2025-05-02, shares: 3.0)

AGNOSTIC: Quantitative Finance: The AGNOSTIC tool, free from assumptions or parametric models, is designed to help tackle dimensionality and overfitting issues in Quantitative Finance. (2025-05-05, shares: 4.0)

Share Repurchase and Corporate Investments: Legalizing stock repurchases enhances equity capital access and investment, indicating that buyback restrictions could hinder efficient capital distribution. (2025-05-04, shares: 3.0)

Expected Returns in Stock Performance: A few stocks significantly impact the performance of cross-sectional asset pricing anomalies, implying that a large part of the returns may be due to mispricing. (2025-05-06, shares: 3.0)

Municipal Bond Auctions and Trading: In municipal bond auctions, dealers bid higher when expecting to acquire new clients, resulting in lower winning yields and broader trading networks. (2025-05-01, shares: 3.0)

Network Basket Loan: The article introduces the Network Basket Loan, a financial tool aimed at funding socially beneficial projects and preventing market failures related to social and partial selection. (2025-04-20, shares: 3.0)

REIT Return Analysis: The paper presents and analyzes six REIT return factors, demonstrating that these factors outperform general equity asset pricing factors and behave differently across economic conditions. (2025-04-09, shares: 2.0)

Unit Roots in G7 Countries: The study reassesses the stability of 40 macroeconomic variables across the G7 economies, finding strong evidence of mean reversion in the U.S. and significant variations between countries. (2025-05-01, shares: 3.0)

Neural Network in Asset Prediction: The paper uses Fourier series expansion on the average asset return function, shedding light on the equity premium puzzle and demonstrating how alpha represents the long-term historical dependence of covariance risk. (2025-05-01, shares: 2.0)

Order Flows and Stock Prices: The research analyzes a large external asset purchase program, discovering that stock prices can be influenced by external order flows even if they don't provide information about future cash flows. (2025-05-01, shares: 2.0)

Venue Participation in China Bonds Market: The article explores bond trading activity and transaction cost differences between Over-the-Counter and Central Limit Order Book venues in China's interbank government bonds market, showing how market structure and trading mechanism impact asset liquidity. (2025-05-01, shares: 2.0)

RePec

Finance

Automated Trading System for EM Portfolios: A new Automated Adaptive Trading System may help stabilize emerging markets during downturns, addressing issues caused by the rise of algorithmic trading and passive investing. (2025-05-14, shares: 27.0)

Dynamic Correlations in Risk Parity Portfolio Optimization: Accounting for fat-tailed returns in risk parity portfolio optimization can lessen sensitivity to volatility shocks, decrease portfolio turnover during market instability, and enhance risk-adjusted returns. (2025-05-14, shares: 16.0)

Adaptive Market Hypothesis & Sharpe Ratio Strategies: The research finds that Sharpe Ratio Minimae and Maximae trading strategies are more profitable than the buy-and-hold strategy in global markets, supporting the Adaptive Market Hypothesis. (2025-05-14, shares: 15.0)

Statistical

BRM for Predictions: The blockwise reduced modeling (BRM) method is introduced for analyzing incomplete data sets, improving predictive performance for both linear and nonlinear models. (2025-05-14, shares: 20.0)

New Momentum Strategy: A new machine learning strategy, momentum-determined indicator-switching (N-MDIS), is proposed to increase the accuracy of equity premium prediction, outperforming existing strategies. (2025-05-14, shares: 19.0)

News Sentiment Impact: A reevaluation of the role of news sentiment shows that accurately measured sentiment significantly impacts intraday stock return volatility. (2025-05-14, shares: 16.0)

Stochastic ML Approach: A version of Stochastic Gradient Boosting is suggested for estimating production possibility sets in Data Envelopment Analysis (DEA), reducing overfitting and aiding in scenarios requiring generalization. (2025-05-14, shares: 16.0)

GitHub

Finance

Paper2Code: Automating Code Generation: Paper2Code is a new tool that can automatically generate code from machine learning scientific papers. (2025-04-25, shares: 1572.0)

EasyStreamingIndicators: Python Library for Analysis: The article introduces a new Python library that can calculate technical analysis indicators on live data. (2023-05-12, shares: 115.0)

ArkFlow: Rust Stream Processing Engine: The article discusses a high-performance data stream processing engine built in Rust, supporting various data sources and processors. (2025-03-01, shares: 896.0)

Trending

ScriptsNotebooks for Medium: The article shares a collection of scripts and notebooks used in the author's Medium posts. (2018-10-28, shares: 467.0)

RealWorld ML System Design: The article reviews 500 case studies of real-world machine learning system designs from 100 companies, with a focus on GenAI. (2025-03-08, shares: 868.0)

Horizontal Scaling PostgreSQL: The article explains how to horizontally scale PostgreSQL using automatic sharding. (2024-12-27, shares: 1362.0)

ContextGem: LLM Extraction: The article presents ContextGem, a tool designed to easily extract LLM from documents. (2024-10-11, shares: 305.0)

Microsoft Playwright: Web Testing: The article introduces Playwright, a framework for web testing and automation that supports Chromium, Firefox, and WebKit through a single API. (2019-11-15, shares: 72492.0)