**6. Econometric methods**

In this study, the degree of stationarity of series are found with Dickey Fuller and Ng -Perron methods. Between series interaction are measured with classic [25] causality test, [26] were analyzed by symmetric latent causality test and [27] asymmetric latent causality test methods. While [27] are developing symmetric and asymmetric implicit causality tests, [25] suggested the analysis which negative and positive shocks can be separated for the cointegration analysis with using the cumulative totals of these shocks. Firstly these series are divided into positive and negative shocks before these causality tests.

If causality relationships between two series such as *y*1*<sup>t</sup>* and *y*2*<sup>t</sup>* series,

$$\mathcal{Y}\_{\mathbf{1}t=Y\_{1,0}} + \sum\_{\mathbf{l}=\mathbf{0}}^{t} \mathbf{e}\_{\mathbf{1}i} \tag{7}$$

Other models will be helping one by one for each dependent variable with lagged independent other variables. When null hypothesis rejected that means there is causality for each taken dependent variable to independent variables. In the Todo yamamato causality test held the extra the lag value is expanded and taken dmax =1

*Relationship between Economic Growth, Unemployment, Inflation and Current Account…*

They [26] refer to this causality analysis performed between the same types of shocks, [27] named this causality test asymmetric causality test performed between different types of shocks. In the study, Unit Root perron test [28] was performed

It was examined by the method of [28]. The MZa and MZt tests are developed from the type of ADF and PP type test whereas the null hypothesis says the variable is not stationary. MSB and MPT tests are KPSS group tests and the null hypothesis

**Table 1** was observed that all of the series were not stationary at level as I(0) but

After the determination of the degree of the level stationary of variables will be used for Granger causality test. In the analysis, causality relationships between the series were first examined by [29] method. The Akaike and Hannan-Quinn information criteria were determined based on VAR analysis. The first differences of the series were used for Granger causality and the results obtained are presented in

**Table 1** results show there is a strong one-way causality relationship between Inflation and Economic growth. The relationship is from inflation to growth ıt means that the inflation is cause of growth as the rejected null hypothesis shows it. Inflation rates are also directly affects the higher economic growth rate in Turkey. The import of the raw materials and semi-finished materials are needed during

*MZa MZt MSB MPT MZa MZt MSB MPT*

�24,12 (�12,42)\*

�21,20 (�12,42)\*

�19,33 (�14,6)\*

�22,56 (�14,6)\*

�31,56 (�12,33)\*

27,45 (�12,33)\*

�24,45 (�12,45)\*

�25,44 (�12,45)\*

�4,54 (�3,11)\*

�4,33 (�3,11)\*

�3,43 (�2,44)\*

�3,66 (�2,44)\*

�3,89 (�2,67)\*

�3,92 (�2,67)\*

�2,44 (�1,16)\*

�2,56 (�1,16)\*

0,45 (0,78)\*

0,34 (0,78)\*

0,64 (0,77)\*

0,67 (0,78)\*

0,22 (0,77)\*

0,50 (0,78)\*

0,33 (0,77)\*

0,34 (0.77)\*

2,34 (3.34)\*

2,15 (3,34)\*

2,11 (3,21)\*

1,77 (3,21)\*

1,56 (1,88)\*

1,42 (1,88)\*

1,58 (2,33)\*

1,37 (2,33)\*

5,54 (21,44)

34,55 (21,44)

22,13 (5,33)

25,45 (5,33)

27,56 (5,13)

31,42 (5,13)

8,55 (3,77)

9,88 (3,77)

*The stationary serial that has at %1 significance critical values. The I(1) all models have trend and constant.*

**Variables** *I*ð Þ **0** *I*ð Þ**1**

0,46 (7,33)

1,42 (7,33)

1,56 (6,33)

�1,57 (6,33)

0.55 (15,7)

0,34 (15,7)

0,21 (9,72)

0.22 (9,74)

when the first differences were taken all variables became stationary as I (1).

and used the k levels with suitable lags levels.

*DOI: http://dx.doi.org/10.5772/intechopen.93833*

refers the series is stationary.

the production effects the economy.

CAD + �2,36

CAD – 1,89

UNE + �4,33

UNE\_ �4,21

INF+ �4,67

INF- �4,68

G+ �8,77

G\_ 7,34

*\**

**295**

**Table 1.**

(18,15)

(18,1)

(�18,7)

(18,7)

(20,3)

(20,4)

(4,22)

(4,24)

*Ng and Perron (2001) unit root test results [28].*

�1,34 (4,76)

�1,67 (4,77)

�2,44 (6,45)

�2,31 (6,46)

�2,44 (11,7)

�1,66 (11,8)

�3,77 (8,33)

�3,59 (8,34)

*The parenthesis shows the %1 significance level of asymptotic critical levels.*

**Table 2**.

and the findings were presented in **Table 1** below.

$$\mathcal{Y}\_{2t=Y\_{2,0}} + \sum\_{i=0}^{t} \mathfrak{e}\_{2i} \tag{8}$$

And the positive shocks are showed,

$$
\boldsymbol{e}\_{1,i}^+ = \max\left(\boldsymbol{e}\_{1,i}, \mathbf{0}\right) \tag{9}
$$

$$
\varepsilon\_{2,i^+} = \max\left(\varepsilon\_{2,i}, 0\right) \tag{10}
$$

The negative shocks are determined:

$$
\epsilon i \mathbf{1}^{\cdot} = \min \left( e\_{1,i}, \mathbf{0} \right) \tag{11}
$$

$$
\varepsilon\_{2,i^{-}} = \min\left(\varepsilon\_{2,i}, \mathbf{0}\right) \tag{12}
$$

The estimated equation will be held in the table with Toda- Yamamato causality;

$$\begin{aligned} cad\_{t} &= \gamma\_{0} + \sum\_{i=1}^{k} a\_{1}cad\_{t-i} + \sum\_{j=k+1}^{k+d\max} a\_{2}cad\_{t-j} + \sum\_{i=1}^{k} a\_{3}G\_{t-i} + \sum\_{j=k+1}^{k+d\max} a\_{4}G\_{t-j} \\ &+ \sum\_{i=1}^{k} a\_{5}\,\textit{u}\,\textit{u}\boldsymbol{e}\_{t-i} + \sum\_{j=k+1}^{k+d\max} a\_{6}\boldsymbol{u}\boldsymbol{n}\boldsymbol{e}\_{t-j} + \sum\_{i=1}^{k} a\_{7}\,\textit{u}\boldsymbol{f}\_{t-i} + \sum\_{j=k+1}^{k+d\max} a\_{2}\boldsymbol{n}\boldsymbol{f}\_{t-j} + \boldsymbol{e}\_{t} \end{aligned} \tag{13}$$
