Journal of Risk Model Validation

Risk.net

A prudent loss given default estimation for mortgages

Bogie Ozdemir

  • We propose a prudent methodology to correct for potential biases in LGD estimations due to historical price appreciations, appraisal biases and depreciation or potential damage to the house. 
  • This methodology can be used to estimate both Through-The-Cycle LGD and forward-looking, Point-In-Time, LGDs based on expected house prices. 
  • We also show that it can be used for Stress Testing as well as estimating a term structure for LGDs as required by IFRS-9.
  • We provide an empirical analysis to demonstrate the application of the methodology. 

ABSTRACT

House price corrections have been a cause of financial crises and continue to be a concern. Both A-paper and alternative A-paper lending are vulnerable to potential
house price corrections. The loan-to-value (LTV) ratio is a key metric used for deal acceptance, and loss given default (LGD) largely drives the capital assigned to the
deal. Yet both the LTV calculations and the LGD estimations under the advanced internal ratings-based approach (A-IRB) may have significant biases. The historical loss data used for calibration may come from a period when house prices were going up. There could also be appraisal biases that need to be taken into account. These biases may result in a significant understatement of LGDs and LTVs, especially during a period when house prices may no longer be going up, or, worse, be subject to a correction. Regulators, concerned about this possibility, are establishing floors for LGDs. We propose a prudent methodology to correct for potential biases in LGD estimations due to historical price appreciations, appraisal biases and wear-and-tear or potential damage to the house. This methodology can be used to estimate both through-the-cycle (TTC) LGD and forward-looking point-in-time (PIT) LGDs based on expected house prices. We also show that it can be used for stress testing as well as estimating a term structure for LGDs as required by the International Financial Reporting Standard (IFRS) 9. We provide an empirical analysis to demonstrate the application of the methodology.

To continue reading...

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an indvidual account here: