Consider the following TSLS model, where Ability is an endogenous independent variable, while Educ and Age are exogenous.

Stage 1 : Ability = 0 + 1 Educ + 2 Age + 3 Expr + 4 ParEduc + 5 KWW + V

Stage 2 : log(Wage) = 0 + 1 Educ + 2 Age + 2 Ability + U

If we use the Sargan test, the rst-stage F test and the Hausman test together to decide whether a TSLS or an OLS would be more appropriate, ll in the following blanks: at the 5% signicance level, the critical value for the Sargan test is blank, and the critical value for the Hausman test is blank;

if we blank (reject or fail to reject) the Sargan test, blank (reject or fail to reject) the rst-stage F test, and blank (reject or fail to reject) the Hausman test, then it is implied that we shall use the TSLS instead of the OLS estimation.