Can the use of proxy variables solve the omitted variable bias in cases where unobservable characteristics might be important?
If X* is a proxy for X and it can be included in building the model as a proxy for X and this can result in decrease in the omitted variable bias. To eliminate omitted variable bias in the regression, X* needs to be very similar to X implying a high correlation between both. This would make sure the relationship between dependent variable and X or X* is similar.
therefore we can say :
YES, the use of proxy variables can solve the omitted variable bias in cases where unobservable characteristics might be important
Get Answers For Free
Most questions answered within 1 hours.