Q on Multicollinearity

Does multicollinearity cause the regression results to be biased and inconsistent? I think not, but just want to confirm. Only a misspecification would make the model unreliable? Thanks.

Nope

Does it cause there to be any problem with the model?

JDV - No to the first question only? second question?

As I understand it, with multi collinearity the prediction of the overrall regression is accurate but the standard errors are inflated (increased Type II errors on the significance of the slope coefficients). We cannot draw any conclusions about the impact of the individual independant variables on the dependent variable.

nattyg Wrote: ------------------------------------------------------- > Does multicollinearity cause the regression > results to be biased? Nope > and inconsistent? Nope > I think > not, but just want to confirm. Only a > misspecification would make the model unreliable? Guess it depends on what unreliable means and there a whole bunch of ways to make biased an inconsistent estimators. Like B1 = 2. Always. > Thanks.