Page 68 - Azerbaijan State University of Economics

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

Briefly,the estimated parameters of the model are not stable in case of being the policy changes

[6]. Box-Jenkins (BJ) methodology is known as ARIMA methodology. These models aren`t been

set on the single equation models or simultaneous equation models. Such models are based on

the stochastic features of times series.ARIMA is more suitable model. So, these kind of models

enable to introduce the more reliable prediction [7].

VAR methodology resembles simultaneous-equation modeling wherein we consider

several endogenous variables together. Whereas each endogenous variable is explained by its

lagged, or past, values and the lagged values of all other endogenous variables in the model. So,

as usual there are no exogenous variables in the model [5,p. 837].

In generally we can write the autoregression processes as follows:

(1)

Here, . But, zero mean and constant variance

uncorrelated random error term(white noise). .

is consumer price index at time . This is also called -th order autoregression

processs and noted as .

process is not the only process generated by . In general form processes

can be shown as follows.

(2)

Here, constant. white noise stochastic error term. .

ARIMA models include avtoregressive and moving average terms. Therefore such

models are called the ARIMA models. Stationarity problem may change depending on times

68