Modeling the Default Probability: the Case of Listed Shipping Companies

During the last decade, the global shipping industry has experienced a severe recession, accompanied by an increase in the number of financial distress. In this study using logistic regressions, models for predicting the probability of default were constructed on a sample of 359 listed shipping companies for the 2000-2015 period. The predictive abilities of the obtained models were tested on the data for 2016 and were estimated using metrics of predictive quality such as an area under the ROC curve (AUC) and specificity at a given sensitivity’s level. Among the 1136 models without a macroeconomic variable, the best predictive qualities have a model that includes a non-financial variable age of shipping company, accounting and market indicators. Its area under the ROC curve and specificity at a given sensitivity's level of 0.90 are 0.87 and 0.82 respectively. It has also been found that groups of accounting variables profitability, activity and size have the greatest contribution to the probability of default of shipping companies. The key finding of this work is the possibility of improving the forecast qualities of the model by including a non-financial variable age of shipping company, market and macroeconomic predictors.