A banking portfolio risk exposure: a case study

This paper assesses the factors that influence the risk exposure. The cross sectional data are from a recent portfolio for bankcards issued throughout Italy. A multinomial logit framework estimates the probability to incur in one of eleven flagged risk categories depending on a set of socio-economic and bank specific control variables. The dependent variable is constructed to estimate the operational, credit and business risk (defined as loss of profitability). The results provide a microeconomic perspective on the factors that are responsible for risk. Women, middle-aged people, residents in the Centre of Italy are more likely to be either victims or fraudsters. Customers who are resident in the Centre and North West of the country, those who own a revolving card and have had a converted or forced card are more likely to find difficulties in paying off their debt. Finally, females, people aged more than 65 years old, clients who are resident in the North East, those who own a gold card, have had a forced card and belong to the II circuit are more likely either to revoke or not to immediately activate their credit card, representing a loss in profitability for the business.