Optimal Execution Liquidity
Research shows Double Deep Q-learning, a Reinforcement Learning technique, can effectively learn optimal trading strategies in fluctuating liquidity conditions.
5 shares11 citations todaySource ↗
Quant LetterNo. 38
February 2024, Week 3
96 items across 9 sections, as sent to readers on 21 February 2024. Paper titles open their ML-Quant page; ↗ goes to the source.
- Items
- 96
- Papers
- 64
- Sections
- 9
- Issue
- No. 38
arXiv
Quantitative-finance and ML-for-finance preprints from arXiv.
14 items
Finance5
Price of Information Asset Prices
A study finds that investors decide to buy additional information about an asset's trajectory at a specific time, based on the indifference price of information.
5 sharesSource ↗
RAGIC: Risk-Aware Stock Prediction
Risk-Aware Stock Prediction: The RAGIC model, using a Generative Adversarial Network, accurately predicts future stock prices with a consistent 95% coverage.
4 sharesSource ↗
Stackelberg Reinsurance MV Criterion
A study on reinsurance Stackelberg game suggests a single, one-time reinsurance contract is more beneficial than continuous or multiple discrete-time contracts.
3 shares8 citations todaySource ↗
Interbank Network Risk Analysis
A new method for reconstructing financial networks can enforce desired sparsity and link reciprocity, enhancing the prediction of various network properties.
3 shares1 citation todaySource ↗
Crypto & Blockchain4
HighFrequency Bitcoin Price Analysis
An analysis of the Bitcoin market index from 2019 to 2022 reveals two periods of volatility, suggesting greater market efficiency at shorter time scales.
4 shares3 citations todaySource ↗
MARL Model for Crypto Market Simulation
A multi-agent reinforcement learning model simulates crypto markets using Binance's daily closing prices of 153 cryptocurrencies from 2018 to 2022, accurately emulating crypto market microstructure.
4 shares5 citations todaySource ↗
Blockchain Anomalies and Fraud Detection
A paper examines key definitions and properties of blockchain, analyzes anomalies and frauds that threaten these networks, and proposes detection and prevention strategies.
2 shares12 citations todaySource ↗
Ethereum 2.0 Reward Dynamics Analysis
A study of consensus reward data from the Ethereum Beacon chain offers insights into reward distribution and evolution, aiding in the assessment and refinement of blockchain systems' decentralization, security, and efficiency.
2 shares25 citations todaySource ↗
Historical Trending5
Optimal Retirement Strategies with Jump Risks
A study suggests that immediate retirement is the optimal decision when de facto wealth surpasses a certain percentage of wage.
74 shares5 citations todaySource ↗
Automated Trading on Energy Markets
Automated trading in the day-ahead energy market using reinforcement learning algorithms yields the highest profits, according to a study.
36 sharesSource ↗
Lambert Monte Carlo
The article introduces a new method using the Lambert function to evaluate reservation price in illiquid markets, improving accuracy and aiding in hedging asset selection.
31 sharesSource ↗
Efficient Pricing under Heston Model
The authors suggest a data-driven method using artificial neural networks for efficient pricing of certain options, reducing computational time and increasing accuracy.
22 shares3 citations todaySource ↗
Correlation in Upstreamness
The paper investigates the correlation between upstreamness and downstreamness in global value chains, attributing the observed correlation to structural constraints rather than economic trends.
22 shares9 citations todaySource ↗
SSRN
Working papers in finance and economics from SSRN.
27 items
Quantitative11
Algorithmic Stock Trading with Ensemble Deep Neural Networks
The study proposes an ensemble network that uses various graph representations and time series data for financial forecasting and investment strategy formulation.
2 sharesSource ↗
Dynamic Tail Dependence Among Assets
The research introduces a new model for better understanding and managing the risk associated with multiple asset classes, improving upon previous models.
8 sharesSource ↗
Hybrid Model for Cryptocurrency Forecasting
A new forecasting model that combines Discrete Wavelet Transform and Long Short-Term Memory network is presented to predict cryptocurrency returns, using specific economic indicators.
