Credit models are essential to control credit risk and accurately predicting bankruptcy and financial distress is even more necessary after the recent global financial crisis. Although accounting and financial information have been the main variables in corporate credit models for decades, academics continue searching for new attributes to model the probability of default. This thesis investigates the use of corporate efficiency and corporate governance measures in standard statistical credit models using cross-sectional and hazard models. Relative efficiency as calculated by Data Envelopment Analysis (DEA) can be used in prediction but most previous literature that has used such variables has failed to follow the assumptions of Variable Returns to Scale and sample homogeneity and hence the efficiency may not be correctly measured. This research has built industry specific models to successfully incorporate DEA efficiency scores for different industries and it is the first to decompose overall Technical Efficiency into Pure Technical Efficiency and Scale Efficiency in the context of modelling financial distress. It has been found that efficiency measures can improve the predictive accuracy and Scale Efficiency is a more important measure of efficiency than others. Furthermore, as no literature has attempted a panel analysis of DEA scores to predict distress, this research has extended the cross sectional analysis to a survival analysis by using Malmquist DEA and discrete hazard models. Results show that dynamic efficiency scores calculated with reference to the global efficiency frontier have the best discriminant power to classify distressed and non-distressed companies. Four groups of corporate governance measures, board composition, ownership structure, management compensation and director and manager characteristics, are incorporated in the hazard models to predict financial distress. It has been found that state control, institutional ownership, salaries to independent directors, the Chair’s age, the CEO’s education, the work location of independent directors and the concurrent position of the CEO have significant associations with the risk of financial distress. The best predictive accuracy is made from the model of governance measures, financial ratios and macroeconomic variables. Policy implications are advised to the regulatory commission.

Credit scoring is critical for banks to evaluate new loan applications and monitor existing customers. Machine learning has been extensively researched for this case; however, the adoption of machine learning methods is minimal in financial risk management. The primary reason is that algorithms are viewed as ‘black box models’ and cannot satisfy regulatory requirements. While deep learning methods such as LSTM have been evaluated for behavioural credit scoring based on performance, research has not holistically evaluated these models on performance and explainability. To answer the research question, How can traditional machine learning and deep learning methods conform with regulatory guidelines for explainable artificial intelligence (XAI), and are they preferable to benchmark methods? this thesis used a public customer credit card dataset to compare the performance and explainability of machine learning and deep learning models against the benchmark statistical model linear regression. Model performance was evaluated using ROC-AUC, accuracy, Brier scores, F1 scores and the G-mean. The McNemar test evaluated whether, through pairwise comparison, the model performances were statistically different. The models were then evaluated on whether local and global explanations could be ascertained using feature/permutation importance and SHAP. The results found that neither the machine learning model, XGBoost, nor the deep learning model, LSTM, produced a statistically superior performance from the benchmark model. While there were performance improvements, only the machine learning model using post-hoc methods could produce local and global explanations. Given the strict regulatory environment, it is understandable that banks are hesitant to implement machine learning or deep learning models that lack the adequate levels of explainability regulators require.

It is essential for a bank to estimate the credit risk it carries and the magnitude of exposure it has in case of non-performing customers. Estimation of this kind of risk has been done by statistical methods through decades and with respect to recent development in the field of machine learning, there has been an interest in investigating if machine learning techniques can perform better quantification of the risk. The aim of this thesis is to examine which method from a chosen set of machine learning techniques exhibits the best performance in default prediction with regards to chosen model evaluation parameters. The investigated techniques were Logistic Regression, Random Forest, Decision Tree, AdaBoost, XGBoost, Artificial Neural Network and Support Vector Machine. An oversampling technique called SMOTE was implemented in order to treat the imbalance between classes for the response variable. The results showed that XGBoost without implementation of SMOTE obtained the best result with respect to the chosen model evaluation metric. / Det är nödvändigt för en bank att ha en bra uppskattning på hur stor risk den bär med avseende på kunders fallissemang. Olika statistiska metoder har använts för att estimera denna risk, men med den nuvarande utvecklingen inom maskininlärningsområdet har det väckt ett intesse att utforska om maskininlärningsmetoder kan förbättra kvaliteten på riskuppskattningen. Syftet med denna avhandling är att undersöka vilken metod av de implementerade maskininlärningsmetoderna presterar bäst för modellering av fallissemangprediktion med avseende på valda modelvaldieringsparametrar. De implementerade metoderna var Logistisk Regression, Random Forest, Decision Tree, AdaBoost, XGBoost, Artificiella neurala nätverk och Stödvektormaskin. En översamplingsteknik, SMOTE, användes för att behandla obalansen i klassfördelningen för svarsvariabeln. Resultatet blev följande: XGBoost utan implementering av SMOTE visade bäst resultat med avseende på den valda metriken.