ChatGPT vs. DeepSeek: Can LLMs Predict the Stock Market and Macroeconomy?
This paper shows that ChatGPT-3.5 can predict aggregate U.S. stock returns using WSJ headlines. DeepSeek and smaller LLMs (like BERT) do not match this predictive performance.
Risk Premia
Capturing persistent risk-return trade-offs, such as carry and volatility risk premium.
Latest posts
When Investors Care About Currencies: A New Twist on Exchange Rates and Sovereign Bonds
When the euro gets stronger, European investors tend to sell government bonds from emerging countries if those bonds are in the local currencies of those countries. They do this because a stronger euro means those investments are now worth less to them, even if countries themselves aren’t riskier.
Empirical Asset Pricing via Machine Learning
This paper compares machine learning methods to predict stock returns and measure equity risk premia. The best-performing model (a 3-layer neural net) earns out-of-sample Sharpe ratios as high as 2.45.
Autoencoder Asset Pricing: Nonlinear Factors with No-Arbitrage Discipline
This paper introduces a new class of asset pricing models using autoencoders. By embedding firm characteristics into a neural network architecture, the authors construct a nonlinear conditional factor model that enforces no-arbitrage. It dramatically outperforms linear benchmarks.
Machine Learning vs. Economic Reality: The Limits of Deep Alpha
Machine learning (ML) signals look great in backtests—but they often break when exposed to economic restrictions. This paper shows that ML-based strategies lose 50–80% of their alpha when excluding microcaps, distressed stocks, or accounting for transaction costs.
Return Predictability Spikes in Bad Times: Disagreement Drives Momentum
This paper explains why return predictability—and momentum—concentrates in bad times. Investors rely on different forecasting models, which causes disagreement to spike in recessions. These disagreement shocks create time-series momentum, especially when uncertainty is high and beliefs polarize.
Trading Costs Demystified: Real Costs for Institutional Traders
Unusual negative news—news that combines rare language with negative sentiment—predicts sharp increases in market volatility. This paper shows that both firm-specific and market-wide volatility rise significantly after such news, with effects that persist for months.
Low-Risk Anomalies? It’s About Skewness, Not CAPM
This paper shows that so-called low-risk anomalies—like betting against beta or volatility—aren’t puzzles at all. They're compensation for skewness risk, especially default risk. High beta and high volatility stocks tend to be more negatively skewed, and markets misprice them if they ignore this.
Predicting Returns with Machine-Learned News Sentiment
This paper develops a machine learning method that builds its own sentiment dictionary from scratch to predict stock returns from news articles. The resulting strategy delivers much higher Sharpe ratios than those based on commercial scores like RavenPack or dictionary-based methods.
Factor Timing: Predicting When Alpha Shows Up
This paper shows that timing factor exposures over time can improve performance significantly. By forecasting returns on key equity anomalies using valuation signals, the authors build a dynamic portfolio that earns higher Sharpe ratios than static factor investing or market timing alone.
Betting Against Beta: How Leverage Constraints Flatten the Risk-Return Curve
This paper shows that investors with limited ability to borrow tend to chase high-risk assets. As a result, low-risk (low-beta) assets offer better risk-adjusted returns. A strategy that goes long low-beta assets and shorts high-beta ones—called “Betting Against Beta”—consistently beats the market.
Speed = Alpha? Measuring the Edge in High-Frequency Trading
This paper shows that in high-frequency trading (HFT), milliseconds matter. The fastest firms earn the majority of profits, not because they make better trades—but because they make more of them, faster. Speed wins.