As we mentioned in our previous blogs, Recursion’s proprietary tools Cohort Analyzer, and Pool Level Analyzer can analyze FED and CMO portfolios recursively down to the “simple pool” level. There are a wide range of applications of these powerful tools. We previously demonstrated how to calculate FED portfolio and CMO lockup rates at the macro level. Another important application is to study the collateral of mortgage bonds directly at the loan level in order to support investor’s trading decisions.
Before we delve into this particular CMO bond, we want to discuss the loan leverage coverage ratio for all agency pools. As we know, Fannie Mae discloses loan level information for all pools issued in and after 2013, and Freddie Mac does this for pools issued in and after 2006, while Ginnie Mae disclosures include loan level information for all pools that are not paid off. When Recursion was founded in 2015, due to its short disclosure history, Fannie Mae pools’ loan level coverage was fairly low. However, as of today, Fannie Mae’s loan level coverage has improved to close to 90%. As time goes on and pools issued before 2013 gradually pay off, the loan level coverage of outstanding agencies pools will reach 100% as will the CMO loan level coverage.
There are many challenges in calculating statistics of a CMO bond directly at the loan level. To name a few:
The good news is that Recursion Analyzers take care of all these challenges. Our users only need to supply the bond identifier and indicate how they wish to slice and dice the collateral to obtain the analytics they want.
Today we will use FHR 4924 BS to showcase Cohort Analyzer’s CMO capacities. This CMO is an inverse IO bond issued in 2019, backed by a mixture of 33 UMBS pools. In order to get the loan level statistics right, Cohort Analyzer will query the Fannie and Freddie loan tables at the same time.
The balance of the CMO pays down over time, and the FNM share steadily declines from 30.8% in March to 28.9% in September. This change in share is directly related to the faster prepayment speeds of the FNM collateral over this period.
This is only the beginning. Starting from this basic query, Cohort Analyzer users can add an unlimited number of different criteria, for example, by bank/nonbank, by servicer, by utilization of property inspection waiver… CMO trading desks equipped with Cohort Analyzer enjoythe benefits of easy access to loan level information to make more profitable investment decisions.