The economic components of a geographic region, such as its economic growth, are influenced by the spillover effect of other regions’ economic components. In addition to these spatial dependences, there are common factors, such as oil price, that induce a correlation among the economic variables of the geographic regions. Thus, it is necessary to distinguish between the dependence created by these factors and spatial dependence; otherwise the role of spatial effects will be biased upward, and creating other problems such as inconsistency of coefficients and violations of some common assumptions. In Spatial econometric models such as spatial auto regression and spatial error models, the common factors are not considered. In this paper, for the first time, the existence of spatial dependence on economic growth of Iranian provinces during 1394-79 is tested and evaluated, using a combination of spatial auto regression models and spatial error models, and a combination of maximum likelihood estimation (MLE) and Common correlated effects (CCE). The CCE method is able to consider the role for common correlated effects. The results show that ignoring Common correlated effects contributes to inaccurate results. Significance of proxy variables used for Common correlated effects and lack of significance of variables related to spatial dependence indicate that existing dependence is due to the influence of Common correlated effects and economic growth data of Iran's geographical regions do not show spatial dependency at least at province’s level.
Madanizadeh S A, Barakchian S M, Pour Jahanbakhsh A. Spatial Dependence with Common Factors:
A Case Study of Growth in Iran's Provinces. qjerp 2020; 27 (92) :45-68 URL: http://qjerp.ir/article-1-2193-en.html