CMStatistics 2022: Start Registration
View Submission - CFE
Title: A multilevel scoring model with pooled cross-sectional and multi-period data including financial agents efficiency Authors:  Vassilis Ioannou - University of Edinburgh (United Kingdom) [presenting]
Raffaella Calabrese - University of Edinburgh (United Kingdom)
Finn Lindgren - University of Edinburgh (United Kingdom)
Abstract: Researchers and practitioners in credit risk often need to use data pooled from various sources, which are clustered in a number of dimensions, such as different geographies, or contributing lenders. We suggest modelling default risk using a multilevel approach in order to capture the complexity of Industry level data to predict accurately across different subsegments of the population. Default risk is modelled using discrete-time survival analysis, in the presence of the competing risk of prepayment, allowing for time-varying coefficients. Finally, extensive national-level data are used to explore whether the efficiency of financial agents, such as originators and servicer companies, has an impact on the conditional probability of default.