ABSTRACT
The problem of multicollinearity arises in econometrics when two of the explanatory variables included in a multiple regression equation are closely related to one another. The problems of multicollinearity tend to increase as additional explanatory variables are included in the regression equation. For this reason it is useful to perform several simple regressions of the dependent variable on each of the explanatory variables in turn, in addition to multiple regressions. Multiple regression formulae applied to the original data are complex and involve a multitude of calculations. Consequently it is preferable to leave multiple regression calculations to computers. The greater the number of explanatory variables included in the regression equation the more complex the ordinary least squares formulae become. The properties and results derived for the simple-regression two-variable case are applied in the case of multiple regressions. Multiple regression models, however, offer more opportunities for testing hypotheses than simple regression models do.
