The credit channel of the Sovereign Spread: A Bayesian SVAR analysis
Description
This repository includes the data and the Matlab replication codes for the analysis in the article: "The credit channel of the Sovereign Spread: A Bayesian SVAR analysis" by Gianluca Cafiso, Alessandro Missale, Giulia Rivolta (2025), published in Economic Modelling. Open the "readme.txt" file for instructions, or read below. The excel file "Data_BEAR" includes the data, all the other files are the Matlab replication codes.
Files
Steps to reproduce
This file contains the instructions to replicate the results in the paper "THE CREDIT CHANNEL OF SOVEREIGN SPREAD: A BAYESIAN VAR ANALYSIS" Authors: Gianluca Cafiso, Alessandro Missale, Giulia Rivolta The codes run in Matlab 2024b and the estimation is based on: - the BEAR Toolbox available on the ECB website (https://www.ecb.europa.eu/press/research-publications/working-papers/html/bear-toolbox.en.html) - the codes of Banbura, Giannone, Lenza (2015) "Conditional forecasts and scenario analysis with vector autoregressions for large cross-sections" (International Journal of Forecasting, vol. 31, issue 3), available online. - the codes of Giannone, Lenza and Primiceri (2015) "Prior Selection for Vector Autoregressions" (Review of Economics and Statistics, vol.97, issue 2) available online. ********************************************************************** **********--- MAIN RESULTS ---********** -> Figures 3, 4 and 5 (Shocks, IRFs and HD from main VAR) These figures can be generated running the code 'main_CMR.m'. The file 'bear_settings_proxySVAR.m' should contain the following endogenous variables (line 16): SPREAD UNEMPR CPI DEBT-GDP BK-LOANS-IT BK-GBOND-IT BK-DEP SR BK-SPREAD SP -> Figure 6 (Conditional forecast) The results can be generated running the code 'CF_BGL_bear.m'. -> Figure 7 (IRFs from VAR extended with data on debt holdings) This figure can be generated running the code 'main_CMR.m'. The file 'bear_settings_proxySVAR.m' should contain the following endogenous variables (line 16): SPREAD UNEMPR CPI DEBT-GDP BK-LOANS-IT BK-GBOND-IT BK-DEP SR BK-SPREAD SP HTB-NRES HTB-RES-HHO HTB-RES-BI HTB-RES-BK HTB-RES-OFI **********--- ROBUSTNESS CHECKS ---********** -> Figure 8 (IRFs from VAR identified with Cholesky) This figure can be generated using running the code of Giannone, Lenza, Primiceri (2015) with our data. Variables are adjusted as follows: - rates (SPREAD, UNEMPR, SR, BK-SPREAD) are divided by 100; - CPI, BK-LOANS-IT, BK-DEP, BK-GBOND-IT, SP are in logarithm and multiplied by 12; The VAR model includes 13 lags. In the IRFs, rates and the DEBT-TO-GDP are rescaled multiplying by 100 while the other variables are divided by 12 and multiplied by 100. NB: Small differences may arise because of the initialization of the random number generator in Matlab. -> Figure 9 (IRFs from VAR with monthly GDP) This figure can be generated running the code 'main_CMR.m'. The file 'bear_settings_proxySVAR.m' should contain the following endogenous variables (line 16): SPREAD GDPR-INT CPI DEBT-GDP BK-LOANS-IT BK-GBOND-IT BK-DEP SR BK-SPREAD SP -> Figure 10 (IRFs from main VAR with high-relevance prior) This figure can be generated running the code 'main_CMR.m'. The option for the prior type in the code 'bear_settings_proxySVAR.m' (line 119) should be set as: s.strctident.prior_type_proxy=2;