Essays on credit risk measurement

Abstract

This dissertation investigates the applications of the Merton KMV model with the conditional volatility (GARCH) to measure credit risk in Thailand. The thesis consists of three essays. First essays calculate the distance to defaults during Mar 2010 to Apr 2018 with various methods and check the validity of the data to provide warning indicator of the default in the 461 listed companies. Result show that Both Merton KMV and conditional volatility GARCH distance to default move with the same direction while conditional volatility GARCH distance to default indicate more conservative credit risk level compare to Merton KMV distance to default model. Second essay applied the distance to default to predict the change in credit rating of the 77 listed companies which has 95 credit rating change during 2010 to 2018. Result found that some of the modelling signal being create from various type of distance to default model could predict the change in credit rating with statistically significant while most of the modelling signal was not. However, the successive modelling signal show about 50-62 percentage of predictive power. The third essay provided the application of the aggregate distance to default to the macroeconomic and financial risk indicator. Result found that the aggregate distance to default from both conditional volatility model (GARCH) and Merton KMV can intuitively represent how much credit risk of the system which It can be used as credit risk indicator. Moreover, It can formulate sectoral credit risk indicator which show consistent sign with the event. The testing shows significant value of predicting the movement of bond spread. Finally, result from this study show sign of implication in using adaptive conditional volatility structural model to capture credit risk. However, it may need further development to improve robustness of result before implementing in the real practice.