Using Latent Root Regression to Identify Nonpredictive Collinearity in Statistical Appraisal Models

The objective of this study is to examine the potential benefits of using latent root regression techniques to improve the stability of appraisal coefficients (hedonic prices) over time. Another related objective is to more clearly identify the nature and implications of collinearity present in most appraisal models. The results indicate that the majority of the collinearity present in the data is of a predictive nature, hence latent root techniques will likely show little or no improvement over OLS models. Copyright American Real Estate and Urban Economics Association.