Statistical Methods in Finance 2023

December 19 to 23, 2023



Markov Regenerative Credit Rating Model

By S Dharmaraja
Indian Institute of Technology, Delhi

Credit rating serves as an important input in the risk management of the financial firms. The level of credit rating changes from time to time because of random credit risk and, thus, can be modelled by an appropriate stochastic process. Markov chain models have been widely used in the literature to generate credit migration matrices; however, emergent empirical evidences suggest that the Markov property is not appropriate for credit rating dynamics. The purpose of this talk is to address the non-Markov behavior of the rating dynamics alone with big data and optimization issues. A model is proposed based on Markov regenerative process (MRGP) with subordinated semi-Markov process (SMP) to obtain the estimates of rating migration probability matrices and default probabilities. Further, the applicability of the MRGP model with the help of historical credit rating data is numerically illustrated.