Capturing the nonlinear impact in distress state: Enhancing scenario design of stress test

Abstract

This study uses a switching model to capture the nonlinearity in deteriorating macroeconomic conditions in Indonesia. This design is very useful in that it enhances the financial authority’s stress testing framework, which usually does not consider the nonlinearity phenomenon and ends up underestimating the loss generated by the stress conditions. A triggering index, incorporating a combination of domestic and international macroeconomic variables, is implemented to switch the economy from a normal to a distress state, and vice versa, using the threshold vector auto-regression (TVAR). We find the in two states responses to shocks are asymmetric. The distress state amplifies a negative shock to the financial system, and also recovers faster to a positive shock. Likewise, the financial system is more resistant to switch to a distress state from a normal state, unless the shock is higher than what the system could withstand.