link to table of contents previous next

TABLE 8-3 Results of Granger Causality Tests Within Cities (commodity = vegetable oil)

Region City
Dependent
Observations		p Value
for H0:
m /=> s		p Value
for H0:
s /=> m

Conclusion

CentralMoscow1250.0180.055feedback
Bryansk 860.0020.405m = > s
Vladimirn.a.n.a.n.a.n.a.
Ivanovon.a.n.a.n.a.n.a.
Tvern.a.n.a.n.a.n.a.
Kalugan.a.n.a.n.a.n.a.
Kostroman.a.n.a.n.a.n.a.
*Oryol 870.1350.406m = > s
Ryazan1310.0000.001feedback
Smolensk 450.0070.044feedback
Tula 840.0640.001feedback
Novomoskovsk 480.0000.000feedback
Yaroslavln.a.n.a.n.a.n.a.
Rybinskn.a.n.a.n.a.n.a.
Volga*Astrakhan1450.0160.888m = > s
Volgograd1430.0100.017feedback
Kamyshinn.a.n.a.n.a.n.a.
Samara1360.5620.675no causality
Syzran 710.0190.009feedback
Togliattin.a.n.a.n.a.n.a.
Penza 500.0030.000feedback
Saratovn.a.n.a.n.a.n.a.
Balakovon.a.n.a.n.a.n.a.
Ulyanovskn.a.n.a.n.a.n.a.
*Kazan1430.3620.204no causality

NOTES: Summary: At the 20 percent (10 percent) significance level, m = > s in 11 (10) of 13 cases; s = > m in 8 (8) of 13 cases; feedback is present in 8 (8) of 13 cases; and no causality is present in 2 (3) of 13 cases.

Cities with asterisks denote cases in which the null hypothesis of no cointegration could not be rejected. Results for these cities were obtained using differenced data. Otherwise, results were obtained using detrended data. The notation x = > y denotes the inference that the variable x Granger causes y; feedback denotes a case in which x causes y, and y also causes x.

n.a. = not available