Measuring firms’ default probabilities with imperfect information

  • Adriana de Almeida Martins Libânio (Student)

Student thesis: Master's Thesis

Abstract

This dissertation aims to estimate default probabilities in an imperfect information setting. This is achieved through the implementation of a model that considers creditors cannot observe the firm asset values directly. Instead, they receive periodic information coming from accounting reports that can be imperfect. The dataset considered covers 18 non-financial companies that belong to the Euro STOXX 50, between 2010 and 2020. Evidence was found that probabilities of default overall behave monotonically when it comes to the degree of data inaccuracy. In periods marked by extreme events, such as the European sovereign debt crisis and the covid 19 pandemic, it is possible to observe a wider impact in the probabilities of default, as the assumption of the degree of accounting noise increases. Lastly, comparing the results to the default probabilities implied by Standard & Poor’s credit ratings, it is possible to conclude that the model’ results are underestimating the probabilities of default. On absolute terms, the difference between the model default probabilities and the ones implied by credit ratings, ranges between 0.46% and 0.75%.
Date of Award27 Apr 2022
Original languageEnglish
Awarding Institution
  • Universidade Católica Portuguesa
SupervisorNuno Silva (Supervisor)

Keywords

  • Credit risk
  • Structural models
  • Reduced form approach
  • Incomplete information
  • Firm default probability

Designation

  • Mestrado em Finanças

Cite this

'