Mortgage Guaranty Insurance Corp. is the latest in the mortgage industry to develop a proprietary "mortgage score" to rate the riskiness of home loans.
The score, based on the performance of 560,000 mortgages in the insurer's portfolio, predicts the likelihood of default in the first four years of the loan's life, MGIC said.
In April, the Milwaukee-based insurer will begin marketing its mortgage scores to lenders. In doing so, it joins the growing ranks of major mortgage players seeking to maintain a competitive edge and keep their customers by providing additional services at little extra cost.
In general, lenders have been slow to sign on, and MGIC does not expect the program to generate significant revenues, according to Stephen L. Blose, vice president of product development and strategic technologies.
MGIC is marketing its mortgage scores to help lenders simplify the process of making home loans and to help them hold lending costs down. The score is based on 12 variables that are available at the very start of the process, when the borrower fills out a loan application.
If mortgage scores are calculated at this stage, lenders can separate the least risky loans and process them quickest, Mr. Blose said. The scores may also be calculated at the underwriting stage, as part of an automated decision.
The statistical technique of scoring, which relies on the actual performance of large loan portfolios to rate future loan performance, is newly popular in the mortgage industry.
Last year, Fannie Mae and Freddie Mac began to require lenders to use credit scores in underwriting mortgages. Credit scores distill borrower creditworthiness to a single number, and have long been used by credit card issuers.
At the same time, the agencies as well as the insurers have developed so-called mortgage scores, which predict mortgage performance more accurately than plain credit scores.
MGIC is using its mortgage score in its quality control efforts, and expects to use it in the underwriting process within six to nine months, Mr. Blose said.