Introduction

Retail lending differs substantially from wholesale lending, and while sophisticated scoring methods are employed for classifying and/or measuring delinquency and default probabilities for individual retail credits, internal economic capital models are less fully developed for retail. Even though retail lending represents a substantial business line for most universal banks, there are several reasons why portfolio credit risk modeling has received less attention on the retail side.

Key findings

• Retail loan portfolios – A tier-2 European bank believes that most of the industry still views retail portfolios as relatively homogeneous sets of small dollar transactions that might have relatively higher, but largely predictable, expected loss characteristics. However, the growing relative size and increased volatility of retail portfolios are attracting increased attention from bank risk managers and regulators.
• Homogenous pools – While significant progress has been made in understanding the risk of commercial credits, far less research has been undertaken on measuring credit risk in retail portfolios. Research uncovered that some banks use a non-parametric ‘recursive portioning model’ which is highly suitable for grouping together the individual retail claims into homogenous pools, according to the probability of default and it overcomes some of the disadvantages of the more common parametric methods.
• Retail loans classification models – Two European banks cited that in recent years there has been a shift away from discriminant analysis in favour of logistical regression, which has the advantages of imposing fewer formal statistical requirements on the operating figures and producing more robust results.
• Future trends – Two leading European banks inferred that going forward banks will initiate a critical review of existing rating and scoring processes. Both respondents added that if participants are to benefit fully from the demonstrated effects of reducing capital cushioning, it is essential that they choose not only an efficient and selective algorithm, but that they also ensure that data quality is high and meets the demands of Basel II.

Conclusion

This report illustrates some of the main non-parametric and parametric models that are currently used by banks and confirms that the ability to separate sound from potentially defaulting borrowers with greater precision through the formation of homogenous pools reduces regulatory capital requirements.

To obtain a free copy of this section in full, please contact us at marketing@lepus.com with your name, job title, firm, phone number and email.

Tagged with: 2009, Basel II, retail, risk, Risk Research Report