16 sharesSource ↗
Corporate Fraud Detection with Linguistic Readability
The study uses Word2Vec, a natural language processing method, to create new readability indexes for Chinese texts, improving the detection of corporate fraud.
4 sharesSource ↗
Upper Bound for Options on Realised Volatility
The study presents a new formula for calculating the price of volatility swaps in uncorrelated stochastic volatility models, providing an upper limit for options on realized volatility.
3 sharesSource ↗
HARX Volatility Model for Asset Price Forecasting
The study finds that low-frequency HARX models using public data can accurately predict asset price volatility, matching the original HAR model's accuracy.
2 shares1 citation todaySource ↗
Investment Dynamics
The article challenges the Q theory of investment, arguing that supply shocks can also influence investment dynamics, not just demand.
2 sharesSource ↗
Hedge Fund Strategy
The research uses machine learning to analyze hedge fund strategies, concluding that most do not align with their reported performance.
3 sharesSource ↗
Credit Portfolio Modeling
The paper presents a simulation tool for assessing credit portfolio risks and CDO strategies, highlighting the role of quantitative methods and machine learning in financial risk evaluation.
3 sharesSource ↗
FRTB Impact on Market Risk
The study examines the impact of the Basel Fundamental Review of the Trading Book on banks' capital requirements, predicting significant increases in regulatory capital.
2 sharesSource ↗
Jump Components in Jump Diffusion Models
A proposed nonparametric test can determine if a jump diffusion process contains a jump component or is a diffusion, with the test statistic showing standard normal distribution if there are no jumps.
2 sharesSource ↗
Financial16
Profitable Day Trading Strategy
Research indicates that day trading, particularly on Stocks in Play, can provide a steady income, with a top 20 Stocks in Play portfolio achieving over 1600 net performance and a Sharpe ratio of 2.
12 shares3 citations todaySource ↗
Optimal Option Market Making
A novel market making model for options trading has been introduced, considering trader's volatility views and incorporating features like trading position limit, risk control, and simultaneous market making of multiple options.
3 shares1 citation todaySource ↗
News Sentiment in Salmon Price
Deep learning models and sentiment analysis can accurately predict salmon spot prices, with sentiment scores reducing prediction errors.
3 sharesSource ↗
Direct Reinforcement Learning
A new online portfolio decision model combines the multifactor model and mean-variance portfolio optimization in one step, enhancing overall performance.
2 sharesSource ↗
Synthetic Beta Determination
The article proposes a new method to estimate the beta coefficient in investment projects using a simulation model, allowing for the calculation of market beta even when it's unobservable.
2 sharesSource ↗
Wheat Futures Forecasting
The article uses a convolutional neural network to predict futures prices by analyzing aerial images of wheat fields and cloud cover, suggesting that unique algorithm and data choice can yield positive alpha in a short time frame.
2 sharesSource ↗
Asset Pricing Test
The paper introduces a statistical test to identify sparsity in high-dimensional factor models, concluding that less than ten factors can explain stock returns and dense models perform better than sparse ones.
4 shares3 citations todaySource ↗
Interest Rates Stochastic Volatility Model
The lognormal stochastic volatility model is introduced in the single-factor Cheyette model for interest rate dynamics, demonstrating robustness and accuracy in fitting market implied volatilities.
1,086 sharesSource ↗
Cryptocurrency Inverse Options Valuation
A study on inverse options, common in crypto exchanges, reveals that they can be equivalent to standard vanilla options under certain circumstances.
844 sharesSource ↗
ETF Tracking Errors Machine Learning
Machine learning methods, specifically Random Forest and Gradient Boosting Decision Tree, are found to be more effective in predicting U.S. ETF’s tracking errors, with U.S. assets and expense ratio being key factors.
14 sharesSource ↗
Commodity Hedging Futures Markets Speculation
A comparison of traditional and selective hedging strategies in commodity futures markets shows that traditional hedging is more beneficial as selective hedging increases risk without additional returns.
