Conducting a “top-down” approach to stress testing provides a useful check on the results on the basis of individual balance-sheet information (the bottom-up approach). Furthermore, financial institutions in some countries may not have the capacity to estimate the effect of a given set of shocks on their portfolio. In this case, the agency coordinating the stress-testing exercise could adopt a “top-down” approach and could apply adjustment parameters that are based on systemwide estimates. For example, a regression model of loan loss rates for the entire banking system could be used to estimate the effect of a macroadjustment scenario on the credit quality of an institution. Examples of this approach include the following:
• Fr0yland and Larsen (2002) modeled losses for Norwegian banks on household loans as a function of household debt, wealth, and unemployment. They also mod-
eled losses on loans to enterprises as a function of risk-weighted debt and collateral. Andreeva (2004) modeled the loan loss ratio (to assets) of loans to Norwegian enterprises as a function of bankruptcy probabilities and a variety of economic factors, including the unemployment rate and the real interest rate.
• Benito, Whitley, and Young (2001) extended the Bank of England’s macromodel by incorporating household and corporate balance sheets. They then performed a stress test by incorporating a fall in housing prices and a rise in interest rates and by examining the effect on a variety of indicators, including mortgage arrears.
• Hoggarth and Whitley (2003) described the process of using the Bank of England’s macromodel, as well as using the top-down approach, to estimate the effect of macrovariables on new provisions by banks.
• Arpa et al. (2000) estimated the effect of macroeconomic factors (real GDP, real estate prices, inflation, and real interest rates) on risk provisions and on earnings for Austrian banks. Kalirai and Scheicher (2002) modeled loan loss provisions in Austria as a function of various macroeconomic indicators and then used the model to conduct a series of sensitivity tests.
• Pesola (2001) examined the Nordic banking crisis by estimating a model of loan losses as a function of GDP, indebtedness, unexpected changes in income and interest rates, and deregulation.
The estimated equations from those papers are all examples of how the authorities or individual institutions can use the top-down approach to approximate the effect of economic developments on individual portfolios or to calibrate the parameters used in their stress tests. Regression-based estimates have their limitations, because they are often providing only a partial equilibrium estimate of some effect; therefore, care should be taken in interpreting the results of such estimates.