Premium
MULTIPLE REGRESSION ANALYSIS APPLIED TO RESIDENTIAL PROPERTIES
Author(s) -
Gloudemans Robert J.,
Miller Dennis W.
Publication year - 1976
Publication title -
decision sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.238
H-Index - 108
eISSN - 1540-5915
pISSN - 0011-7315
DOI - 10.1111/j.1540-5915.1976.tb00676.x
Subject(s) - variance (accounting) , real estate , econometrics , sample (material) , regression analysis , computer science , regression , residual , value (mathematics) , property (philosophy) , business , economics , statistics , mathematics , accounting , philosophy , chemistry , finance , chromatography , algorithm , epistemology , machine learning
An interesting and promising innovation in the assessment of property for tax purposes is the application of multiple regression analysis. Sales prices are regressed on various housing characteristics, and models developed from sold properties are used to generate value estimates for unsold properties. Two important issues relating to the development of such models are (1) the extent to which real estate markets are stable in terms of structural relationships over time, and (2) the extent to which sales prices reflect actual market values. The first issue is important because it affects how current sales must be before they can be used in model development, as well as the amount of information the assessor must collect and maintain on properties. The second issue is important because it concerns the amount of measurement error in the models and the extent to which the assessor/analyst can unambiguously interpret residual variance. These issues are investigated by comparing regression models developed from a sample of properties in Eugene, Oregon, which sold in each of two well separated time periods.