Revision-Aware Proper Scoring Rules for Real-Time Macroeconomic Forecasting: Public Replication Package

Source
DataCite
Published
2026
License
Creative Commons Attribution 4.0 International
Themes
Green Economy, Machine Learning & AI

Get the data from DataCite

About this dataset

This archive provides the public replication package for the study “Revision-Aware Proper Scoring Rules for Real-Time Macroeconomic Forecasting”. The study develops a revision-aware framework for evaluating probabilistic macroeconomic forecasts when official targets are preliminary releases, finite-maturity vintages, or complete official release paths. The package supports the construction of leakage-safe forecast-origin records, release-path records, forecast-distribution records, score records, timing audits, and the tables and figures used in the manuscript and supplementary materials.The empirical applications cover U.S. real GDP release paths, GDP maturity and sample-robustness checks, U.S. industrial production vintages, and quarterly real GDI validation exercises. The methodology implements target-aware forecast evaluation for revised official macroeconomic data, including...

How the data was collected

The methodology implements target-aware forecast evaluation for revised official macroeconomic data, including scalar and release-path scoring, CRPS, RMSE, log-score diagnostics, marginal release scores, joint energy scores, composite release-path scores, score-differential diagnostics, Newey–West sensitivity checks, moving-block bootstrap flip-rate intervals, PIT and prediction-interval calibration diagnostics, and synthetic revision-environment validation.The archive contains source-download scripts where public providers allow automated retrieval, source manifests and provider URLs, manifest hashes, processed audit records, forecast records, score records, configuration files...

Browse sustainability datasets on SusanHub