Default Prediction For Commercial Mortgage Backed Securities

Abstract

Commercial mortgage default has become a topic of interest for a large number of parties due to the emergence and continued growth of the secondary mortgage market. With the multitude of parties holding a vested interest, it is important to develop a highly efficient method of monitoring collateral performance and ultimately be able to confidently predict or anticipate default. This study shows the correlation between appearance of a loan on a Watchlist and its potential to become delinquent in the future. While testing this hypothesis, a model is created that incorporates several other variables readily used to predict collateral performance for commercial mortgages and specifically commercial mortgage backed securities. Implementing the use of logistic regression, two models are created to show the level of correlation and significance with delinquency. Also, a model with a high level of explanatory power from a selected group of variables is created. The results are provided and analytical commentary on their impact is discussed in detail.