New dataset enables policy-driven agentic financial simulations

VERA Brief

AI-generated. Grounded in the article and its cited sources.

Researchers have introduced PAWS, a dataset that links U.S. financial and economic policy episodes with news records and stakeholder actions. This dataset supports multi-agent simulations by providing structured information on policy impacts and associated events.

Key facts

• PAWS is a dataset designed for financial multi-agent simulations.
• The dataset includes U.S. financial and economic policy episodes, news records, and stakeholder actions.
• Each action in the dataset is linked to supporting news and includes details about its interaction mode and financial-action family.
• Case studies of the 2008 short-selling ban and 2001 decimalization are part of the dataset.
• A replay study within the dataset identified challenges in detecting rare actions and calibrating agent responses.