The NPV test is a "black box" in which data goes in, a decision comes out, and borrowers aren't privy to details about how the decision was made. The Federal Housing Finance Agency maintains secrecy over its formulas, updated at the beginning of each quarter, for calculating projected home values and REO discounts. Mortgage servicers don't disclose the probabilities that they assign for redefault and self-cure rates and the rest.
Go public with calculationsConsumer advocates have lobbied, in vain, for this information to be made public. "Without access to the NPV analysis, homeowners are entirely reliant on the servicer's good faith," Julia Gordon, senior policy counsel for the Center for Responsible Lending, told the House Financial Services Committee Dec. 8.
Gordon added that "servicers should be required to allow borrowers to review the property valuation used in the NPV calculation, as it is one of the inputs with the greatest effect on the results."
But servicers and government agencies have rebuffed such requests. They say the NPV calculation is beyond the comprehension of many borrowers.
"There's always some desire for a level of transparency with the consumer," says Joanne Gaskin, director of mortgage scoring solutions for FICO. "I think it's important to educate them as far as the key variables that impact the decision, but actually giving the net present value engine to the consumer is probably not the best approach. It is a fairly complex process and decision for most average homeowners to understand."
Maybe more important, secrecy allows servicers to avoid appeals and objections from borrowers who disagree with the home values and default probabilities embedded in the NPV test.
The federal government has come up with two NPV models that, in their simplest format, can be deployed as spreadsheets. (The Treasury and FHFA developed the NPV model used in the Home Affordable Modification Program, or HAMP, and the FDIC built an NPV model dubbed "Mod in a Box.") Some servicers customize these basic NPV models, transforming them into databases that pull information from many sources. Response Analytics and FICO belong to this modification customization industry.
Gaskin, of FICO, says almost half of servicers use spreadsheets to make modification decisions. Someone manually inputs information into the spreadsheet, which yields a decision. It's a static, relatively primitive method.
The biggest servicers, Gaskin says, have more sophisticated NPV decision engines with "all sorts of data input at the ready," including FICO score, total debt obligations, household income and the probability that the borrower will walk away from the loan. The walk-away probability is based on factors such as location and how much more the borrower owes than the house is currently worth. The data can be refined and updated frequently, based on results of the servicer's past mortgage modification decisions.
FICO and Response Analytics develop borrower profiles. "What you do is define groups of different types of borrowers," says Lippman, of Response Analytics, "so you can infer behavior from others like them."
This analysis includes "psychographic factors," including the depth of the desire to stay in the home. For example, Lippman says, if you tell the servicer that you have deep roots in your hometown, with lots of family and friends, and you don't want to be embarrassed by foreclosure, you will be deemed less likely to redefault.
This likely marks the best opportunity to influence the mortgage servicer's decision. Loan modification applications include a hardship letter, in which you explain why you need a lower monthly house payment. If you want to keep the house, it's important to make this clear in the hardship letter. Explain why you want to keep the house: for example, to avoid the embarrassment of foreclosure, or because you grew up in the house and it has sentimental value.
Create a news alert for "mortgage"