Learning the Macroeconomic Language

2025-12-24

AI Summary

Large language models (LLMs) struggle to effectively model macroeconomic dynamics under small-sample regimes. Method: This paper proposes a novel paradigm—“theory-driven synthetic data + temporal large models”—where million-scale, theory-consistent synthetic panel data are generated via dynamic stochastic general equilibrium (DSGE) modeling and Bayesian posterior sampling, then used to train a temporal Transformer architecture. Contribution/Results: The approach synergistically integrates the structural rigor of macroeconomic theory with the representational power of LLMs, enabling the model to learn a generalizable “macroeconomic language.” Empirical results demonstrate that the resulting hybrid predictor significantly outperforms both purely statistical models and conventional DSGE models in forward-looking forecasts through 2025, achieving, for the first time, a unified balance between theoretical interpretability and data-driven predictive accuracy.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Computer Vision: Large Vision Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

We show how state-of-the-art large language models (LLMs), seemingly inapplicable to the small samples typical of macroeconomics, can be trained to learn the language of macroeconomy. We estimate a large-scale dynamic stochastic general equilibrium (DSGE) model on an initial segment of the data and obtain a posterior distribution over structural parameters. We sample from this posterior to generate millions of theory-consistent synthetic panels that, when mixed with actual macroeconomic data, form the training corpus for a time-series transformer with attention. The trained model is then used to forecast out-of-sample through 2025. The results show that this hybrid forecaster, which combines the theoretical coherence of DSGE models with the representational power of modern LLMs, successfully learns the macroeconomic language.

Problem

• Train LLMs on synthetic data for macroeconomic forecasting
• Combine DSGE models with transformers to predict future trends
• Learn macroeconomic language using limited real-world data samples

Innovation

• Generating synthetic panels from DSGE posterior distributions
• Training time-series transformer with mixed synthetic and actual data
• Combining DSGE theoretical coherence with LLM representational power