Using automated valuation models for real estate valuation

Abstract

Asset valuation is not only a fundamental tool in economic relations but also an integral part of asset utilization, management, and planning for all participants in these relationships. Until now, traditional approaches relying on 100% involvement of appraisers have remained dominant, but with technological advancement, there is an increasing trend toward automation. Automated valuation models offer professional appraisers advantages in reducing costs, improving predictive capabilities, and increasing computational precision, while mathematical calculations and statistical modeling enable the forecasting of real estate market prices based on historical transactions. This research examines the theoretical concepts, historical development, current state, and trends in real estate valuation and automated valuation models. The research has developed a four-stage process mapping that defines the boundaries between appraiser activities and automated operations in developing value predictions for real estate valuation (hereinafter REV). It also determines how to utilize values predicted by automated valuation models (hereinafter AVM). Using the AVM calculations, the appraiser makes decisions on selecting the actual value by applying coefficients that consider the property's specific characteristics and materiality. This research work is valuable for banks, insurance companies, financial institutions, valuation companies, and professional appraisers in preparing information for REV, developing pipeline processes for automated valuation models, and applying the proposed model for calculating real estate values, risk assessment, and calculating tax bases.