Stress Testing Credit Risk: Comparison of the Czech Republic and Germany

FSI Papers
October 2008

Introduction1

In quantitative terms, credit risk is the most important risk in banking books. This has recently been evidently shown again in the example of the US subprime crisis. Moreover, the crisis occurred despite various improvements in credit risk management, for example the progress in the field of credit risk analysis applied by banks on the portfolio level - spurred by Basel II 2 - as well as the increasing availability of a widerange of instruments that make credit risk more liquid, for example securitization and credit derivatives. Hence, credit risk remains a major threat to financial stability in a globalised financial world, where the cross-border contagion of crises particularly threatens countries with a weak banking sector.

Credit risk analysis for the financial sector as a whole can be perceived as a crucial means to prevent the occurrence of financial crises. This can be realised by means of a regular robustness test on a country's banking sector against credit risk, for example by means of stress tests being carried out by supervisory bodies and central banks providing hints to detect financial system fragility.3

Within the framework of credit risk modelling and stress testing, this study investigates and compares two countries, a new EU member state - the Czech Republic, and the largest EU economy - Germany. We seek to provide answers to various questions, notably: which macroeconomic variables are the most important to explain credit risk; whether there are country-specific differences; and what impact the occurrence of unfavourable macroeconomic circumstances can have on the macro and micro (portfolio) level, respectively.

The data employed in this study cover a period of eight years from 1998 to 2006 for the Czech Republic and twelve years from 1994 to 2006 for Germany. The investigated time horizon covers multiple periods of severe macroeconomic stress, namely the consequences of the Asian crisis in 1997 and the Russian crisis in 1998, as well as the crisis in the financial markets after 11 September 2001.

When it comes to credit risk analysis for the Central and Eastern European transitional economies (and also for many other transformation economies and developing countries), a key limitation is the availability of data as time series are usually (still) relatively short with various structural breaks. Accordingly, the data in this study have been selected very carefully to prevent misinterpretation.4 In order to arrive at ameaningful comparison between the Czech Republic and Germany, equivalent data has been sought.

We investigate both the corporate sector and the household sector and find that credit risk for corporates can be modelled based on similar macroeconomic variables for both countries, despite fundamental differences in the default rate (time) pattern. Credit risk modelling for the household sector turns out to be more challenging for both countries as there are apparently other variables than solely economic ones to explain the default rate pattern. These findings do, in general, confirm previous studies.

In the second step, we use the outcome of our credit risk modelling for macro stress testing purposes. In the third step we translate the outcome of the macro stress tests into a Basel II-type micro stress test of a hypothetical, but realistic, credit portfolio. We find that the impact of the macroeconomic shocks we study are substantially higher in the Czech Republic than in Germany, both on the macro and micro level. For a stress test of medium severity, we find an increase in aggregate corporate default rates of more than 100% and a rise in Basel II internal ratings-based (IRB 5) minimum capitalrequirements on the credit portfolio level of up to 60% in the Czech case within one year. The figures for Germany are an increase in corporate default rates of 40% and an augmentation of IRB minimum capital requirements of roughly 30% for the same credit portfolio. For the household sectors of both countries, the impact is much less pronounced compared to corporates, with a higher impact in the German case; but the outcome is less robust and has to be therefore carefully interpreted.

Our contribution to the literature is two-fold: first, we provide a comprehensive framework for stress testing; second, we directly compare a new EU member state with one of the large "old" EU economies. While there is no directly comparable study in the literature, we confirm the finding of previous studies that stress events have a more material effect in less developed economies.

The paper is organised as follows. Section 2 provides an overview of related studies. Section 3 describes the model used to analyse credit risk for the corporate and household sectors. In Section 4, the underlying data are presented. Next, the corporate and household models are calibrated for the Czech Republic and Germany in Section 5 and 6, respectively. Section 7 is devoted to stress testing. Finally, section 8 provides conclusions.


1 The findings, interpretations and conclusions expressed in this paper are entirely those of the authors and do not represent the views of any of the authors' institutions. We thank Miroslav Singer, Vladimír Pikora, Thilo Liebig, Christoph Memmel and Pierre Tychon for their valuable comments and support. The financial support for this project by the Czech National Bank is gratefully acknowledged. Petr Jakubík: Czech National Bank and the Institute of Economic Studies of Charles University in Prague. Contact: Petr.Jakubik@cnb.cz. Christian Schmieder: Deutsche Bundesbank and European Investment Bank. Contact: C.Schmieder@eib.org.
2 See BCBS (2006), European Commission (2006) and the respective national transposition of Basel II.
3 The importance of this mission is underlined, for example, by the fact that the Basel Committee on Banking Supervision mandated a working group of the Research Task Force to further investigate stress testing.
4 We draw on the findings of the analysis carried out by Jakubík (2007).