ECO 310  Econometrics Dr. Robert Jantzen

Homework 12:  Multicollinearity


1.  Create a data set using Microsoft Excel which contains the following information on consumer spending (C), Income (X2) and Wealth (X3) for 10 individuals.  Click here for tips on how to create the data set and to load it into the EALimdep program.

C      X2    X3
70     80     810
65    100    1009
90    120    1273
95    140    1425
110  160    1633
115  180    1876
120  200    2052
140  220    2201
155  240    2435
150  260    2686

Save the file as a Comma Delimited (*.csv) file.  Then start up the EALimdep program and import the Excel file.

2.  Generate correlation coefficients for the three variables using EALimdep (click here for help).  What do they show?

3.  Run a regression of C on X2 and X3 (click here for help).  Run another regression of C on just X2.  Run a third regression of C on just X3.
     a.  Compare the three regression R squareds.  What do they show?
     b.  Compare the regression coefficients on X2 and X3.  What do they show?
     c.  Conduct appropriate t-tests on the three regression's coefficients.  What do they show?
     d.  Conduct an F test on the first regression including both the X2 and X3 variables.  What does it show?

4.  a.  What does multicollinearity affect estimated regression results?
     b.  How can multicollinearity be "diagnosed?"
     c.  How can multicollinearity be "cured?"