Block Exogeneity

[qwerty87]

Hello there,

I am trying to determine which variables are exogenous and which are endogenous in a VAR. In order to do so, I set up a VAR which I estimated, and went to VAR Granger Causality/Block Exogeneity Wald Tests. I am however unsure of how to interpret the results; If I have high p-values for a dependent variable, does that mean that the excluded variable is exogenous to the dependent variable? Which is the one I should be looking at, individual p-value or All? If the variable is exogenous according to this test, should it be estimated as such or completely taken out of the VAR?

Any Help is greatly appreciated. Here's an example of what I am getting:


Dependent variable: D(LPG)
Excluded Chi-sq df Prob.

D(INF) 5.057097 5 0.4090
D(R) 5.174518 5 0.3950
D(SP) 1.415238 5 0.9226
D(LDX) 7.411691 5 0.1918
All 20.39006 20 0.4338


Dependent variable: D(INF)

Excluded Chi-sq df Prob.

D(LPG) 11.99482 5 0.0349
D(R) 26.36822 5 0.0001
D(SP) 4.030944 5 0.5450
D(LDX) 4.696067 5 0.4541

All 49.66257 20 0.0002

[mr.ngoc]

Table 5. VECM Granger Causality/ Block Exogeneity Wald Tests
Independent variables
Dependent variables ∆WF VALUE ∆REAL GDP ∆PRICE WF ECT
∆WFVALUE - 1.207463 ( 0.5468) 7.413161** (0.0246) 7.650118 (0.1053)
∆REAL GDP 4.371514 (0.1124) - 0.309538 (0.8566) 5.891249 (0.2074)
∆PRICE WF 12.09127*** (0.0024) 3.188488 (0.2031) - 19.11370*** (0.0007)
Note: above values are statistics, number in parentheses are value of probability
number in parentheses of ECT are t-statistics
***, **, * significant at level 1%, 5%, 10%, respectively.
The results show that, error correction term for cointegrating equation with total value wooden furniture import of Japan from Vietnam (WF VALUE) as a dependent variable. There is no existence for granger causality from real GDP in Japan (REAL GDP) and import-price of wooden furniture (PRICE WF) to total value wooden furniture import of Japan from Vietnam in long-run. Import-price of wooden furniture exhibits an evidence of Granger causality to total value wooden furniture import of Japan from Vietnam in short-run, which is positive throughout. As the results, price plays a important role in wooden furniture import of Japan from Vietnam. Yet there is no evidence for Granger causality running from real GDP in Japan to total value wooden furniture import of Japan from Vietnam in short-term. The result also showed that there is long-run Granger causality relationship running from total value wooden furniture import of Japan from Vietnam and real GDP in Japan to import-price of wooden furniture imported to Japan, which is positive throughout. The real income level rises, so does consumption of goods. A percentage of that increase in consumption will most likely reflect an increased demand for foreign goods (Madura 2005, 34). There is short-run Granger causality running from total value wooden furniture import of Japan from Vietnam to import price of wooden furniture imported to Japan and plays a significant positive role. In related to price and total value, import price also exhibits relationship with total value wooden furniture import of Japan from Vietnam. However, there is no evidence for Granger causality running from real GDP in Japan to import price of wooden furniture imported to Japan in short-run.

ECT=all ( Granger causality in long-run,)
∆WF VALUE ∆REAL GDP ∆PRICE WF (independent variavles) = Granger causality in short-run