Skip to content
All library documents

Using LLM Agents to Predict Federal Funds Rate Decisions

Article arXiv papers · Author: Yuhan Hou et al.
Read the original

Summary

FedSight AI uses large language model agents to simulate Federal Open Market Committee deliberations and predict federal funds target rate decisions. Individual agents review structured economic indicators and unstructured material such as the Beige Book, discuss policy alternatives, and vote, aiming to reflect the committee’s decision process.

The system also applies Chain-of-Draft, a concise multistage reasoning approach. For meetings in 2023 and 2024, the paper reports 93.75% prediction accuracy and 93.33% stability, and says the system outperformed MiniFed and an ordinal random forest baseline. Its stated strength is transparent reasoning aligned with FOMC communications. The supplied description does not specify the number of evaluated meetings, evaluation protocol, or whether the results generalize beyond that period, so the reported figures should be read within the stated evaluation setting.

Key ideas

  • FedSight AI assigns LLM agents to analyze economic indicators and policy communications.
  • Agents debate policy options and vote to simulate FOMC deliberations.
  • Chain-of-Draft adds concise, multistage reasoning to the framework.
  • The paper reports accuracy and stability results for 2023–2024 meetings and comparisons with named baselines.
  • The brief description does not provide evaluation details needed to judge broader generalization.

Tags

Fixed incomeMachine learningStatistics
Full text
# FedSight AI: Multi-Agent System Architecture for Federal Funds Target Rate Prediction


# FedSight AI: Multi-Agent System Architecture for Federal Funds Target Rate Prediction









The Federal Open Market Committee (FOMC) sets the federal funds rate, shaping monetary policy and the broader economy. We introduce \emph{FedSight AI}, a multi-agent framework that uses large language models (LLMs) to simulate FOMC deliberations and predict policy outcomes. Member agents analyze structured indicators and unstructured inputs such as the Beige Book, debate options, and vote, replicating committee reasoning. A Chain-of-Draft (CoD) extension further improves efficiency and accuracy by enforcing concise multistage reasoning. Evaluated at 2023-2024 meetings, FedSight CoD achieved accuracy of 93.75\% and stability of 93.33\%, outperforming baselines including MiniFed and Ordinal Random Forest (RF), while offering transparent reasoning aligned with real FOMC communications.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.