Price Adjustment on Higher-Valued Comps Are Frequently Small and Downwardly Rigid
A previous blog reported that appraisals are often identical to or exceed the contract price. In it, the analysis showed that nine out of 10 appraisals were either at the contract price (30 percent) or above it (60 percent). If the appraisal’s role is to justify the loan or the sale price, then this valuation pattern may be quite understandable. However, that is not the institutional role prescribed to the appraisal; rather, it is supposed to provide a credible, independent third-party valuation. What might contribute to appraisal valuations commonly exceeding the subject property’s price?
A review of 750,000 purchase-loan appraisals conducted in 2015 and 2016 showed that 69.1 percent of comps had a price – averaging 12 percent – above the subject property’s price, and 30.9 percent were below at about 6.4 percent on average.  See Figure 1.
While a more frequent use of higher-valued comps could lead to higher appraisals, value adjustments made by the appraiser to the comp in order to make the comp ‘similar to’ the subject property could negate this effect. However, if the adjustments to the comp property’s value are potentially inadequate or inconsistent, then this could partly explain why appraisal valuations tend to be reported above the subject property’s sales price.
In Figure 2, comparable sales are first separated into higher- and lower-priced comps (relative to the subject property’s contract price). Then, these comps are grouped according to their price difference with the subject property: 0-5 percent, 5-10 percent, 10-20 percent, 20-30 percent, 30-40 percent, and more than 40 percent. These six price brackets are intended to provide an indication of potential underlying characteristic differences between the subject property and the comp. The larger the price gap, the more likely there are greater physical and economic dis-similarities between the two, thus requiring a larger price adjustment. The Initial Price Difference percent is the average price difference, and Average Adjustment percent summarizes the average price adjustment received by each group.
There are two distinct patterns. First, when the Average Adjustment percent is compared between higher and lower comps, value adjustments received by lower comps are typically much larger than those received by higher comps. For instance, at a 20-30 percent subject-comp price differential, downward adjustments on higher comps average 8.1 percent, compared to 19.9 percent upward adjustment on lower comps. While a 19.1 percent upward adjustment on lower comps appears to track with the 23.4 percent pre-adjustment price differential, an 8.1 percent downward adjustment on higher comps is much less than the 24.2 percent average price difference pre-adjustment. Similar disparate adjustments between lower and higher comps hold true for other groups.
Second, if one looks at changes in the Average Adjustment percent as the subject-comp price differential widens, one notices another important asymmetry. For instance, moving from a 20-30 percent price difference to 30-40 percent, Average Adjustment on higher comps only increases from 8.1 percent to 10.1 percent. In contrast, adjustment on lower comps goes up correspondingly to the category, from 19.9 percent to 29.2 percent, to reflect potentially widening underlying characteristic differences between the subject and comp properties. A 29.2 percent upward price adjustment on lower comps seems roughly consistent with an initial 30-40 percent price difference, but a 10.1 percent adjustment on higher comps does not.
By themselves, these patterns alone are not sufficient to directly link to the appraisal outcome at the loan level, and more research is needed. However, they do reveal aspects of underlying inconsistency in the appraisal development, for which more sophisticated data analytics techniques are likely better suited. Undoubtedly, as the industry continues to embrace the opportunities that new data and technology have presented, traditional appraisal will likely incorporate big data techniques to position itself as a superior, unique alternative to automated valuation models.
 The years 2015-2016 mark a period of geographically broad and relatively fast appreciation in home prices across the U.S. Appraisers may have selected higher-priced superior comps rather than make a time adjustment of market trends.
 Figure 1 uses net adjustments; although not reported, gross adjustments exhibited similar asymmetrical patterns. While the net adjustment adds up all itemized adjustments to result in an adjusted price, the gross adjustment is used to assess the frequency (as well as size) of itemized adjustments. Gross adjustment sums the absolute value of all adjustments and then divides it by the initial comp price.
© 2017 CoreLogic, Inc. All rights reserved.
[related_posts_content limit="5" title="Related Posts"]