The authors conduct a comprehensive study of some parametric models that are designed to fit the unusual bounded and bimodal distribution of loss given default (LGD).
This paper presents a method for approximating the current loan-to-value (CLTV) and remaining principal structures of heterogeneous mortgage loan pools.
Estimating credit risk parameters using ensemble learning methods: an empirical study on loss given default
This study investigates two well-established ensemble learning methods: Stochastic Gradient Boosting and Random Forest, and proposed two new ensembles.
Improving credit risk modelling assumptions could soften Basel's push for input floors
Risk.net analysis finds PD floor would hit a swath of low-risk corporate loans at the biggest EU banks
Biggest share of bank capital at stake as regulators take aim at credit models
This paper compares two methods of estimating LGD: a beta regression model and a multinomial logit (MNL) model.
Regulators argue a backstop is needed to avoid too-low modelled numbers
The simple link from default to LGD
Systematic risk factors redefined