2 sharesSource ↗
Price Discovery for Derivatives
The study investigates price discovery in a model where an agent has private information about state probabilities, extending the setting to Arrow-Debreu securities and analyzing the impact of informed demand price and information efficiency of prices.
227 sharesSource ↗
Option Characteristics and Stock Returns
The research analyzes options-implied information for predicting stock returns, finding that only a few option characteristics significantly predict returns after controlling for firm characteristics, and these are linked to asset mispricing, future tail return realizations, and short-selling costs.
1,810 sharesSource ↗
Mutual Fund Derivative Use Revealed
The paper studies fund derivative use and its impact on performance using new SEC data, revealing that despite small portfolio weights, derivatives significantly contribute to fund returns, with most funds using derivatives to amplify rather than hedge equity returns.
2 sharesSource ↗
Portfolio Performance Ratios
The article proposes four coherence axioms that portfolio performance ratios should meet, arguing that performance ratios with fixed thresholds other than the risk-free rate do not meet these axioms, allowing portfolio managers to manipulate performance ratios by altering the proportion of the risk-free asset in the portfolio.
2 sharesSource ↗
Central Clearing and Interest Rate Swap Spreads
The study investigates how the determinants of interest rate swap spreads have changed since the implementation of Title VII of the Dodd-Frank Act of 2010, finding that increases in swap volatility correspond to a tightening of swap spreads and that the Treasury liquidity premium no longer significantly influences swap spreads after the implementation of SEF trading.
2 sharesSource ↗
RePEc
Economics working papers from RePEc's NEP field reports.
18 items
Finance6
Stock Opening Price Gaps
An AI and big data study found that negative gap openings are more frequent than positive ones, and price adjustments for bad news occur faster than for good news.
11 sharesSource ↗
Regime Switching in Commodity Prices
A study from 1959 to 2022 using a 3-state Markov-switching model found that oil prices are more volatile than copper prices, reacting more to market cartelization, war episodes, and global demand shifts.
11 sharesSource ↗
VIX and SPX Futures Relation
The correlation between VIX and SPX futures strengthens during high market volatility, but transaction costs prevent profitable trading strategies.
20 sharesSource ↗
Bank System Volatility and Innovation Quality
Chinese non-financial firms' innovation quality positively correlates with banking sector volatility risk, but this effect is lessened for bank-related firms and during high economic policy uncertainty.
17 sharesSource ↗
Individual Investors' Mutual Funds Strategy
Mutual fund individual investors tend to be momentum buyers and contrarian sellers, with older and experienced investors leaning towards momentum investing, and those with smaller transactions leaning towards contrarian investing.
14 sharesSource ↗
Monetary Policy and Stock Volatility
Research from 2009 to 2021 shows that the Federal Reserve's monetary policy actions led to market volatility in the USA, especially for value stocks.
12 sharesSource ↗
Machine Learning5
Spatial Autocorrelation in Hedonic Models
The study proposes a spatially aware cross-validation strategy to correct bias in tree-based algorithms used in real estate data analysis.
23 sharesSource ↗
Financial Machine Learning with R
The paper critiques the use of machine learning in finance, offering guidance on method selection and referencing R libraries for computation.
20 sharesSource ↗
Dimensionality Reduction with Machine Learning
The paper presents a technique that merges dynamical mechanisms and machine learning to simplify high-dimensional complex systems, proving its effectiveness through tests and simulations.
19 sharesSource ↗
Euro Area Sovereign Bond Market Resiliency
The research examines the resilience of the euro area sovereign bond market, finding a negative correlation with spreads and a positive correlation with depths.
12 sharesSource ↗
Machine Learning for Finance
The article introduces a new deep learning algorithm designed to solve complex financial models. This algorithm offers accurate computations at a low cost, and can provide new economic insights.
11 sharesSource ↗
Historical Trending7
Adaptive Portfolio Selection with Peer Impact
The paper introduces an adaptive moving average method and an adaptive mean-variance model for online portfolio selection, considering the influence of other risky assets for better return prediction.
