Empirical performance of loss given default prediction models

Two models of stochastic loss given default

We propose two structural models for stochastic loss given default that allow the credit losses of a portfolio of defaultable financial instruments to be modeled. The credit losses are integrated into a structural model of default events accounting for correlations between the default events and the associated losses. We show how the models can be calibrated and analyze the impact of correlatio...

Estimating Conservative Loss given Default

The new Basel Capital Accord (Basel II) is going to be embedded in the risk management practices at many financial institutions shortly, but the academic and financial world are still discussing about several topics related to the new capital adequacy rules. One of the most important and prominent examples among these topics is the link between loss given default (LGD) and the economic cycle. I...

Estimating bank loans loss given default by generalized additive models

Loss given default models incorporating macroeconomic variables for credit cards

Based onUKdata formajor retail credit cards,we build severalmodels of Loss GivenDefault based on account level data, including Tobit, a decision tree model, a Beta and fractional logit transformation. We find that Ordinary Least Squares models with macroeconomic variables perform best for forecasting Loss Given Default at the account and portfolio levels on independent hold-out data sets. The i...

Loss given default as a function of the default rate

A recently derived function ties a portfolio’s loss given default rate (LGD) to its default rate. This study compares the predictive performance of the LGD function to that of linear regression using simulated data. The data are simulated using a linear model. Even though this confers an advantage to linear regression, the LGD function produces lower mean squared error over a meaningful range o...