Re(Visiting) Time Series Foundation Models in Finance

TL;DR

This study evaluates the effectiveness of time series foundation models in financial forecasting, revealing that models trained from scratch on domain-specific data outperform pre-trained models in accuracy and economic value.

Contribution

It provides the first comprehensive empirical analysis of TSFMs in finance, highlighting the importance of domain-specific training over off-the-shelf pre-trained models.

Findings

Pre-trained TSFMs perform poorly in zero-shot and fine-tuning.

Models trained from scratch on financial data improve forecasting accuracy.

Data augmentation and hyperparameter tuning further enhance model performance.

Abstract

Financial time series forecasting is central to trading, portfolio optimization, and risk management, yet it remains challenging due to noisy, non-stationary, and heterogeneous data. Recent advances in time series foundation models (TSFMs), inspired by large language models, offer a new paradigm for learning generalizable temporal representations from large and diverse datasets. This paper presents the first comprehensive empirical study of TSFMs in global financial markets. Using a large-scale dataset of daily excess returns across diverse markets, we evaluate zero-shot inference, fine-tuning, and pre-training from scratch against strong benchmark models. We find that off-the-shelf pre-trained TSFMs perform poorly in zero-shot and fine-tuning settings, whereas models pre-trained from scratch on financial data achieve substantial forecasting and economic improvements, underscoring the value of domain-specific adaptation. Increasing the dataset size, incorporating synthetic data augmentation, and applying hyperparameter tuning further enhance performance.