Search for answers or browse our knowledge base.
Corporate EAD Modelling
Publication in the Journal of the Royal Statistical Society.
Credit granting institutions are in the business of lending money to customers, some of whom subsequently fail to repay as promised. For these events, accurate loan balance estimates—termed exposure at default (EAD)—provide quantification of potential losses and form a required input to minimum credit capital calculation under the Basel II Accord. Most available EAD research estimates the credit conversion factor (CCF), which is a transform of EAD, but as we highlight this has substantial deficiencies: an inherent singularity rendering the CCF undefined or numerically unstable and it often provides EAD estimates that fail economic intuition. We build a descriptive model for EAD—without relying on the CCF—using the Global Credit Data database, advancing the literature in three important ways. First we identify, like other studies on revolving facilities, that balance and limits drive EAD and we therefore develop our model to capture these joint dynamics flexibly. Second we find evidence in the data of riskbased line management where lenders tend to decrease limits for riskier obligors. Third we confirm results from other studies of mild EAD countercyclicality, whereby EAD is lower during a subdued economy.
(Publication on a Westpac’s data set)