Journal of Property Tax Assessment & Administration
Abstract
As part of the assessment process, assessors are responsible for collecting data about all parcels within their jurisdiction. Whether valuations come from a cost system, a regression model, or another approach, they can be only as good as the underlying data, a concept known as Garbage In, Garbage Out, or GIGO. As such, high-quality data are imperative to ensuring all property is valued fairly and equitably. Using before and after data from the Maricopa County Assessor’s Office desktop review program, this paper illustrates how parcel data quality affects valuation regression models. Improvements in measures of goodness of fit and measures of variable importance and reliability are discussed and deconstructed to show how improved data quality results in less model error and higher confidence in coefficients. As valuation models seek to reflect appraisal theory, this paper also identifies how improved data quality results in more sensible calibrations. Finally, this paper illustrates how poor data on one parcel not only affect that parcel’s value, but adds a value burden to neighboring parcels with correct data as well. In short, better data quality not only provides more accurate valuations for each individual parcel but improves fairness and equity across all parcels, too.
Keywords
Data processing in assessment - Arizona - Maricopa County
Recommended Citation
Jennifer, R. (2021). Garbage in, garbage out : Implications of data quality for valuation models. Journal of Property Tax Assessment & Administration, 18(1). Retrieved from https://researchexchange.iaao.org/jptaa/vol18/iss1/1
