EBA

March 2018

Contents

Repository stores data, programs, and results for bottom-up analysis of the default risk of European banks under the baseline and adverse scenarios of the 2018 EU-wide stress tests designed by the European Banking Authority (EBA).

Banks analyzed are headquartered in the following countries:

Analytical Engine

• The default analysis was performed using the Matlab-based tool BuDA: A Bottom-Up Default Analysis Framework, version 2.0, Octover 2010, 2017 release, developed by J.-C. Duan, W. Miao, J.A. Chan-Lau, and the Credit Research Initiative Team of the National University of Singapore.

• Data, scripts, and programs used the following software:

  • Matlab 2017b
  • R version 3.4.3
  • RStudio version 1.1.423
  • LibreOffice Calc version 6.0.1.1 (x64)
  • Microsoft Office 365 version 1708

Data files

• EBA Data.xlsx Main data file. Retrieves macroeconomic and financial time series from Thomson-Reuters. The main series used in the analysis are national time series of the following variables, except for the 3-month EURIBOR rate:

  - Real GDP (quarterly)
  - Unemployment (monthly)
  - CPI (quarterly)
  - Long term interest rates (10-year government bonds, monthly)
  - EURUSD (monthly)
  - Oil prices (WTI, monthly)
  - Stock price returns (National stock index)
  - 3-month EURIBOR rate
  

• rpp.xlsx Residential property prices, quarterly series. Data retrieved from the European Central Bank (ECB) Statistical Data Warehouse (SDW); and file produced using the R script prgGetECBData.R.

• EBA Tables directory This directory contains separate Acrobat files describing baseline and adverse scenarios assumptions for each country. Data retrieved using the R script prgReadPDFTable.R.

• Testing directory The directory stores the user-supplied macro scenario xlsx files needed to run BuDA. Files prepared using the R script prgReadPDFTable.R.

• Training directory The directory contains the CSVfiles used to train the BuDA models for each country. Files prepared using the R script prgBuDATraining.R.

• Internal Macros directory The internal macros directory contains the xlsx files with the scenario projections for stock index returns and the 3-month EURIBOR rate. Files prepared using the R script prgReadPDFTable.R.

Program files

Programs should be run sequentially:

• prgGetECBData.R Uses routines in CRAN package ecb to retrieve residential property prices directly from the ECB SWD and creates the rpp.xlsx file. Access to the SWD may be intermittent and retrieving the data may require running the script several times.

• prgBuDATraining.R Prepares the training files using data from the files EBA Data.xlsx and rpp.xlsx, and stores them in the training directory.

• prgReadPDFTable.R Scrapes scenario information from PDF document, Adverse macro-financial scenario for the 2018 EU-wide banking sector stress test, European Systemic Risk Board (ESRB), January 16, 2018. Once it reads the data, it combines it with data from the testing scenarios to create the user-based macro scenarios and the internal macro-scenarios. The scenarios are stored in the Testing and Internal Macros directories respectively.