Blog

When all required assumptions for the ordinary least square…

When all required assumptions for the ordinary least square (OLS), the coefficient for the independent variable is the best linear unbiased estimator (BLUE). However, if there is an unobserved confounder (U) that is correlated with the dependent variable (Y) and the independent variable (X), the coefficient of X will be biased. In this case, we can use the instrument variable (IV; Z) method to get a consistent estimator for the coefficient of the endogenous independent variable (X). The required conditions for IV (Z) are: IV (Z) is (1)__________ (a. correlated, b. uncorrelated; 2 points) with the endogenous independent variable X. (Hint: Relevance condition) IV (Z) is (2) __________(a. correlated, b. uncorrelated; 1 points) with error terms. (Hint: Exogeneity condition) IV (Z) (3) ____________(a. have, b. do not have; 1 point) a direct path toward the dependent variable Y. (Hint: Exclusion restriction)