Chronologically Consistent Large Language Models

TL;DR

This paper introduces ChronoBERT and ChronoGPT, large language models trained with only data available at each time point, reducing lookahead bias and improving the credibility of social science and finance applications.

Contribution

The authors develop a novel training framework for chronologically consistent large language models that effectively mitigate lookahead bias in social science and finance tasks.

Findings

Models outperform or match standard models like BERT on NLP benchmarks.

Real-time outputs achieve Sharpe ratios comparable to larger models in finance.

Framework ensures more credible backtests and predictions in social sciences.

Abstract

Large language models are increasingly used in social sciences, but their training data can introduce lookahead bias and training leakage. A good chronologically consistent language model requires efficient use of training data to maintain accuracy despite time-restricted data. Here, we overcome this challenge by training a suite of chronologically consistent large language models, ChronoBERT and ChronoGPT, which incorporate only the text data that would have been available at each point in time. Despite this strict temporal constraint, our models achieve strong performance on natural language processing benchmarks, outperforming or matching widely used models (e.g., BERT), and remain competitive with larger open-weight models. Lookahead bias is model and application-specific because even if a chronologically consistent language model has poorer language comprehension, a regression or prediction model applied on top of the language model can compensate. In an asset pricing application predicting next-day stock returns from financial news, we find that ChronoBERT and ChronoGPT's real-time outputs achieve Sharpe ratios comparable to a much larger Llama model, indicating that lookahead bias is modest. Our results demonstrate a scalable, practical framework to mitigate training leakage, ensuring more credible backtests and predictions across finance and other social science domains.