Advanced

Exposure At Default During Financial Stress - A Comparative Study

Haglund, Susanna and Ripa, Julia (2016) FMS820 20161
Mathematical Statistics
Abstract
In recent years the capital requirements for banks have been updated which has complicated
the pricing procedure for derivatives. Nordea has developed a proxy model that approximates
the risk measure Exposure At Default, which is an important component in the
recently updated requirements. The purpose of this thesis is to validate the accuracy of
Nordea’s model.
In order to fulfill this purpose, models for the Exposure At Default calculations are developed
and implemented in both the risk neutral probability measure and the real world
probability measure. Both models are based on time consuming Monte Carlo simulations.
To improve speed a third, analytical solution in the risk neutral probability measure, is developed
as well. The... (More)
In recent years the capital requirements for banks have been updated which has complicated
the pricing procedure for derivatives. Nordea has developed a proxy model that approximates
the risk measure Exposure At Default, which is an important component in the
recently updated requirements. The purpose of this thesis is to validate the accuracy of
Nordea’s model.
In order to fulfill this purpose, models for the Exposure At Default calculations are developed
and implemented in both the risk neutral probability measure and the real world
probability measure. Both models are based on time consuming Monte Carlo simulations.
To improve speed a third, analytical solution in the risk neutral probability measure, is developed
as well. The result shows that the two models in the risk neutral probability measure
converge towards the same value, as the number of simulations in the Monte Carlo model increases.
There is a difference in the results generated in the two measures which is assumed
to depend on the absence of the risk premium in the real world probability measure.
The final conclusion is that Nordea’s proxy model does not generate trustworthy results.
However, considering this conclusion, Nordea was able to improve their proxy model to
generate accurate results. (Less)
Please use this url to cite or link to this publication:
author
Haglund, Susanna and Ripa, Julia
supervisor
organization
course
FMS820 20161
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Exposure At Default, EAD, Interest Rate Swap, Kalman Filter, Monte Carlo, Real World Probability Measure, Risk Neutral Probability Measure, Vasicek Model.
language
English
id
8883843
date added to LUP
2016-06-21 11:51:15
date last changed
2016-06-21 11:51:15
@misc{8883843,
  abstract     = {In recent years the capital requirements for banks have been updated which has complicated
the pricing procedure for derivatives. Nordea has developed a proxy model that approximates
the risk measure Exposure At Default, which is an important component in the
recently updated requirements. The purpose of this thesis is to validate the accuracy of
Nordea’s model.
In order to fulfill this purpose, models for the Exposure At Default calculations are developed
and implemented in both the risk neutral probability measure and the real world
probability measure. Both models are based on time consuming Monte Carlo simulations.
To improve speed a third, analytical solution in the risk neutral probability measure, is developed
as well. The result shows that the two models in the risk neutral probability measure
converge towards the same value, as the number of simulations in the Monte Carlo model increases.
There is a difference in the results generated in the two measures which is assumed
to depend on the absence of the risk premium in the real world probability measure.
The final conclusion is that Nordea’s proxy model does not generate trustworthy results.
However, considering this conclusion, Nordea was able to improve their proxy model to
generate accurate results.},
  author       = {Haglund, Susanna and Ripa, Julia},
  keyword      = {Exposure At Default,EAD,Interest Rate Swap,Kalman Filter,Monte Carlo,Real World Probability Measure,Risk Neutral Probability Measure,Vasicek Model.},
  language     = {eng},
  note         = {Student Paper},
  title        = {Exposure At Default During Financial Stress - A Comparative Study},
  year         = {2016},
}