Investigating the role of government size in the impact of institutional variables on income distribution in selected countries

The diversity of economic activities on the one hand and the increasing role of governments to expand public services, social security, economic and social obligations of the state in an effort to achieve economic growth and the equitable distribution of income on the other hand, has transformed the role of institutional components into an important and influential issue. Since institutional economists believe the weakness of the structure and function of institutions is one of the causes of underdevelopment, this study aimed to investigate the effect of the institutional variables of the governance quality index, economic freedom and facilities granted by banks to the private sector on the distribution of income from 2004-2015. Data modeling was performed based on the threshold regression method, which yields more consistent estimates compared with other estimation methods. Later, data stationary was examined using the Levine Lin Cho unit root test. The results show that during the research period, the variables of good governance, facilities given to the private sector and the Gini coefficient are stationary at level, and the variables of government size and economic freedom are stationary in the first differences. In order to estimate the model by using the threshold regression test, considering that in order to specify a nonlinear model, in the first step the linear model should be specified to create a basic model to test the hypothesis that the model is linear. The linear model of feasible generalized least squares (FGLS) has been estimated in the framework of diagnostic tests (correlation, heteroscedasticity, and normal distribution), F Limmer test, and Hausman test as panel data regressions with fixed effects. According to the results, the Ramsey test to model specification is not satisfactory, indicating the error of specifying the linear model, which reinforces the possibility that the nonlinear relation is closer to the actual model. The estimation of the threshold regression model has been calculated By using threshold levels, and threshold effect tests. The values of the threshold levels obtained for the government size variable are 7.85 and 8.16. The obtained threshold levels show that when government size reaches 7.85%, the income distribution function fails and the effect of model variables on income distribution varies. In the next step, the search for the second threshold shows that when the size of the government is 8.16 percent, the way in which the variables of economic freedom, private sector facilities and good governance affect the distribution of income changes again. Also, due to the fact that according to the threshold effect test, threshold values are significant, the effect of the coefficients obtained in the model is affected by the threshold values obtained. Analysis of the results obtained from the estimate of the threshold regression model shows that in the regime of the big government, the variables of the index of quality of governance and economic freedom have the most significant negative effect and the private sector facilities have a significant positive effect on income distribution. The findings of this study confirm the necessity of planning and orientation in the countries under study in order to improve indicators of government size, good governance and economic freedom, regulations, efficiency and effectiveness, and support for the private sector.