* R. Carter Hill, William E. Griffiths and Guay C. Lim, * Principles of Econometrics, Fourth Edition, Wiley, 2011. * Chapter 7.2 Applying Indicator Variables READ (cps4_small.dat) / names STAT / ALL GENR blfem=black*female * Wage Equation - Table 7.3, page 265 OLS wage educ black female blfem * F-test page 266 TEST TEST black=0 TEST female=0 TEST blfem=0 END * Wage Equation with Regional Indicator Variables * Table 7.4, page 267 OLS wage educ black female blfem south midwest west * Test the null hypothesis of no regional differences (top of page 268). TEST TEST south=0 TEST midwest=0 TEST west=0 END * Chapter 7.2.3 The Chow Test * Test for differences between the wage regressions for the south * and the rest of the country. * Create interaction variables GENR south1=educ*south GENR south2=black*south GENR south3=female*south GENR south4=blfem*south * Model estimation Column (1) Table 7.5, page 269. OLS wage educ black female blfem south south1-south4 * Get coefficient estimates for workers in the south * Column (3) Table 7.5, page 269. * Note: the standard errors are calculated using the error variance * estimate for the full model. TEST constant+south TEST educ+south1 TEST black+south2 TEST female+south3 TEST blfem+south4 * The F-test (Chow test) statistic, page 270. TEST TEST south=0 TEST south1=0 TEST south2=0 TEST south3=0 TEST south4=0 END * Now estimate a separate wage equation for southern workers. * This replicates the standard errors reported in * Column (3) Table 7.5, page 269. SET NOWARNSKIP SKIPIF (south.EQ.0) OLS wage educ black female blfem DELETE SKIP$ * Chapter 7.3 A log-dependent variable * Equation estimation - middle of page 272. GENR lwage=LOG(wage) OLS lwage educ female / LOGLIN * Estimate a percentage differential between male and female workers * top of page 273. TEST 100*(EXP(female)-1) STOP