70.2K

Verified Solution

Link Copied!

= 14. Which of the following test results would tell you that the errors in a regression model, Y; = Bo + B1X + t, where t = 1,...,52, are heteroskedastic? a) The variant of the White's test statistic, calculated from the regression 6 = 0.502 + 0.735% + ut (0.163) (0.208) Ob) The variant of the White's test statistic, calculated from the regression &} = 0.502 + 0.735? + ut (0.163) (0.452) = Oc) The Durbin-Watson statistic value of 2.015; d) The Durbin-Watson statistic value of 1.251; O e) None of the above. 15. Given your answer to Q14, which of the regressions below would provide better estimates than OLS? a) Estimate a model with quasi-differenced variables, Y1 = Y; - pyt-1,Xit = X1t pX1.t-1 and X2t = X2t - PX2.t-1 using population autocorrelation coefficient p if it is known or an estimate of p if it is not known; b) Estimate a model with quasi-differenced variables, Yx = Y; VpYt-1 and Xit = X1t - "PX1.t-1 and Xt = X2t - VPX2.t-1 using population autocorrelation coefficient p if it is known or an estimate of p if it is not known; c) Estimate a model with weighted variables "lui lore. 9/yxie and *alyx Xit X a if it is determined that var(t) = 02X1t; Xit Xit xt Melvis ve kalves Xit d) Estimate a model with weighted variables Yt) if it is determined that var(xt) = 02/Xiti * | , . and X24) JX1 e) None of the above

Answer & Explanation Solved by verified expert

See Answer