Page 57 - Azerbaijan State University of Economics

THE JOURNAL OF ECONOMIC SCIENCES: THEORY AND PRACTICE, V.74, # 2, 2017, pp. 51 - 63

3.2. Econometric Model and Methodology

We investigated how the indicators of the corruption and public governance affect the total tax
revenues in 35 OECD economies during 2002-2015 periods. Our dependent variable is total tax
revenues as a percent of GDP (TAXREV), while our independent variables are the indicators of
control of corruption (COC), government effectiveness (GE), regulatory quality (RQ), and rule
of law (ROL). Therefore, we consider the following model:

= + + + + + (1)
'
-
'(
2
'(
'(
'(
'(
4
'(
0

First the cross-sectional dependence among the series was examined with Pesaran (2004) LM
CD test, because cross-section dimension of the dataset (N=35) was found to be smaller than
time dimension of the dataset (T=14). Furthermore, the homogeneity of the slope coefficients
was analyzed by adjusted delta tilde test of Pesaran and Yamagata (2008). Then Pesaran (2007)
panel CIPS unit root test was used to examine the stationarity of the series. Later we will benefit
0
from Chow (F) and BP (c ) test to select the estimation method of panel regression. Finally
serial correlation problem will be tested by Wooldridge (2002) autocorrelation test and
heteroskedasticity problem will be tested by Greene (2003) heteroskedasticity test after
estimating the panel regression.

4. Empirical Analysis

4.1. Results of Cross-Sectional Dependence and Homogeneity Tests

We employed Pesaran (2004) LM CD test, because N (=35) was found to be higher than T (=14)
and the test results were displayed in Table 3. The null hypothesis (cross-sectional independence)
was rejected at 1% significance level, and we revealed a cross-sectional dependence among the
variables. Furthermore, the homogeneity of the slope coefficients was analyzed with adjusted
delta tilde test of Pesaran and Yamagata (2008) and the results were presented in Table 3. The
null hypothesis was rejected and the slope coefficients were found to be heterogeneous.

Table 3: Results of Cross-Sectional Dependence and Homogeneity Tests
Cross-sectional dependency tests
Test Test statistic p-value
LM (Breusch and Pagan (1980)) 887,5 0,0000
LM adj* (Pesaran and Yamagata, (2008)) 7,935 0,0000
LM CD* (Pesaran (2004)) 7,048 0,0000
Homogeneity Tests
Test Test statistic p-value
Delta_tilde 19,674 0,001
Delta_tilde_adj 21,445 0,015
*two-sided test

57