Overcoming multicollinearity by deducting errors from the dependent variable

A method to overcome multicollinearity, based on deduction of the stochasti c disturbance from the dependent variable, was developed. This method resul ted in a marked increase in the coefficient of determination - R-2 associat ed with the "corrected dependent variable". Simulation results revealed tha t the proposed method yielded good estimation of those stochastic disturban ce. The resulting "corrected" dependent variable may be incorporated into s ome of the existing methods for overcoming multicollinearity. Application o f least-squares regression to the "corrected" dependent variable yields est imators of normal size, and the predictive ability of the derived equation is excellent. (C) 2001 Elsevier Science Ltd. All rights reserved.