Replication package for "Common and Country-Specific Uncertainty Shocks in Europe: Why Their Nature Matters for Policy"
Description
This replication package contains the data and the code for the decomposition of the Economic Policy Uncertainty index (EPU) on the common and country-specific components using the dynamic connectedness (EPU_disaggregation), and for estimation of the effects of shocks into both components on economic activity, inflation and monetary policy setting (PVAR). The estimates of shocks are conducted using panel VAR identified with zero and sign restrictions. We show that only synchronized uncertainty shocks significantly affect Europe’s macroeconomy, while purely national shocks have limited impact.
Files
Steps to reproduce
The replication package consists of two parts: 1. EPU Disaggregation – Replicates the disaggregation of the Economic Policy Uncertainty (EPU) index into common and country-specific components. 2. PVAR Estimation – Estimates a Panel Vector Autoregression (PVAR) using the BEAR toolbox. Folder: EPU_Disaggregation The estimation of total connectedness for EPU indices is conducted in RStudio. Replication Procedure • To replicate the results, run the file driver_epu_disaggregation.R. This script generates a CSV file (GFEVDforComputationYoY_baseline.csv) containing the Generalized Forecast Error Variance Decomposition (GFEVD) estimates and total connectedness. This file serves as an input to EPU_disaggregation_baseline.xlsx, where the proposed disaggregation is computed. • The file driver_lags_selection.R is used for lag selection. • Figures 1–5 can be generated using the MATLAB script Figures.m in the Figures subfolder. Folder: PVAR The PVAR model is estimated in MATLAB 2022b using the BEAR toolbox (version 5.2). Replication Procedure • To estimate the PVAR, run EPU_BVAR.m in the examples folder. • Before running the estimation: 1. Change the paths in the addpath and dataIn to the location of your downloaded replication folder. 2. Verify the settings in bear_settings_PANEL_EPU.m, where the variable names are specified. Adjust these settings for estimations beyond the baseline scenario. • The dataset is stored in the Excel file bear_data_TC_common_demeaned.xlsx, with each country assigned a separate sheet. • Zero-sign restrictions are specified in the sheet "sign res values", while the duration of the restrictions is defined in "sign res periods".