3 1. You suspect Model 2.1 suffers from heteroscedasticity problem. To test this you obtain the following output presented in Figure 3.1 and 3.2. Conduct these tests at 5% significance levels, outlining your procedure clearly to show if the model has heteroscedasticity problem. Compare the conclusions from both tests for heteroscedasticity, which test is more robust? (40 marks) Figure 3.1 White's test for heteroskedasticity OLS, using observations 1-49 Dependent variable: uhat^2 coefficient std. error t-ratio p-value const 0.325003 0.261433 1.243 0.2212 EDUC 0.00240949 0.0341902 0.07047 0.9442 EXPER 0.00511256 0.0134154 0.3811 0.7052 AGE 0.0179820 0.0108340 1.660 0.1050 sq_EDUC 0.00416545 0.00201270 2.070 0.0452 ** X2_X3 5.70749e-05 0.000940249 0.06070 0.9519 X2_X4 0.000970127 0.000640802 1.514 0.1381 sq_EXPER 0.000326519 0.000374188 0.8726 0.3882 X3_X4 3.40700e-05 0.000301526 0.1130 0.9106 sq_AGE 0.000139715 0.000137919 1.013 0.3173 Unadjusted R-squared = 0.425378 Figure 3.2: Breusch-Pagan test for heteroskedasticity OLS, using observations 1-49 Dependent variable: scaled uhat^2 coefficient std. error t-ratio p-value const 0.176098 0.911987 0.1931 0.8478 EDUC 0.266790 0.0750897 3.553 0.0009 *** EXPER 0.00645373 0.0308108 0.2095 0.8350 AGE 0.0127826 0.0180802 0.7070 0.4832 Explained sum of squares = 20.3879 Unadjusted R-squared = 0.237068326 2. What is autocorrelation? What consequences does it have for the OLS estimators? Outline one procedure that can be used to detect autocorrelation. (30 marks) 3. Figure 3.3 presents the summary statistics of the residuals from Model 2.1. Conduct the Jarque- Bera test, outlining your procedure clearly. What inference can be drawn from the result? (30 marks) Figure 3.3: Summary Statistics, using the observations 1 - 49 for the variable uhat4 (49 valid observations) Mean Median Minimum Maximum 3.0814e-016 -0.049917 -0.45310 0.62512 Std. Dev. C.V. Skewness Ex. kurtosis 0.25644 8.3222e+014 0.40919 -0.28922