Mortgages: August 2012 Archives

It has become a trend for real estate agents who think they're being "smart" to require an automated underwriting approval.

These are automated underwriting programs from Fannie Mae and Freddie Mac saying that Fannie or Freddie will buy the loan providing that everything is precisely as represented. The advantage to automated underwriting is that it will often approve people who might not qualify under manual underwriting rules, but usually due to a particularly stirling credit score Fannie and Freddie will move someone who's marginal to an acceptance. The problem with automated underwriting is that absolutely nothing can change or it is no longer valid.

Let me tell you a true story that has happened to me twice now with different processors. In both cases, I ran automated underwriting on loan and got a full regular approval. Then my processor, for reasons known only to them that neither one of these two women were able to articulate to me, decides to run automated underwriting again on exactly the same refinance and gets a level 3. This is not a good thing. Level 3 acceptance is not the third level up the corporate food chain approving the loan. Think of it like life insurance, where level 3 means you're getting three bumps up the cost ladder because you're a riskier bet for the insurance company. That's what level 3 is. They'll still take you, but they want to charge extra. In each case, it could just as easily moved from "accept" all the way to "caution" (Freddie Mac's code word for "No, we won't buy it") What Level 3 meant in practical terms was that instead of making money on the loan, I lost money but completed the loan anyway because that's good business and the right thing to do for the client who trusted me. However, not every lender follows that business model.

If anything about the assumed scenario changes, automated underwriting that was previously done is useless. The two classics are if the purchase contract is for a little bit more or if the tradeoff in rate and cost gets a little higher rise a tad before they are locked. If the down payment is a couple hundred dollars less, or a slightly lower percentage of the purchase price. If one of the buyer's credit cards lowers the credit limit, resulting in a credit score a couple of points lower, that could trigger a change. The list of possible reasons for a change goes on and on.

There are exceptions and points in the process where as long as something is still within the same basic band of guidelines, you don't have to run automated underwriting again. For instance, if an appraisal for a refinance comes in slightly low but you're still within the same loan to value ratio band, I've funded loans without re-running automated underwriting.

The thing to take away from this is not to put your faith in automated underwriting from Fannie and Freddie. Above the cutoffs for manual underwriting, it is extremely finicky. It can be finicky even below those guidelines, as one of the above mentioned processors found out. Truthfully, if lenders didn't give price breaks for automated underwriting, I wouldn't do it except in those circumstances where the buyer doesn't qualify under manual underwriting rules.

In fact, the real Gold Standard for preliminary approval is manual underwriting. Going through manual underwriting isn't sexy, and it doesn't generate a result that looks like it was Handed Down From On High. "Hey, I put this information into the computer and it said I was approved!" as voices from heaven sing "Hallelujah!" (at least in the mind of that deluded individual). But if a borrower qualifies under manual underwriting rules, then they qualify. Maybe that lender won't give their loan officer that quarter of a discount point for automated underwriting, but they will fund the loan provided everything checks out and there aren't any loanbusters. Somebody will approve it and it will fund.

If there are loanbusters present, automated underwriting won't catch that any better than manual. As a matter of fact, manual underwriting is better at catching loanbusters before it gets that far. If the buyer's ratios are tight and qualification depends upon rates that might not be there tomorrow at a cost they can afford to pay, that shows up quite well under manual underwriting. As a listing agent, if I see someone with a 44.9% debt to income ratio and just barely enough cash to close under the listed assumptions, I know that's a shaky deal at best. Automated underwriting doesn't tell you how close to the line it is, it just tells you the result. Manual underwriting lets you know how resilient the buyer's ability to carry through on the purchase is likely to be if something goes a little bit differently that projected. I don't know about you, but in my experience, transactions where everything goes precisely according to the initial plan are about as common as battle plans that survive contact with the enemy.

(Note however, that the originator of that quote strongly believed in planning the whole campaign out in an extensive and detailed manner beforehand so that when issues happened, he and his officers knew what their options were and were not. As a result, he was the most successful general of his day even if most Americans have never heard of him. While Lee and Grant were mucking about mostly over a small patch of Virginia for years, Helmuth von Moltke the Elder planned and executed two successful winning wars in a single campaign each)

As a listing agent, I will not accept automated underwriting results attesting to the buyer's qualification. I want to know how subject to failure this offer is. As a buyer's agent, I don't write them unless clueless listing agents demand them. The object, after all, is to get the property for the buyer at a price they are willing to pay, and beating the listing agent up on this subject is counterproductive to that, no matter how stupid it is. If you're a seller and want to know how qualified a buyer really is, insist upon seeing the manual underwriting numbers.

Caveat Emptor

Original article here