Generalized additive modelling of the repayment performance of Korean borrowers

Young Ah Kim and Peter Moffatt
Additional contact information
Young Ah Kim: University of Essex

No 2016-03, University of East Anglia School of Economics Working Paper Series from School of Economics, University of East Anglia, Norwich, UK.

Abstract: Data from a sample of around 32,000 customers taking out personal loans from a Korean bank, are analysed. The focus of analysis is a binary variable indicating default, defined as any sort of failure to meet the obligations of the loan during a time period up to a fixed reference date. Around 1.5% of the sample defaulted. The Generalized Additive Modelling (GAM) framework is used to investigate the combined effect of a number of factors on the likelihood of default. The GAM framework allows flexibility in the effects of continuously- distributed predictors. The B-spline smoothing approach is used for each of these effects. An extensive model-selection process is implemented. It is found that the statistical fit improves if some, but not all, of the predictors are modelled with the B-spline. The two variables found to have non-linear effects are amount borrowed and age of borrower. The predicted probability curve obtained for the former shows that borrowers least likely to default are those that have borrowed at the 85th percentile of the amount-borrowed distribution. The prediction curve for the latter shows that the default probability declines in a stepwise fashion with age, falling abruptly at certain ages, but appearing to level off for significant periods within the life-cycle.

Date: 2016-05-12
New Economics Papers: this item is included in nep-rmg
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://ueaeco.github.io/working-papers/papers/ueaeco/UEA-ECO-16-03.pdf main text (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:uea:ueaeco:2016_03

Ordering information: This working paper can be ordered from
Reception, School of Economics, University of East Anglia, Norwich Research Park, Norwich, NR4 7TJ, UK

Access Statistics for this paper

More papers in University of East Anglia School of Economics Working Paper Series from School of Economics, University of East Anglia, Norwich, UK. Contact information at EDIRC.
Bibliographic data for series maintained by Cara Liggins ().