Stop leaking revenue.

AuthorDeitch, James M.
PositionMANAGEMENT - Mortgage Bankers Association

On a recent flight back from a Mortgage Bankers Association (MBA) event, I was thumbing through cost-per-loan statistics over the past five years. One set of numbers from MBA's Quarterly Mortgage Bankers Performance Report caught my eye. [paragraph] There were four quarters in which the average midsize bank and mortgage banker produced about $300 million in a quarter--the second quarter of 2009, fourth quarter of 2010, first quarter of 2012 and first quarter of 2014. [paragraph] The shocker: Total production expenses jumped from $3,581 per loan in the second quarter of 2009 to $8,025 per loan in the first quarter of 2014. This is an increase in cost of 125 percent. The cost per loan declined to $6,779 as volume increased by 40 percent in the third quarter of 2014--but the cost per loan was still almost twice that of 2009. [paragraph] I happened to be sitting next to an Intel (the semiconductor manufacturer) executive on the flight, and asked him what happened to the cost of a microprocessor from 2009 to 2014. His answer: A dollar's worth of processing in 2009 cost about 9 cents five years later, or a decrease of about 90 percent. [paragraph] As we talked about our industries, the Intel executive stated the obvious.

"Your industry has a productivity problem," he said. "You get 10 times the performance for every dollar spent on computer technology versus 2009, but closing a loan costs twice as much as in 2009."

How should the mortgage industry best tackle these significant productivity issues? That question spawned this article.

The 2015 and 2016 residential mortgage market volume projections are about $1.2 trillion, according to the most recent MBA forecasts. Refinance volume is expected to be down considerably, but purchase volume is expected to rise to about $750 billion.

The next two years may look remarkably like 1999--about $800 billion of purchase business and about $400 billion of refinance volume.

Current loan quality is very high, due to a tight government-sponsored enterprise (GSE)/lender credit box and some degree of self-selection by borrowers (i.e., borrowers know poor credit, lack of documented earnings or lack of stable employment is disqualifying). One might expect that per-unit costs to originate would be falling, not increasing, under these circumstances because only high-credit-quality, fully documented loans are being underwritten. Fewer loans are being closed overall, resulting in higher fixed costs per loan, especially when factoring in higher compliance costs.

The Dodd-Frank Wall Street Reform and Consumer Protection Act has imposed many prescriptive compliance requirements on mortgage bankers, but for many lenders, the loan manufacturing process has not changed.

Many lenders use a business process that was designed years ago in a different regulatory, product and cost environment. The process has been modified to meet the prescriptive requirements, but the basic business process has not changed for most lenders. This observation applies to retail, wholesale and correspondent lending channels.

Retail consumers tell many of us in the mortgage industry that the process needs improvement: Why does it take 45 days to get a mortgage when I can get a car loan or credit card in under an hour? Why do I have to keep supplying more information throughout the process? Why do I feel lost during the process?

Brokers tell us that our communication and integration with their systems is poor.

What can mortgage bankers learn from other industries? Talking to a Southwest Airlines captain on another flight recently (he was deadheading to his duty station) he said, "Southwest Airlines has 6,500 pilots, and every pilot can fly every one of Southwest's aircraft. We fly in pairs, a captain and a first officer. I may fly with nine to 15 different first officers each month, but we know exactly what is expected of each other and the aircraft. Our procedures and operational methods are called crew resource management, where we back each other up and cross-check each...

To continue reading