I have seen the future

[djd09]

I'd venture that at least 40-50% of residential GSE/FHA/VA mortgages don't need appraisals. The quantity and quality of data available to collateral underwriters is sufficient (now) to mitigate collateral loss risk for most homes in conforming neighborhoods. And, while condition can be a significant impact on value, the PDR option, I think, gives users adequate control over that unknown as well. And then, if you factor in ANY credit side metrics into the equation - you potentially need even fewer appraisals (e.g. low LTV ratio, significant assets, 800+ credit scores, etc).

I'm not poo-pooing our profession. I'm being realistic. 30 years ago, the internet didn't exist, digital photography didn't exist, MLS and property DNA information was housed at local appraisal districts and RE offices, etc. The users of our services depended on us as much to supply the data they needed as they did our analysis. They no longer need the data - they have access to it now. And the automated models are quickly rising to the credibility levels of traditional appraisals.

Now, I think the folks who choose to brave 3.6 will see a pretty good uptick in volume - I just don't know how long it will last.

the mortgage broker will estimate value without a license or USPAP. what could go wrong

 

[George Hatch]

WRT most residential, multi-family and common non-resid'l property types in the urban and suburban communities, there is no alternate universe where the machine doesn't eventually catch up to the performance level of the appraisers as a group.

And the comparison doesn't hinge on the analytics themselves. It hinges on the quantity and quality of the data. No matter how weird the property is, if there is an adequate number of reasonably similar sales data then the analysis from there is easy. It's when the quantity and quality of the data is inadequate that the analysis becomes more difficult.