Multicollinearity:

The multiple regression model covered  in Block 1  is  the workhorse of the applied literature in economics. It has been perhaps the single most common tool used for - analysis. Therefore it makes sense  to consider some  of  the problems that often eses when using it. This unit as well as the  two following  'will  take up some of  the more common econometric issues that arise in  this context as well as give you the tools to test for  and  fix some of  these. We take up the  issue of multicallinearity. In a multiple  regression model, we can include many explanatory variables. These explanatory variables are expected  to be unrelated among themselves.  In emperical estimation, however, some of these kariables may be related.

In ordinary least squares model we assume that sample  observations are measured without error, which  is  always not true. When this assumption does not hold, OLS estimators are biasedand  inconsistent. Errors may appear in the measurement of dependent variable, independent  vhable  or both. When there is  error  in dependent variable,  this does  not destroy the unbiased property of the OLS estimators  but the estimated variances are larger than  the case where  there is no such errors of measurement.