The Review of Real Estate Appraisals Using Multiple Regression Analysis

Abstract

A technique is presented that enables review appraisers to test the work of other appraisers. The technique is based upon the simple assumption that the covariance between the adjusted selling price of comparable properties and the property characteristics for which adjustments are properly made is near zero for a given subject property. In a test of the underlying theory of the technique using pooled appraisal reports, the theory is supported. Also, another set of appraisals is pooled to demonstrate how the technique can identify the presence of improper adjustments. Finally, applications and limitations of the technique are discussed.