Credit Risk
Credit Risk
Credit Risk
a practical guide
Stacia Howard 1
1. Introduction
With the ever-increasing diversification and inter-connectedness of financial systems,
regulators have been dedicating more resources to understanding the relationships within
their financial systems and investigating any inherent vulnerability. As such, a number of
methodologies have been developed to analyse the stability of the financial sector. One of
the more popular ways in which financial stability is assessed is through the use of stress
tests. Stress tests, in the commercial banking literature, refer to assessing the impact of a
rare but plausible shock to the financial system. Countries have to determine whether the
financial institutions to be included in the analysis would conduct their own individual stress
tests and then the regulator would aggregate the submitted data to arrive at a macro stress
test; or whether the regulators would collect the necessary data and conduct one stress test
based on the information received. Based on the stability reports published by various
regulatory bodies throughout the world, the preference seems to be to collect the data to
perform the stress test rather than rely on individual institutions to submit their results.
The first step is usually to determine what risks will be stressed and specifying the scenarios,
as this will assist in determining the methodology to be used and the data requirements. In
most instances historical data is employed to evaluate the sensitivity of commercial banks’
balance sheets to various shocks to macro fundamentals and then utilising the estimated
coefficients to simulate the impact on the financial system of possible stress scenarios in the
future. Three broad techniques have been used to implement the stress testing approach:
(1) time series analysis; (2) panel data regressions, and, (3) structural models (see Sorge
and Virolainen, 2006 for more details).
Time series models are perhaps the simplest technique to apply. Kalirai and Schicher (2002),
Hoggarth and Zicchino (2004) and Delgado and Saurina (2004) estimate models of the
determinants of loan write-offs or non-performing loans. The coefficients from these
regression equations were then employed to assess by how much one of the macro
fundamentals would have to change before the system experiences severe stress. Rather
than focus on one or two indicators of financial stress, Hanschel and Monnin, (2005) develop
a stress index using market price, aggregate balance sheet, non-public information and other
structural data. After estimating the stress index, the authors then try to forecast the index by
using macroeconomic imbalances.
One of the drawbacks of aggregate or time series models is that they aggregate the
microeconomic defaults that lead to financial stress. Panel data regressions can account for
bank-specific factors that may be highly correlated with financial stress at a particular
institution. Bangia et al. (2002) use credit migration matrices (which show the expected
changes in credit quality of borrowers) to provide the linkage between macroeconomic
conditions and asset quality. Using the credit ratings history of 7328 borrowers between 1981
and 1998, largely corporate institutions, the authors attempt to estimate the migration matrix,
or the probabilities of being in a particular debt-rating grade. The stress test is therefore done
1
Economist at the Central Bank of Barbados.
2
Principles for Sound Liquidity Risk, Basel Committee on Banking Supervision, BIS.
3.1 Introduction
Barbados is a small island economy in the Caribbean of about 274,000 inhabitants. It is an
open economy and therefore is fully exposed to the risks tied to international developments.
In addition, its membership of the CARICOM Single Market and Economy (CSME) exposes
the country to the risks common in high levels of regional integration. The existence of both
international and regional financial institutions in the country’s financial sector also adds
another level of global exposure. As with many countries in the world, the Barbadian financial
sector is heavily bank-based, with commercial banks accounting for 66 percent (2006) of
total financial system assets. In addition to the commercial banks, there are merchant banks,
trust and finance companies, mortgage finance companies, credit unions, pension funds,
insurance companies, asset management companies and a stock exchange. The activities of
these institutions are monitored by four regulatory bodies: the Central Bank of Barbados
(commercial banks, merchant banks, trust and finance companies and mortgage finance
companies), the Department of Cooperatives (credit unions), the Supervisor of Insurance
(insurance companies and pension funds), and, the Securities Commission of Barbados (the
stock exchange and associated brokerage firms).
The country of Barbados was chosen as the case study for a number of reasons. Firstly, it is
one of the highest ranked developing countries in the world (as per the Human Development
Index of the UNDP) and, as such, is dealing with some of the financial problems faced in
advanced economies while still hampered by the institutional framework of many developing
countries. In the opinion of the author, this offers a good example of what the average
developing country is facing, or could face in the near future, in their financial systems.
Secondly, it is a small open economy and is therefore exposed to international financial
developments with limited scope to impact global developments. Stress testing such an
economy requires a slightly different approach than commonly found in the existing literature
and often calls for additional data that may not be incorporated in the stress tests of
advanced economies. Thirdly, as part of a common market with goals of a monetary union –
the CARICOM Single Market and Economy (CSME) – it is part of a regional integration
movement similar to that of the European Union and this presents its own set of challenges,
especially with respect to the financial sector. Fourthly, the country is currently involved in an
effort to improve its monetary and financial statistics, with the assistance of the International
Monetary Fund, and the findings thus far may be of use to other countries.
