Question

Jane is doing a test on the effect of foreign direct investment and unemployment
on the US GDP Growth performance, There are 33 observations.
Growth = 13.00 + 0.456 FDI + 0.277 Unemployment
se: (5.98) (0.132) (0.116)
Durbin-Watson Stat: 1.768
R2: 0.427
Note: The estimated regression is linear model with Growth is GDP growth, FDI is inflow of foreign direct
investment in percentage of GDP, Unemployment is unemployment rate.
a) Test (at 5%) the significance of all regressors from the above model.
b) Explain whether the above model suffered autocorrelation problem (at 5%).
c) What are some limitations of the Durbin-Watson d statistic?

Answer 1

Testing significance of la) all regressors from The model The respective Hylotnesis ho: model is insignificant (that is R²0 hi: mover is significant that is pro fobs R²1(K-1) 1341 (1-R²) Inok 0.427 6.573 133-3 = 0.2135 Sll 110178 0.0191 Fobs 110178 Ecritical F(K-1)(n-K) Fooos, 2,30 = 3.32 Fobs Foritical implies we have evidence of rejecting to the fore model is significant at 5 % Level

(b) Durbin watson Test = 10768 Statistic D-statistic c 2 Then we Then we have respective (1.768) Hypothesis Ho there is tue Autocorrelation NO MI there is tue Auto correlation reject i Accept Ho HO ၂ 0 di 2 (1026) (1065) 10768 D- statistic critical value for x=0.05 k's h=33 3 and di = 1026 du= 1005 D- Statistic 10768 Lies in the region of Accept Ho of No tue Autocorrelation So Here is No Auto correlation
(c) limitations of durbin watson d statistic

•The test is not applicable for regression through origin

•Not applicable if the AR(1) scheme in not followed

•not applicable if model contains lagged value of dependent variable as a aggressor