Modeling SME credit ratings using non-homogenous backward semi-Markovian approach

Abstract

Considering the growth in SME lending in Kenya and the obvious risks it posses to the banking sector, we establish a credit risk model that is responsive to the jumps in the economy. This is based on simulation of implied values of credit worthiness over a period of 12 months for 1000 SMEs, in which case we establish a case for the discrete time non-homogeneous semi-Markov approach as a proxy for internal rating model for a portfolio of SME loans. While viewing credit risk as a reliability issue, the model provides a credit indicator which gives a prospective measure of credit risk for an SME portfolio. Banks seeking to comply with the new IFRS9 guidelines can espouse this model to adequately measure impairment of financial instruments.