As part of its surveillance of the financial systems, the CBB conducts stress tests in
conjunction with the IMF and World Bank during the Financial Sector Assessment
Programmes (FSAP) conducted by these international financial institutions. For the last
FSAP (in 2008), two different stress tests frameworks were utilised by the CBB: (1) a macro
3
Monetary and Financial Statistics Manual, International Monetary Fund.
3.2.1 Introduction
The framework chosen is adapted from Čiják (2007) and is built in Microsoft Excel.
This basic analysis was augmented using a technique formalised in Worrell (2008). For each
macroeconomic shock, the variable was stressed until the system failed to ascertain how
long or how much stress it would take before systemic failure. This is a useful exercise
because it permits policymakers a further gauge of the resilience of the financial system as
well as contributing to the development of early warning systems.
The Barbadian banking sector has six banks, none of which are domestically owned. For
three of the institutions, there was no useful estimate of capital, with the largest bank
(measured by total assets) included. Given that the results of stress tests tend to be
expressed in terms of their impact on capital, the absence of this information for half of the
institutions and more than half of the assets of the banking sector presented a significant
challenge. In order to surmount this problem and still arrive at useful results, the results of
the stress tests were expressed in terms of their impact on the return on assets of the
institutions and as a percentage of nominal GDP. In this way, the macroeconomic as well as
profitability impacts of the various vulnerabilities could be easily identified, communicated
and understood. It, still however, does not allow for an assessment of whether the stress was
sufficient to cause the bank to fail. Therefore, comparisons were made to times in the past
when similar levels of loss or bank failures were experienced to give an idea of the severity
of the stress.
The settlement date, maturity date, coupon, yield and frequency of coupon payment of each
fixed-income asset held by reporting institutions
As corporate bonds are rare in the Barbadian financial system, the data focused on
government paper as it represented the vast majority of fixed income assets of the banks
(based on conversations with commercial banks). As such, the information was sourced from
the department within the CBB that deals with government debt but this information could
also have been sourced directly from either government or the financial institutions included
in the sample. The settlement date, maturity date, coupon, yield and frequency of coupon
payment were collected for each fixed income asset and each bank. Based on this data, the
duration was calculated for each instrument, and then averaged for each bank.
Performing loans (pass loans and special mention loans) and non-performing loans
(substandard loans, doubtful loans and loss loans)
This data was already being collected by the CBB.
4
The CBB requested reporting institutions to submit this information from the beginning of their financial year.
This date varied from one institution to another.
Total deposits, divided into demand (demand and savings deposits) and time deposits
This data was already being collected by the CBB.
4. Conclusion
The data requirements of stress testing are one of the main hindrances for many countries
when trying to assess the vulnerabilities within their financial systems. This paper attempted
to outline the main variables for which data needed to be collected in order to measure the
most commonly analysed risks. In addition, it gave examples of how some data constraints
could be surmounted using Barbados as a case study. Further work, however, needs to be
done in the area of coming up with proxies for missing data, especially financial statement
data, as the implementation of new data collection forms is a complicated process that often
takes more time than is available. In addition, measurement of equity and commodity price
risks should be incorporated into future work.
References
Bank for International Settlements (2001) “A Survey of Stress Tests and Current Practice at
Major Financial Institutions” Report by a Task Force established by the Committee on the
Global Financial system of the central banks of the Group of Ten countries.
Bank for International Settlements (1999) “Credit Risk Modelling: Current Practices and
Applications” Basle Committee on Banking Supervision.
Bangia, A., Diebold, F.X., Kronimus, A., Schagen, C. and Schuermann, T. (2002) “Ratings
Migration and the Business Cycle, with Application to Credit Portfolio Stress Testing” Journal
of Banking and Finance 26: 445–474.
Basu, R., Choueiri, N. and Pascual, A. G. (2006) “Financial Sector Projections and Stress
Testing in Financial Programming: A New Framework” IMF Working Paper WP/06/33.
Basurto, M. A. S and Padilla, P. (2006) “Portfolio Credit Risk and Macroeconomic Shocks:
Applications to Stress Testing Under Data-Restricted Environments” IMF Working Paper
WP/06/283.
Blaschke, W., Jones, M.T., Majnoni, G. and Martinez Peria, S. (2001) “Stress Testing of
Financial Systems: An Overview of Issues, Methodologies and FSAP Experiences” IMF
Working Paper WP/01/88.
Boss, M., Krenn, G., Puhr, C. and Summer, M. (2006) “Systemic Risk Monitor: A Model for
Systemic Risk Analysis and Stress Testing of Banking Systems” Financial Stability Report 11,
Oesterreichische Nationalbank.
Central Bank of Barbados (2006) Annual Statistical Digest 2006.
Chan-Lau, J. A., Mitra, S. and Ong, Li Lian (2007) “Contagion Risk in the International
Banking System and Implications for London as a Global Financial Center” IMF Working
Paper WP/07/74.
Čiják, M. (2004) “Stress Testing: A Review of Key Concepts” CNB Internal Research and
Policy Note /2/ Czech National Bank.