GNLGTRDec 29, 2025

A Test of Lookahead Bias in LLM Forecasts

arXiv:2512.23847v17 citationsh-index: 5
Originality Incremental advance
AI Analysis

This provides a cost-efficient diagnostic tool for assessing the validity of LLM-generated forecasts, addressing a specific issue in economic forecasting.

The authors tackled the problem of detecting lookahead bias in economic forecasts from large language models by developing a statistical test based on Lookahead Propensity, and applied it to tasks like predicting stock returns and capital expenditures.

We develop a statistical test to detect lookahead bias in economic forecasts generated by large language models (LLMs). Using state-of-the-art pre-training data detection techniques, we estimate the likelihood that a given prompt appeared in an LLM's training corpus, a statistic we term Lookahead Propensity (LAP). We formally show that a positive correlation between LAP and forecast accuracy indicates the presence and magnitude of lookahead bias, and apply the test to two forecasting tasks: news headlines predicting stock returns and earnings call transcripts predicting capital expenditures. Our test provides a cost-efficient, diagnostic tool for assessing the validity and reliability of LLM-generated forecasts.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes