Page 74 - Azerbaijan State University of Economics

THE JOURNAL OF ECONOMIC SCIENCES: THEORY AND PRACTICE, V.71, # 2, 2014, pp. 66-80

Table.3. Model description
ARIMA Parametr Standardd Akaike Schwarz
structure s Coefficients error P-value criteria criteria

C 100.2926 0.090057 0.0000
AR(1) 1.394135 0.181354 0.0000
ARIMA(2,0,1) 0.388184 1.949043 2.077528
AR(2) -0.586309 0.110206 0.0000
MA(1) -0.783051 0.194569 0.0001

C 100.2836 0.094647 0.0000
AR(1) -0.364031 0.130477 0.0071
AR(2) 0.451225 0.116792 0.0003
AR(3) 0.017287 0.132519 0.8967
ARIMA(6,0,1) 0.476243 1.930067 2.195480
AR(4) -0.060884 0.128596 0.6377
AR(5) -0.323912 0.110715 0.0049
AR(6) -0.222988 0.106125 0.0400
MA(1) 0.999949 0.048372 0.0000

C 100.2928 0.085467 0.0000
AR(1) 1.116979 0.128963 0.0000
AR(2) -0.612542 0.116807 0.0000
AR(3) 0.856034 0.081601 0.0000
AR(4) -1.160392 0.116454 0.0000
ARIMA(5,0,4) 0.581795 1.763970 2.093029
AR(5) 0.271697 0.129597 0.0405
MA(1) -0.779660 0.032446 0.0000
MA(2) 0.295043 0.042696 0.0000
MA(3) -0.767063 0.032021 0.0000
MA(4) 0.925151 0.024526 0.0000

Let`s look at diagnostic test results of ARIMA model. There is heteroscedasticity problem

in the model SARIMA (1,0,1)(0,0,1)x12. Stationarity requirement is provided by

test.(Annex.1.Table.4.)

Table.5. ARIMA model diagnostic test results.
Testlər ARIMA(2,0,1) ARIMA(6,0,1) ARIMA(5,0,4)
F-stat. P-val. F-stat. P-val. F-stat. P-val.
Breusch-Godfrey Serial
Correlation LM Test:
2.157050 0.1240 0.989137 0.3783 1.671701 0.1973
Heteroskedasticity
Test: ARCH

74