Forecasting Is a Science
For loss forecasting on subprime mortgages, some organizations rely too heavily on models, while others only rely on its management's intuition. The best loss forecasting comes from some sort of combination of the two, said a presenter from Deloitte & Touche during a presentation from the firm called "Subprime Consumer Finance: Know Your Blind Spots for 2007."
Niall Lynas said that a sound loss forecasting process includes such issues as pricing and underwriting; an allowance for loan and lease losses; the expected loss, calculated using the process established under Basel II; the collections workflow; a portfolio valuation; and management reporting.
But when using a model, it is important to remember that a statistical model is only as good as the data it was built on and many do not account for current conditions in the market. He gave some reasons why this is the case. Until recently there have been very low defaults, plus before 2002 the current broad array of new products such as option adjustable-rate mortgages and negative-amortization mortgages were not common.
Furthermore, nontraditional mortgage portfolios have never been stressed under harsh conditions such as negative house price growth. There is a five-step process in creating an accurate loss forecast. These steps are having clearly defined policies and procedures in place; having an understanding of how changes to business practices, policies, procedures and strategies can impact losses; the use of reliable and relevant data; having a sound statistical methodology to use in developing the forecast; and results and reporting. (c) 2007 Mortgage Servicing News and SourceMedia, Inc. All Rights Reserved. http://www.mortgageservicingnews.com http://www.sourcemedia.com