Lifetime Probability Of Default Modeling For Hungarian Corporate Debt Instruments

The paper attempts to provide forecast methodological framework and concrete models to estimate long run probability of default term structure for Hungarian corporate debt instruments, in line with IFRS 9 requirements. Long run probability of default and expected loss can be estimated by various methods and has fifty-five years of history in literature. After studying literature and empirical models, the Markov chain approach was selected to accomplish lifetime probability of default modeling for Hungarian corporate debt instruments. Empirical results reveal that both discrete and continuous homogeneous Markov chain models systematically overestimate the long term corporate probability of default. However, the continuous nonhomogeneous Markov chain gives both intuitively and empirically appropriate probability of default trajectories. The estimated term structure mathematically and professionally properly expresses the probability of default element of expected loss that can realistically occur in the long-run in Hungarian corporate lending. The elaborated models can be easily implemented at Hungarian corporate financial institutions.