Fake Date Tests: Can We Trust In-sample Accuracy of LLMs in Macroeconomic Forecasting? The views expressed herein are solely those of the authors. The results presented are preliminary and are published to stimulate discussion and receive feedback for possible further improvements to the article. Th
Alexander Eliseev, Sergei Seleznev · 2026
Read Fake Date Tests: Can We Trust In-sample Accuracy of LLMs in Macroeconomic Forecasting? The views expressed herein are solely those of the authors. The results presented are preliminary and are published to stimulate discussion and receive feedback for possible further improvements to the article. Th by Alexander Eliseev, Sergei Seleznev (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-07992.txt.
Description
Alexander Eliseev and Sergei Seleznev question backtests that grade language models on data they already trained on. Economists now apply language models to tasks like macroeconomic forecasting. They often backtest the models, even though training data include the same period being forecast. The authors call their check fake date tests. Their 2026 discussion paper asks whether in-sample accuracy of such forecasts can be trusted. The results are presented as preliminary, to invite feedback.
A model that memorized past news can look prophetic without forecasting skill. Fake dates expose that by moving the claimed present. Honest evaluation matters before models inform economic decisions. The paper serves economists adopting language models and evaluators who audit them. It supplies a simple test that separates memory from genuine prediction. The authors publish through a discussion series for working papers.
Catalog record
| CREATOR | Alexander Eliseev, Sergei Seleznev |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 120 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-07992