Editor: David Reiss
Brooklyn Law School

July 9, 2013

Thrilla in Vanilla: Freddie v. Jumbo

By David Reiss

Kroll BondRatings issued an RMBS Commentary, Mortgage Credit Trends:  Freddie Mac vs. Prime Jumbo. This commentary is important because it offers some help in evaluating the proposed “risk sharing” securitizations that Fannie and Freddie are considering. It is also important because if compares plain vanilla agency issues with comparable private label jumbos as well as not-so-comparable limited doc jumbos.  The differences are revealing.

Kroll’s key findings are

  • Freddie Mac default and loss rates were much higher for vintages that experienced severe home price declines. The worst vintage was 2007, which experienced an estimated aggregate home price decline in excess of 18% with 12.3% of the original vintage balance liquidated to date.
  • A rapid and significant improvement in credit characteristics sharply curtailed Freddie Mac mortgage liquidation rates, which fell from 8.3% for the 2008 vintage to 0.9% for the 2009 vintage.
  • Current Freddie Mac originations continue to be of very high credit quality, with a weighted average (WA) FICO score of 767 and a WA loan-to-value ratio (LTV) of 70% for the 2012 vintage.
  • Credit performance of jumbo prime mortgages and Freddie Mac mortgages is highly comparable when controlling for characteristics such as FICO, LTV, balance, and income and asset documentation. (2)

I am most interested in the last finding. While Kroll controlled for many characteristics, the fact remained that Freddie, as a general rule, allows for fewer high risk characteristics like low doc loans and high CLTV [combined loan to value]. For instance, almost all of Freddie’s loans were full doc while about half of the private label loans were low doc. So, while Kroll appears to be correct in stating “that the credit analysis tools developed for analyzing jumbo prime loans can be productively applied to agency mortgages,” we should not take that to mean that the two products are effectively the same. (7)


| Permalink