WE ARE SURROUNDED BY DATA. It seems to be everywhere in the mortgage banking industry, yet some surprises remain for mortgage professionals who might assume that mortgage data is plentiful--particularly when it comes to valuations.
In most cases, automated valuation models (AVMs) provide highly reliable value estimates, but it's important to be cognizant of where the tools have natural weak spots as a result of inherent market characteristics.
Everyone knows that you just enter an address and you can instantly obtain accurate automated values, correct? Not so fast: Some markets are tough to value from an AVM perspective because of marked similarities, marked dissimilarities, their expansive parcel size or their prevailing transaction types. Your AVM provider should be able to discuss their approaches in these scenarios.
In big cities, such as New York or San Francisco, AVMs are challenging from a data perspective because housing units can seem similar and the subtle factors may not be revealed in the data.
Similar square footage, bed and bath count and location may make two units appear identical. But if one is on the 48th floor of a highly desirable building and the other is on the 10th floor of a less-desired property, how do they truly compare when this nuance can't be immediately gleaned from the available data? How does sound bleed through walls in one building versus the other, and is there a doorman on duty 24/7?
It is typical to find addresses with virtually the same data (and across the street from one another) varying widely in value due to factors that rarely show up in AVMs, but which local appraisers will take into account.
In New England, homes built more than a century ago are often located near new construction, therefore square footage and bed/bath count may not be as predictive as in other markets. Has the older structure been modernized, and if so, how do the improvements compare with new construction in terms of the appeal to buyers?
Homes built in 2011 or 1911 could be remarkably different, or remarkably similar depending on the degree of modernization done over the years. With so many varying ages and styles blended together in neighborhoods, a more subjective approach to valuations is required in many places, despite having a wealth of data.
In rural areas, such as those in Montana and Nebraska, large amounts of land come into play, and acreage can limit the number of true comparables available. …