How Severe Are Euro Area Regulatory Stress Test Scenarios? Evidence from a Growth-at-Risk Framework

This dataset supports the analysis of medium-term downside risks to euro area growth using a Growth-at-Risk (GaR) framework. The objective is to assess how macro-financial conditions affect the distribution of future real GDP growth, with particular emphasis on left-tail outcomes relevant for macroprudential surveillance and supervisory stress testing.

The dataset combines a broad set of macro-financial indicators capturing domestic financial stress, monetary conditions, balance-sheet vulnerabilities, external competitiveness, foreign demand, and global commodity shocks. Variables include real GDP, the ECB Composite Indicator of Systemic Stress (CISS), real short- and long-term interest rates, credit to the private non-financial sector, house prices, equity prices (EURO STOXX 50), the real effective exchange rate (REER), extra-euro area exports of goods, and oil prices (WTI). The dependent variable is three-year cumulative real GDP growth.

The sample spans 1980Q1–2024Q4 at quarterly frequency. To ensure a sufficiently long historical coverage, several series are extended backward through splicing and chaining procedures using harmonized euro area aggregates and historical proxies. Real GDP and interest rates combine ECB Area Wide Model data with later ECB and Eurostat releases. Equity prices are back-cast before the availability of EURO STOXX 50 data using euro-converted S&P 500 series. Credit aggregates are extended using M3 growth proxies, while oil prices are reconstructed using historical exchange rates prior to euro adoption. All nominal quantities are deflated using the euro area GDP deflator.

Variables are transformed to preserve economic interpretability and consistency with the macro-financial literature. Financial stress (CISS) enters in levels, interest rates in year-on-year changes, and quantity or price variables in year-on-year growth rates. Predictors are aligned at time t to explain GDP outcomes over t+1 to t+12 quarters, avoiding look-ahead bias.

The dataset is designed for quantile-regression-based estimation of conditional growth distributions and parametric density interpolation using skewed-t distributions. It supports the computation of Growth-at-Risk measures, downside risk indicators, recession probabilities, uncertainty measures, and stress-test severity benchmarking exercises.

Data sources include Eurostat, the European Central Bank, the ECB Area Wide Model database, BIS, IMF Direction of Trade Statistics, FRED, and the ECB Data Portal.