Testing the Constancy of Regression Relationships Over Time Using Least Squares Residuals

Abstract

A test of the stability over time of the coefficients of a linear regression model is developed using the cummulative sum (cusum) of squares of the ordinary least squares residuals. An example of the application of the test is given. The power of the test under two systems of coefficient variation is examined using Monte Carlo simulation and compared with the power of the original Brown-Durbin-Evans variant of the test based on recursive residuals.