Evaluating Inequities in Property Assessments for Single Detached Homes in Dallas County, Texas

Authors: John Fell*, University of Texas - Dallas
Topics: Spatial Analysis & Modeling, Economic Geography, Quantitative Methods
Keywords: endogeneity, two-stage least squares, hedonic house price, appraised value, model calibration, price elasticity
Session Type: Paper
Day: 4/14/2018
Start / End Time: 8:00 AM / 9:40 AM
Room: Grand Chenier, Sheraton, 5th Floor
Presentation File: No File Uploaded

County appraisal districts are required to accurately estimate fair cash value of homes on which property tax assessments are based. Due to incomplete information, an appraiser faces challenges potentially leading to vertical inequities. Such imbalances are usually evaluated by comparing the appraisal estimate against a subsequent sale price using a log-log regression model to generate an elasticity. With appraised value as the dependent variable, inherent uncertainty is quantified in the error term. Observed sales prices are potentially influenced by uncertainty themselves and correlated with the disturbances of appraised values. These errors lead to endogeneity ,and ultimately, spurious conclusions due to biased elasticity estimates. A common solution to the endogeneity problem is two-stage least squares, where a set of instrumental variables substitutes the endogenous regressor. Initially, a hedonic home price model with its geographic variables and home characteristics provided the set of relevant instruments. Furthermore, the selection of homes under investigation requires careful consideration if they are to reflect fair cash value; for instance, properties linked to foreclosures are unrepresentative. Additionally, since appraisals, which are ceteris paribus predictions, are lagged by one year from sales, adjustments for annual home price index changes are required. Finally, while the model was calibrated on the log-log scale, judicious comparison is required in the original scale with an insightful interpretation of estimated parameter

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login