23 sharesSource ↗
Centred Expected Shortfall for Asset Managers
The article suggests using Centred Expected Shortfall instead of Expected Shortfall in asset management, explaining how to accurately decompose it.
22 sharesSource ↗
Asset Diversification in High Volatility Markets
The article advises on optimizing the risk-return ratio of an investment portfolio by choosing suitable investment proportions for each asset, using G. Markowitz's theory.
21 sharesSource ↗
Linear Empirical Bayes Method for Covariance Estimation
The paper introduces a group of best linear empirical Bayes estimators for improving covariance matrix estimation in portfolio analysis and risk management, proving their effectiveness with numerical examples and simulations.
16 sharesSource ↗
Portfolio Risk and Return in the Crypto Market
The study explores the risk-return tradeoffs of the most tradable cryptocurrencies, concluding that a portfolio of 10 cryptocurrencies provides better optimization than one with five.
16 sharesSource ↗
Independent Directors and Fraud Risk
Research indicates that independent directors can detect a company's financial fraud risk, with more dissenting opinions on financial proposals in higher risk companies.
9 sharesSource ↗
Factor Models Robustness
Research shows that the robustness of factor models changes with factor formation breakpoints, with extreme sorts yielding higher returns and centered breakpoints resulting in less risk.
8 sharesSource ↗
Machine learning
The general machine-learning papers the letter carried in 2023-25.
5 items
Recently Published5
Unprompted Reasoning in LLMs
Altering the decoding process in large language models can enhance their reasoning abilities and performance without needing specific prompts.
185 shares302 citations todaySource ↗
Hierarchical State Space Models for Sequences
Hierarchical State-Space Models, a new method for continuous sequential prediction, outperforms existing models in predicting sequences from raw sensory data, showing efficient scaling to smaller datasets and compatibility with existing data-filtering techniques.
46 shares23 citations todaySource ↗
Causal Models
The research suggests that intelligent agents must learn an approximate causal model to generalize to new domains, impacting fields like transfer learning and causal inference.
32 shares102 citations todaySource ↗
Important Structures in Turbulence with Deep Learning
A deep-learning method has been used to analyze energy structures in turbulence, finding that the most significant structures aren't always the ones contributing most to Reynolds shear stress.
94 shares88 citations todaySource ↗
Graph Learning with State Space Models
Graph Mamba Networks, a new type of Graph Neural Networks, have been introduced, which achieve excellent performance in various benchmark datasets despite lower computational cost.
39 shares167 citations todaySource ↗
GitHub
Repositories the letter featured.
7 items
Finance7
CleverCSV: Messy CSV Python Package
Messy CSV Python Package: The article introduces CleverCSV, a Python package that enhances management of complex CSV files with better dialect detection and a handy command line application.
1,188 shares
Scalping Algorithm
The article presents an algorithm for implementing a scalping strategy on multiple stocks at once using Python asyncio.
699 shares
Postgres Analytics
The use of Postgres for search and analytics functions is the main focus of the article.
3,060 shares
PandasAI Chat
The article introduces PandasAI, a tool that enables conversational data analysis through various AI models and supports multiple data formats.
9,757 shares
Local LLMs Analysis
The article investigates the use of local LLMs, particularly the Llama2 model, for automatic categorization of bank transaction data.
422 shares
Observable Data Framework
Observable Framework, a static site generator for data apps, dashboards, and reports, is introduced in the article, highlighting its combination of JavaScript and any backend language for data analysis.
1,136 shares
openwebui: ChatGPT WebUI
ChatGPT WebUI: The piece announces the rebranding of Ollama WebUI to ChatGPTStyle WebUI.
5,236 shares
News
Industry news: funds, hiring, markets and regulation.
5 items
Quantitative5
SEC Cyber Risks for Hedge Funds
New regulations requiring increased disclosure of investment strategies have raised cyber security concerns among hedge fund managers at the SEC and CFTC.
