ECONOMETRICS FINAL

List the benefits of omitting a RHS model when trying to address multicollinearity

By omitting a RHS variable, we may reduce multicollinearity and lower standard errors

List a cost of omitting a RHS model when trying to address multicollinearity

Can cause bias in the remaining coefficients and lead to an incorrect model

Signs of multicollinearity prior to running auxiliary regression

• High R²

• One variable is statistically insignificant

• One variable is statistically significant

ei hat formula?

yi-y-hat

y-hat formula

b1+b2xi+bkxk

(if more than 2 variables, extend it to however much there is)

After an auxiliary regression is run, what do we look at to determine multicollinearity?

• If R² is high

• F-hat is high

• T-hat is significant

  Then the variables are highly correlated and could be a source of multicollinearity

Possible sign of multicollinearity (not 100% sure until run auxiliary regression)

• High R²

• One individual variable is significant

• One individual variable is insignificant

Benefit of acquiring additional data and/or new sample to address MC

Getting more data can reduce MC by increasing variation among the explanatory variables which makes estimates more precise

Cost of acquiring additional data and/or new sample to address MC

Data can be costly, and it may not fix MC if patterns don’t change