Paper: arXiv 2309.05560
Authors: Paul Glasserman, Harry Mamaysky, Jimmy Qin
Abstract
An increase in the novelty of news predicts negative stock market returns and negative macroeconomic outcomes over the next year. We quantify news novelty - changes in the distribution of news text - through an entropy measure, calculated using a recurrent neural network applied to a large news corpus. Entropy is a better out-of-sample predictor of market returns than a collection of standard measures. Cross-sectional entropy exposure carries a negative risk premium, suggesting that assets that positively covary with entropy hedge the aggregate risk associated with shifting news language. Entropy risk cannot be explained by existing long-short factors.
Complexity vs Empirical Score
- Math Complexity: 6.5/10
- Empirical Rigor: 8.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper introduces an innovative NLP-driven entropy metric using recurrent neural networks to capture news text shifts and predict market returns. Its empirical methodology is rigorous, featuring cross-sectional asset pricing tests and macroeconomic forecasting, though replication depends heavily on access to large news text datasets.
Research Flowchart
flowchart TD
A[Research Goal: Impact of News Novelty on Market Returns & Economy] --> B[Data Input: Large Corpus of Financial News Text]
B --> C[Methodology: Train Recurrent Neural Network on News Corpus]
C --> D[Compute Entropy Measure to Quantify Text Novelty/Language Shifts]
D --> E[Outcome 1: High Novelty Predicts Negative Market Returns & Macro Outcomes]
D --> F[Outcome 2: Novelty Outperforms Standard Predictors & Unexplained by Existing Factors]
D --> G[Outcome 3: Entropy Exposure Carries Negative Risk Premium / Hedges Risk]