Smoothing in Commercial Property Valuations: Evidence from Individual Appraisals

Abstract

Research into the appraisal-smoothing phenomenon has primarily focused on methods of adjusting aggregate return series. Based on assumed models of appraiser behaviour, previous research has posited relationships between appraisal-based returns and "true" real estate returns and employs various statistical means to "correct" for the lag. This study offers new insights into the causes and extent of appraisal smoothing using a "bottom up" approach. In contrast to the more traditional "top down" analysis we analyse individual appraisal reports to examine appraisal smoothing caused by systematic appraisal errors. We employ a unique data set to test the validity of the partial adjustment model of disaggregate property valuations that underlies the statistical techniques employed to "unsmooth" aggregate benchmark real estate return indexes. By accessing the raw data used by each appraiser, we can explore how the appraisers use the raw data available at the time of the appraisals. We find significant evidence of an appraisal lag. The degree of lagging varies over time, as both the quantity and quality of information available to appraisers changes. This suggests that lagging is attributable in part to rational partial adjustment on the part of appraisers working with noisy, uncertain information. We also find, however, that there is a behavioral component to lagging, as appraisers valuing the same property in consecutive periods anchor (place relatively higher weight) on their previous appraised values.

JEL Classification: G1, G2, G12