Default Probabilities in the Merton Set-up: A Study of the Default Boundar

Abstract

Understanding when companies default matters when calculating default probabilities. With a foundation in the Merton model of risky debt this thesis investigate di erent methods of estimating default probabilities and distance to default. From the standard set-up in the Merton model with Moody's KMV as a default boundary, an expansion is put forward to investigate if a more accurate estimate of the default probability of companies can be found. This set-up includes two coupons and distinguishes between short and long term debt. For the two coupon set-up a compound call-on-call option is used to price equity at time 0. Using this option it is possible to set up two models, one where asset sales are allowed to service debt and one where asset sales are not allowed to service debt. This provide a total of three models that are tested and compared empirically on 47,878 monthly observations from 499 companies. The sample data used include observations from the time period January 2001 to December 2013. Ultimately, the thesis reach the conclusion that the two coupon model set-up does not estimate default probabilities more accurate than the original one coupon model using Moodsy's KMV as the default boundary.