Stress Testing for Default Probabilities in Banking Industry; An Application of Credit Portfolio Approach

Abstract

Because of prevalence of non-performing loans in Iranian banking sector, it is important to estimate the default probability of borrowers in order to effectively manage credit risk. This paper conducts stress testing for default probabilities in banking industry of Iran. We apply the credit portfolio approach model developed by Wilson (1997) and analyze the impacts of various macroeconomic shocks on default rates of banks. In the constructed model, we first estimate the effects of macroeconomics variables on default rate. Then the dynamic relationship between selected macroeconomics variables is estimated by a VAR model. Residuals obtained in the two previous steps were used to construct the covariance matrix for system of equations. Finally, using the Monte-Carlo method, a path of default probabilities is simulated in a one-year horizon under different scenarios. We compare default rates under different stress scenarios with baseline scenario to identify the effects of different shocks. The results of simulation show that unemployment rate shock has been the most harmful factor for default probabilities, followed by exchange rates shock. A shock to GDP growth also affects default rates significantly. Inflation shock generates the least important effect on default rates, consistent with the insignificant coefficient of inflation rate in the estimated default probability equation. A simultaneous shock to all macroeconomic variables has higher impact on the default rates in lower tails than upper tails. The results also show the effects of shocks decrease with the passage of time.

Keywords

• credit risk
• stress test
• wilson model
• simulation
• banking industry