University of Southampton – School of Management, 
in collaboration with SAS UK: 

PHD OPPORTUNITY IN THE AREA OF CREDIT SCORING, CREDIT RISK MODELLING, AND 
BASEL II

Applications are invited for an EPSRC Industrial CASE studentship to 
support doctoral study in the area of credit scoring, credit risk 
modelling, and Basel II.  The three and a half-year studentship covers 
tuition fees and includes payment of a maintenance stipend of not less 
than £14,600 per year and research support.  Applicants must meet EPSRC 
eligibility requirements and have (at least) an upper second class honours 
degree (or equivalent) in a suitable subject.  The student will have the 
opportunity to work in an established credit risk research team within the 
school (supervised by Prof Lyn Thomas, Dr Christophe Mues and Dr Bart 
Baesens), which is part of an EPSRC-sponsored research consortium, the 
Quantitative Financial Risk Management Centre (QFRMC).  He/she will 
collaborate closely with the industrial sponsor, SAS, on problems that are 
of direct interest to their work in the credit risk domain, and that 
currently attract much attention from the financial sector.  

Project Description: Basel II-Compliant Credit Risk Modelling for Low-
Default Portfolios
Credit scoring involves the use by the finance industry of operational 
research and statistical models to assess and manage credit risk (which 
includes among other things the probability of borrowers defaulting on 
their credit obligations).  Originally introduced as a handy tool for 
deciding whether to accept or reject individual applications for retail 
credit, credit scoring techniques are nowadays used to administer and 
follow-up default risk across a whole range of credit portfolios.  In 
particular the introduction of the new Basel II regulatory capital 
framework has created an imminent need to develop new or further refine 
existing models in this area.  However, a common problem for building 
statistically reliable models is that, for several of the bank's 
portfolios, the available data only contain a very low number of defaulted 
exposures or loss events.  Typical examples of these so-called low default 
portfolios (LDPs) include some forms of corporate exposures, bank and 
sovereign exposures, or even prime mortgage loans.  The aim of this 
project is to develop and/or further investigate several potential 
statistical or machine learning approaches for dealing with credit risk 
estimation in LDPs, and to empirically validate their effectiveness on 
real-life data.  

We are looking for motivated candidates with a good 
quantitative/analytical background in one or more of the following areas: 
management science, OR, statistical analysis, mathematics for finance, 
data mining & machine learning.  

For application details and further information please contact Dr 
Christophe Mues (email: [log in to unmask]).  More background information 
on the School of Management and the Quantitative Financial Risk Management 
Centre (QFRMC) can be found on www.management.soton.ac.uk and 
www3.imperial.ac.uk/mathsinstitute/programmes/research/bankfin/qfrmc, 
respectively.  

Please submit applications by 18th June 2007.