Actual losses on loan portfolios are driven by many factors; accordingly, projecting such losses needs to reflect this plethora of factors. The factors that have been shown to be strongly predictive include LTV, FICO, housing price changes, and occupancy types.  These obviously need to be considered in projecting expected losses.  What is often overlooked, however, are the less understood delinquency migration patterns.

The statistically significant dynamic - illustrated below - confirms the belief that high FICO conforming loans have a lower probability of becoming one or two months delinquent.  What is less understood, though, is that once these good credits hit two months delinquent, they are much more likely than poorer credits to go through default.

The Level1Analytics CECL model takes these migration patterns into consideration, along with the other predictive factors.

Why guess when you could know? Take your first step towards better data, and fuel better decision making for your organization.