Bob is carrying out a cross-country analysis on the relationship between population and GDP. The following regression is considered: GDP a+apopu + a(popu)+u, (3) where GDP and popu denote the GDP (in million USD) and population (in million) of a given country, respectively. To detect the presence of heteroskedasticity, Bob also ran the following auxiliary regression where u is the residual from regression (3). + popu+(popu)+v, (4) The following information was extracted from the ANOVA table associated with the auxiliary regression (4): Regression R: 0.38 Number of observations: 45 Sum of squared residuals: 6.81 (a) 16 marks) Based on the auxiliary regression (4), carry out a test for heteroskedastic errors in regression (3) at the 5% significance level. Show all your steps. (b) [4 marks] Suppose the errors are heteroskedastic in regression (3). Comment on the accuracy of the OLS estimates of an, and oy and their OLS standard errors. (c) (6 marks) Bob thought that it is reasonable to assume that sd(u popu)=ax popu for the errors in regression (3). Suppose this assumption is correct. Explain how Bob can transform regression (3) so that the regression errors become homoskedastic.