The article describes a new model called LossCalc™ version 2.0 that predicts how much money will be lost if a borrower defaults on a loan. The model uses various factors like collateral, industry, and economic conditions to make this prediction. The researchers found that measuring how close a company is to default can help predict the loss. They also discovered that recovery rates after defaults can be predicted worldwide. The model was tested using data from over 3,000 loan defaults from 1981 to 2004 and was found to be more accurate than using historical averages.