N4

To uncovered the relationship we could use OLS. For that approach to generate unbiased estimates, we need strict exogeneity to hold. That often is not the case. An alternative estimator is IV. What are assumptions we need for IV to produce unbiased/consistent estimates?

Group of answer choices

Relevance (cov(z,x)!=0)

No serial correlation

Guassian (Normal) Standard Errors

Homoskedasticity

Exogeneity (E(u|z)=0)