7 shares
Hedge Fund Top Earner
In 2023, Izzy Englander of Millennium Management outearned Ken Griffin of Citadel, becoming the highest earner in the hedge fund industry, as per Bloomberg.
3 shares
Tech Firm Hires Security Officer
Marlena Efstratopoulou, previously Chief Risk Officer and Chief Security Officer at Options Technology, has been promoted to Chief Information Security Officer.
3 shares
China Tightens Trading Rules
Regulators have introduced a new regime following the penalization of a major trading house for automated share dumping.
3 shares
Nvidia's Fund Gains
The Financial Times reports that hedge funds like Arrowstreet Capital and Bridgewater Associates made significant profits from increased bets on chipmaker Nvidia in late 2023.
3 shares
Podcasts
Episodes on markets, quant methods and economics.
9 items
Quantitative4
AI Evolution with Melanie Mitchell
AI researcher Melanie Mitchell explores the development of AI from cybernetics to neural networks and deep learning, and delves into the concept of intelligence.
3 shares
Risk Resilience Trends
Julie Muckleroy and Abraham Izquierdo discuss 2024's risk management trends like high interest rates, inflation, and Middle East conflicts in a podcast.
13 shares
US Rates Inflation Data
Phoebe White and Michael Feroli analyze the stronger than expected inflation data from January 2024 and forecast a gradual decrease in inflation in a podcast episode.
8 shares
EM Fixed Income Challenges
In a podcast, Jonny Goulden and Saad Siddiqui discuss the impact of recent market developments on the emerging markets fixed income asset class.
8 shares
Related5
Global Commodities US Natural Gas
The US natural gas market may experience a price drop due to unusually mild weather, despite already low prices not seen since 1995.
4 shares
Fixed Income Focus: Technicals vs Fundamentals
Technicals vs Fundamentals: Mat Rees talks about the macro events affecting fixed income, such as China's deflation and European growth divergence, emphasizing the need for thorough research.
4 shares
Investment Strategies in Inflation Volatility
The RAM Crew examines the complexities of monetary policy and inflation, and their global economic impact, focusing on the role of the Federal Reserve.
3 shares
Global FX Dollar Followthrough
Patrick Locke and James Nelligan discuss the implications of the US CPI and inflation data on the dollar in a podcast.
6 shares
MEBISODE Dividend Approach
Meb discusses the Cambria Shareholder Yield ETF and reads a paper on dividend investing in a podcast episode.
6 shares
Videos
Talks, lectures and tutorials.
3 items
Quantitative3
Quant Finance
Quants are predicted to drive growth in quantitative finance, particularly in the areas of machine learning, AI, and updating models to reflect changes in human behavior and finance.
37 shares
OCaml Efficiency
The fifth video in the OCaml series shows how locals can minimize garbage-collected allocations, providing viewers with step-by-step instructions and code.
0 shares
Business Schools vs Banks
A gap exists between banks and academia, with banks often uninformed about academic advancements, leading to surprise among academics and students.
2 shares
X / Twitter
Posts from quant researchers on X.
8 items
Quantitative4
High vs Low Investment Performance
Research indicates high-investment firms perform worse than low-investment firms, with a new predictive tool created using ChatGPT.
2 shares
Alpha in Low Volatility
The article explores the profitability potential in low-volatility investing.
2 shares
Quant Research Recap
The author provides a summary of significant quantitative research from the previous week.
1 shares
Improved Weather Forecasts by Windborne
Windborne enhances weather predictions for different sectors using its unique balloom sensors, which gather more data per dollar than conventional techniques.
1 shares
Miscellaneous4
Opening Range Breakouts Paper
A paper highlighting the importance of relative volume in opening range breakouts is reviewed in the article.
0 shares
Stocks and Macro Factors
The article posits that macro factors influence stocks during low or zero interest rates, but firm-specific factors take precedence when rates increase.
0 shares
Regime-Based Investing Insights
The article talks about regime-based investing, focusing on the predictability of inflation and the significance of US headline CPI.
0 shares