Data for "Bayesian Forecast Combination with Diversity-based Forward-Looking Priors"

Realvintage_dataSet_WITHnowcast_0225.mat is the real-time crude oil price dataset spans from January 1973 to August 2024, retrieved from the open-access GitHub repository (https://github.com/YunyiJake/Real-time-data-set-for-forecasting-crude-oil-prices). It is an extended and updated version based on the original oil price data from Garratt, A., Vahey, S. P., and Zhang, Y. (2019), “Real-Time Forecast Combinations for the Oil Price,” Journal of Applied Econometrics, 34, 456–462 and Aastveit, K. A., Cross, J. L. & van Dijk, H. K. (2023), ‘Quantifying time-varying forecast uncertainty and risk for the real price of oil’, Journal of Business & Economic Statistics, with the time scope of the original data systematically extended to August 2024 while retaining the consistent real-time data attributes and variable definitions of the two foundational studies, serving as a reliable empirical resource for crude oil price forecasting and related econometric empirical research. Total_medium.mat is the inflation and GDP data employed in the analysis are identical to the dataset used in Billio, M., Casarin, R., Ravazzolo, F. & van Dijk, H. K. (2013), ‘Time-varying combinations of predictive densities using nonlinear filtering’, Journal of Econometrics, 177(2), 213–232, without any modifications to the original data attributes, variable specifications and statistical calibers.