Speaker:
Aram Asatryan
Affiliation:
UNSW Australia
Date:
Fri, 24/10/2014 - 1:30pm to 2:00pm
Venue:
OMB-150, Old Main Building, UNSW (Kensington Campus)
Abstract:
Aram will present his honours work, supervised by Assoc/Prof Spiridon Penev.
Banks wish to quantify their expected losses from active loans. Loss Given Default is one of three parameters that combine to measure this expected loss and has historically not been modeled in much depth. A better understanding of Loss Given Default will allow banks to have a better grasp of the risks they face which can lead to competitive advantages and avoiding adverse selection in loans. A large factor in Loss Given Default is whether a defaulted account will cure or not. We will look at methods to predict this probability of curing with logistic and survival methods as well as a logistic regression with a MIDAS component added to it for the first time.
