Sovereign credit risk in a hidden Markov regime-switching framework. Part 1: Methodology

Standard approaches to estimating credit default probability estimation have certain drawbacks, most importantly regarding the underestimation of the true default probability which remains an undesirable property in sovereign risk management. As an alternative, this research applies a discrete-time Markov-modulated model to default probability estimation and applies it to Merton’s contingent claims approach, offering an attractive combination of possibly resolving the underestimation inherent in most standard structural models with a more conservative approach when predicting valuable information from a sovereign’s economic balance sheet. The crucial advantage of the estimation is that it backs the hypothesis that a regime-switching framework that allows for structural shifts can substantially improve default risk estimators, and proves that the methodology can be tractably extended to a contingent claims approach. Moreover, there are likely practical situations with certain policy implications when the predictions of the model could be used to detect systematic sovereign risk