Macro Stress Testing Credit Risk: Case of Madagascar Banking Sector

Abstract

Abstract This study proposes to assess the vulnerability of banking sector’s credit portfolio under macroeconomic shocks and to evaluate its impact on banking system capitalization. Our method uses the Global Vector Autoregressive (GVAR) Model to generate adverse macroeconomic scenarios. The GVAR model is combining by the satellite credit risk equation to find the non-performing loan under stress conditions. The advantage of using GVAR model is that on the one hand, it captures the transmission of global, external and domestic macroeconomic shocks on banks non-performing loans. On the other hand, this model considers the nonlinear pattern between business cycle and the bank credit risk indicator during the extreme events as highlighting by the macro stress test literature. The forecast of non-performing loan is then used to obtain stress projections for capital requirement for the banking system level. This article attempts to fill the lacks concerning the stress testing works about Madagascar which study is a recent framework, whose no study on dynamic macro stress testing was treated before. The Results outline the interaction of aggregate non-performing loan with macroeconomic evolution. The horizon of capital prediction shows that banking sector reacts most to a GDP shock. Also, Madagascar banking sector is quite resilient and remains sufficiently capitalized under all macroeconomic scenarios designed with a solvency ratio higher than the minimum regulatory CAR ratio.