A novel approach to commercial property valuation

real estate sector is crucial for enhancing competitiveness, as they
provide stakeholders with the accuracy and adaptability needed to
navigate and capitalize in an increasingly complex market landscape.
This paper aims to improve initial valuations provided by real estate
agents by updating their timeliness while also enhancing their
accuracy. Leveraging statistical techniques based on hedonic
regression, it introduces a novel mechanism, successive term
indexing, enabling the reassessment of widely varying commercial
properties with initial assessments that are at least one term old,
where a term typically encompasses one year. The model's novelty
lies in its approach to indexing individual property characteristics
across two successive terms, using a k-means algorithm to categorize
all numerical variables and stepwise selection. The former enables
differing index values for varying sizes, while the latter allows
dynamic evaluation of the significance of regressors in time. Applied
to data from a banking institution, the model showcases strong
predictive accuracy with unique reassessment ratios ranging from -
3.8% to 5.2% for 2022. With its nuanced analysis of market dynamics
and indexing capabilities, the mechanism intersects the elements of
PPIs and AVMs, presenting a significant methodological
advancement and a practical, simple-to-use tool for valuation.