Title Workshop 12 - Exercises.docx ✅

200916/EFM/MEV&AEPage 1 WORKSHOP 12 – ENDOGENEITY & INSTRUMENTAL VARIABLES Goals By the end of this workshop, you should be able to do the following:  Test for endogeneity  Perform IV regression in STATA Tools  You need to bring the following to class  Downloaded copy of wage2.dat  Statistical tables  Ruler  Calculator Dataset variables The following are the variables in the dataset.  wage - monthly earnings  hours - average weekly hours  IQ - IQ score  KWW - knowledge of world work score  educ - years of education  exper - years of work experience  tenure - years with current employer  age - age in years  married - =1 if married  black - =1 if black  south - =1 if live in south  urban - =1 if live in SMSA  sibs - number of siblings  brthord - birth order  meduc - mother's education
200916/EFM/MEV&AEPage 2  feduc - father's education  lwage - natural log of wage Group Work  All of these questions should be completed with others in your table. 1. Write an econometric model showing the effect of years of education on log(wage). 2. Estimate the model in stata and report the results in an equation form. 3. It is thought that years of education is affected by the number of siblings. Write an econometric model showing this relationship. 4. A researcher suspects that years of education is the source of endogeneity. Type the commands below and test the significance for v1 on your paper. If the coefficient of v1 is significant, endogeneity exists. Include a statement testing whether there is endogeneity or not. regress educ sibs predict v1, residuals regress lwage educ v1 test v1 5. If there is endogeneity, perform an instrumental variable regression by typing in the commands below. Report your results in an equation form. ivreg lwage (educ = sibs) 6. Comment in writing how the estimates in part 5 differs with those in part 1. 7. Another researcher suspects that years of education is also the source of endogeneity. However, they believe that birth order should be used as an instrumental variable for years of education. Test for endogeneity by following the commands in part 4. You will need to replace variables and create new ones for the residuals. 8. If endogeneity exists, perform an instrumental variable regression using the commands below. Report your results in an equation form. ivreg lwage (educ = brthord) 9. Write how the estimates in part 1, 5 and 8 differ from one another.
200916/EFM/MEV&AEPage 3