Assessing the default risk of a public company

: the application of structural credit risk models

Abstract

Assessing default risk is a key concern many stakeholders have, let it be as a supplier, as a large customer, or most straight-forwardly as bond or equity holder. Over many years academia worked on models which help to predict credit risk inherent in a company. By applying a contingent claim framework on a firm’s capital structure it enables one to derive credit characteristics and security values. A thorough understanding of the models and drivers of credit spreads that are used to summarize a firm’s risk is important. It allows to actually applying a credit risk assessment on a sample corporation. That example is in this thesis General Electric Company. This being said, it becomes clear that empirical evidence of prior studies is supported, more precisely, after implementation of two structural credit risk models the outcomes underpredict credit spreads. Consequently this may also underestimate the probability of default as the obtained results are compared to credit default swap-spreads. Additionally, there were clear problems during the calibration process that the models did not sufficiently portray reality by any means. The Merton model did not include any tax or bankruptcy considerations whereas the Leland model resorted to the assumption of an infinite-time horizon for the implied maturity and an endogenous bankruptcy level. Nonetheless, it also proved that these firm value-based models work particularly unsatisfactory for investment grade companies. If this thesis had been looking at a speculative grade company, empirical evidence suggests improved results. The rest of this dissertation is organized as follows. First, a literature review will survey across sources and determinants of credit risk and credit spreads whereby the latest empirical findings will be discussed. The subsequent section focusses in detail on structural models. The evolution of these models, starting with the one developed by Merton (1974), is outlined and existing empirical performance evidence is presented. The second part of this dissertation forms an application of the Merton (1974) and Leland (1994) model which results are ultimately used to assess relative performance. I conclude afterwards with a brief summary of this research as well as with an opinion about the application of structural credit risk models.

Date of Award	24 Feb 2015
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Mark Shackleton (Supervisor)