مدلسازی رتبه بندی اعتباری اوراق قرضه اروپایی و احتمال پیش بینی تنزل رتبه

This article proposes and empirically tests a two-step model to forecast the downgrade probability of sterling-denominated Eurobonds. In the first step, the conditional expectation of credit rating is estimated, employing an ordered probit. In the second step, the likelihood of downgrade is modeled using credit rating, as obtained from the conditional mean in the first step, alongside with traditional operating measures in a binary-probit framework. By parameterizing a system of two equations, we are able to accommodate the disentangled effect of credit quality and company financial information on the downgrade risk. We find evidence of a nonlinear response to shifts in both credit rating and leverage. The model's forecasting performance is ascertained by means of cross validation and is benchmarked against both a naive model and a neural network model.