Valuation of Reverse Mortgages with Default Risk Models

This chapter discusses trading book models. Risk models come in a variety of types. However, for market risk purposes there have been a number of types which may be used within the framework. The simplest is the ‘CAD 1’ model — named after the first Capital Adequacy Directive, which permitted such models to be used in the calculation of regulatory capital. VaR models, permitted by Basel 2, were more complex, and this complexity was increased by Basel 2.5, which required the use of ‘stressed VAR’. In due course all of this will be replaced by the Basel 3 FRTB calculation, which rejects VAR and is based on the calculation of an expected shortfall (ES) market risk charge, a VaR based default risk charge (DRC) (for those exposures where the bank is exposed to the default of a third party), and a stressed ES-based capital add-on.

Download Full-text

Peer-to-peer loan acceptance and default prediction with artificial intelligence

Royal Society Open Science ◽

10.1098/rsos.191649 ◽

2020 ◽

Vol 7 (6) ◽

pp. 191649

Author(s):

J. D. Turiel ◽

T. Aste

Keyword(s):

Artificial Intelligence ◽

Small Businesses ◽

Default Risk ◽

Deep Neural Networks ◽

Support Vector ◽

Second Phase ◽

Risk Models ◽

Two Phase ◽

Test Set ◽

Default Prediction

Logistic regression (LR) and support vector machine algorithms, together with linear and nonlinear deep neural networks (DNNs), are applied to lending data in order to replicate lender acceptance of loans and predict the likelihood of default of issued loans. A two-phase model is proposed; the first phase predicts loan rejection, while the second one predicts default risk for approved loans. LR was found to be the best performer for the first phase, with test set recall macro score of 77.4 % . DNNs were applied to the second phase only, where they achieved best performance, with test set recall score of 72 % , for defaults. This shows that artificial intelligence can improve current credit risk models reducing the default risk of issued loans by as much as 70 % . The models were also applied to loans taken for small businesses alone. The first phase of the model performs significantly better when trained on the whole dataset. Instead, the second phase performs significantly better when trained on the small business subset. This suggests a potential discrepancy between how these loans are screened and how they should be analysed in terms of default prediction.

Download Full-text

Understanding performance measures for validating default risk models: a review of performance metrics

The Journal of Risk Model Validation ◽

10.21314/jrmv.2007.005 ◽

2007 ◽

Vol 1 (2) ◽

pp. 61-79

Author(s):

Jorge Sobehart ◽

Sean Keenan

Keyword(s):

Performance Measures ◽

Default Risk ◽

Performance Metrics ◽

Risk Models

Download Full-text

An examination of the Environmental Protection Agency risk assessment principles and practices: a brief commentary on section 4.1.3 of the EPA March 2004 Staff Paper

Human & Experimental Toxicology ◽

10.1191/0960327106ht580oa ◽

2006 ◽

Vol 25 (1) ◽

pp. 19-21 ◽

Cited By ~ 1

Author(s):

K A Mundt

Keyword(s):

Risk Assessment ◽

Environmental Protection ◽

Environmental Protection Agency ◽

Default Risk ◽

Assessment Practices ◽

Protection Agency ◽

Risk Models ◽

The Us ◽

Agency Risk ◽

Current Risk

The US Environmental Protection Agency (EPA) recently issued a Staff Paper that articulates current risk assessment practices. In section 4.1.3, EPA states,“...effects that appear to be adaptive, non–adverse, or beneficial may not be mentioned.” This statement may be perceived as precluding risk assessments based on non–default risk models, including the hormetic–or biphasicdose–response model. This commentary examines several potential interpretations of this statement and the anticipated impact of ignoring hormesis, if present, in light of necessary conservatism for protecting human and environmental health, and the potential for employing alternative risk assessment approaches.

Download Full-text