There has been a plethora of research into company credit risk and financial default prediction from both academics and financial professionals alike. However, only a limited volume of the literature has focused on international shipping company financial distress prediction, with previous research concentrating largely on classic linear based modelling techniques. The gaps, identified in this research, demonstrate the need for increased effort to address the inherent nonlinear nature of shipping operations, as well as the noisy and incomplete composition of shipping company financial statement data. Furthermore, the gaps illustrate the need for a workable definition of financial distress, which to date has too often been classed only by the ultimate state of bankruptcy/insolvency. This definition prohibits the practical application of methodologies which should be aimed at the timely identification of financial distress, thereby allowing for remedial measures to be implemented to avoid ultimate financial collapse. This research contributes to the field by addressing these gaps through i) the creation of a machine learning based financial distress forecasting methodology and ii) utilising this as the foundation for the development of a software toolkit for financial distress prediction. This toolkit enables the practical application of the financial risk principles, embedded within the methodology, to be readily integrated into an enterprise/corporate risk management system. The methodology and software were tested through the application of a bulk shipping company case study utilising 5000 bulk shipping company-year accounting observations for the period 2000-2018, in combination with market and macroeconomic data. The results demonstrate that the methodology improves the capture of distress correlations, that traditional financial distress models have struggled to achieve. The methodology's capacity to adequately treat the problem of missing data in company financial statements was also validated. Finally, the results also highlight the successful application of the software toolkit for the development of a multi-model, real time system which can enhance the financial monitoring of shipping companies by acting as a practical "early warning system" for financial distress.
|Date of Award||21 May 2020|
- University Of Strathclyde
|Supervisor||Evangelos Boulougouris (Supervisor) & Sukie Lee (Supervisor)|