Labor Economics

Suppose that you are working at a think tank after graduating. One of your co-workers estimates the following model using a large sample of individual- level data:

wi= b0+b1femalei+b2IQi+b3agei+b4educi+ei

ui= a0+a1femalei+a2IQi+a3agei+a4educi+i

where w is the wage, female is a binary gender dummy variable, IQ is IQ score (a measure of cognitive ability), age is age, educ is years of schooling, and e and are distributed bi-variate normal. Assume all explanatory variables are exogenous. The wage is only observed

if ui> 0.

(a) Your co-worker has limited knowledge of econometrics, but was able to estimate the above model (with these exact variables, as the equations are written) using the Heckman two-step command in Stata. However, she does not understand what

Stata did. Carefully explain (1) what the Heckman two step method is, (2) what the purpose of it is. In your answer, compare and contrast this methodology to OLS.

(b) Evaluate the econometric approach used by your co-worker. Assume that the model is "correct," in the sense that the equations above are actually the population model