Advances in computer technology are expected to cut the cost of making, or “originating,” a mortgage in half, perhaps within the next two years.
Whether any of the savings will find its way to the borrower’s side of the ledger at closing remains to be seen. After all, automation is an extremely costly endeavor.
But because of intense competition among lenders, most industry experts believe financing costs are sure to come down by at least several hundred dollars.
It now costs roughly $2,600 on average to produce a single mortgage. “As a percentage of total cost,” says William Kelvie, chief information officer at the Federal National Mortgage Association, or Fannie Mae, “it costs more to get a mortgage than to take a company public.”
The effort to automate lending procedures is taking place on two fronts: Electronic processing and automated underwriting.
Although many lenders are already communicating electronically within their own operations, they still cannot “talk” externally to such important service providers as mortgage insurance companies, credit bureaus and appraisers.
Consequently, even the most technologically advanced lender often must revert back to a time-consuming, paper-intensive flow that requires constant re-entering, revalidation and reprinting of data.
But the industry is now exploring ways to exchange information via an electronic data interchange, a form of communication that has been used successfully in many other industries to increase efficiency and productivity, reduce costs, shorten the process and improve customer service.
An EDI for the mortgage industry will identify common information and define standards for exchanging the data electronically so that it can move smoothly and efficiently between insurers, appraisers and the like, regardless of what computer program is being used at either end of the exchange.
“Building electronic data interchanges is a significant step toward the goal of paperless transactions, which we anticipate will become the norm in the loan origination and delivery process,” says Duane White of FBS Mortgage, a Minneapolis-based lender which originates billions of dollars in home loans in 46 states.
In one of several industry initiatives to go paperless, FBS, in conjunction with the PMI Mortgage Insurance Co. in San Francisco, is currently testing one of the first EDI systems for processing mortgage insurance transactions.
To the consternation of some, however, both Fannie Mae and its cross-town rival, the Federal Home Loan Mortgage Corp., or Freddie Mac, also are working on cost-cutting technology to originate loans.
Fannie Mae and Freddie Mac operate in the secondary mortgage market, where they buy loans from lenders, package them into securities and sell them to investors on Wall Street. Together, the federally chartered but publicly owned giants probably put their imprint on two-thirds of all mortgages.
Other than to set underwriting guidelines, however, Fannie Mae and Freddie Mac don’t ordinarily get involved in the origination side of the business. And some major players are worried that if the two extremely deep-pocket companies do now, they will have to junk their own major investments in technology.
Their primary concern is that their systems won’t be able to communicate with Fannie’s and/or Freddie’s, but the two quasi-government agencies have promised that the systems they are building will be open rather than proprietary.
To no one’s objection-in fact, to nearly everyone’s pleasure-Fannie Mae and Freddie Mac also are working on automated underwriting systems that will not only reduce costs and speed approvals but perhaps eliminate discriminatory lending as well.
Some lenders and mortgage insurers already use computers to evaluate potential borrowers. But, for the most part, their systems are not terribly sophisticated.
In most cases, they simply look at applications, pass those that fit within certain parameters and kick those that don’t over to a human underwriter. This way, the human doesn’t waste time pouring over the “slam dunks.”
Even when a lender claims to have the technology necessary to “commit” to a mortgage within five minutes, chances are the promise is “subject to” an appraisal and verifications of your income, credit and employment. And, as anyone who has ever bought a house knows, all that can take weeks or even months.
The artificial intelligence system Freddie Mac is developing is far more advanced. It will combine a rule-based system based on such loan standards as housing-costs-to-income and debt-to-income ratios with a statistical modeling system that, based on historical data, will predict the likelihood of default.
Under the system, which is being tested now and could be widely available in 18 months, lenders will be able to sell loans to Freddie Mac within minutes of taking the application.
Just as important, says Peter Maselli, Freddie Mac’s vice president of business re-engineering, the system can also remove any potential bias.
