The Monetary Policy Statement Database (MPSD) is a comprehensive collection of central bank communications designed for research on global financial conditions and monetary policy transmission. It provides a standardized, machine-readable dataset of policy statements from major central banks worldwide.
- Version: v0.1 (Beta Release)
- Date: March 2026
- Status: Preliminary release. Future updates may modify structure and content.
- Official web page
- Official Registration Form
The MPSD is a semi-public resource. This is a beta (early access) release; the full version will be released upon journal publication. To access the dataset and accompanying codebase, please follow these steps:
- Complete the Sign-up Form: Fill out the Official Registration Form to provide your name, affiliation, and research intent.
- Provide GitHub Username: Ensure your GitHub handle is included correctly in the form.
- Wait for Authorization: Once your request is processed, your GitHub account will be granted access to this repository.
Users may be contacted regarding updates, corrections, and related research outputs.
By requesting access, you agree to cite the following reference in any research using this dataset. The dataset may not be redistributed or publicly shared without permission.
Note: Access is granted for academic research purposes only.
- Scope: 6,693 statements from 51 central banks.
- Timeframe: 1990 to 2024.
- Methodology: Advanced pipeline combining standard NLP preprocessing with Large Language Model (LLM) tools for aspect-based sentiment and question answering.
- Reproducibility: Includes full codebase for cross-country analysis and indicator generation.
The database is actively maintained to ensure it remains a relevant tool for the research community:
- Update Frequency: Data is refreshed on a quarterly or semi-annual basis.
- Expansion: The list of jurisdictions is continuously growing, with 57 countries currently in the expansion pipeline.
- Communication Trends: Policy statements have increased in length since the 2008 Global Financial Crisis, with modest improvements in readability.
- Global Synchronization: Inflation references show significant comovement across countries during global inflation episodes.
- Sentiment Impact: LLM-derived sentiment indicators suggest that central bank communication can predict the Global Financial Cycle rather than merely reacting to it.
If you use this database or the associated methodology in your research, please cite the following research paper:
Baird, C., Benchimol, J., Vyshnevska, V., Sohn, W., & Vyshnevskyi, I. (2026). The Monetary Policy Statement Database. In Data Science in Central Banking, IFC Bulletin No. 67, Chapter 14. Bank for International Settlements.
@incollection{BairdBenchimolVyshnevskaSohnVyshnevskyi2026a, author={Baird, Cory and Benchimol, Jonathan and Vyshnevska, Vira and Sohn, Wook and Vyshnevskyi, Iegor}, editor={{Bank for International Settlements}}, title={{The Monetary Policy Statement Database}}, booktitle={Data Science in Central Banking}, publisher={Bank for International Settlements}, year={2026}, volume={67}, series={IFC Bulletins chapters}}
Detailed documentation and technical specifications can be found in the BIS IFC Bulletin No 67.
A journal version of this work is currently under review. Users will be notified when the final publication becomes available.
Use of this dataset implies agreement to cite the above reference in any resulting research.
This repository (data and text) is licensed under CC BY-NC 4.0. Use is permitted for academic research only; commercial use is prohibited.
For questions, bug reports, or access issues, please open a GitHub Issue.