Multicollinearity and Reduced‐Form Price Equations For Residential Markets: An Evaluation of Alternative Estimation Methods*

Publisher Website
Google Scholar
Abstract
One of the problems that has plagued researchers in their estimation of reduced‐form price equations for specific housing markets has been multicollinearity—the lack of statistical independence of the explanatory variables. This paper evaluates the suitability for structural analysis and prediction of stepwise regression and principal components regression as alternatives to the standard regression model in the estimation of equations with interdependent data. In general, the results of this study indicate that under conditions of multicollinearity principal components regression is the superior estimation technique.

This publication has 0 references indexed in Scilit: