Content area
Full text
Mortgage modification is a topic that has generated a lot of press and a lot of discussion. We find that much of the press and discussion occurs without a thorough understanding of the numbers. In this article, we take a careful look at data on modification volume, types of modifications, and the success rate on each. This careful look at the data allows us to draw policy conclusions from our results. We find the following:
1. There are three key ingrethents for successful modification programs: principal reduction, substantial pay relief, and modifying early in the delinquency cycle. Each is a major contributor to modification success. When used in combination, the success rate on modifications has been quite high. The interaction between these ingrethents has not received adequate attention.
2. Modification volume is down dramatically, because most non-performing loans that qualify have already received modifications. Activity from here on will consist disproportionately of loans that are now current; this category includes loans that have previously been modified. We expect first time modifications to decrease and re -modification activity to rise.
3. The market's methods for quoting success rates on modifications are overstated. They do not take into account loans that have liquidated or been re-modified.
4. It is difficult to make sweeping statements about modification success over time, due to both the changing composition of modification and the introduction of trial modification periods.
MODIFICATION VOLUME
The left-hand panel of Exhibit 1 shows Amherst modification volume (using our proprietary methodology) versus volumes from CoreLogic LoanPerformance and Intex. Note that all three sources count actual modifications as modifications. The differences lie in the logic for inferring modifications. We are very comfortable with our own logic and will use that to form the basis for the remainder of this article. Note that CoreLogic LoanPerformance provides a reasonable match for our numbers, except for the last few months. We believe this reflects the fact that CoreLogic uses a longer (fivemonth) window to establish and confirm modifications, thus their last few observations will be subject to upward revision.
This exhibit clearly shows modification volume looking like two hills - increasing through early 2009, declining through Q3 2009, then up again, peaking in mid-2010, followed by another decline. The valley in...