Is there an income convergence across provinces of Turkey?

In this paper, whether there is a convergence of per capita incomes across Turkish provinces during 2004-2014 period is examined following the availability of per capita incomes of Turkish provinces for this period as of December 2016. Considering that firms and households of different regions within a country tend to have access to similar technologies, share a common central government, therefore have similar institutional setups and legal systems, and they can be assumed to have roughly similar tastes and cultures, absolute income convergence is expected to hold across regions of a country. The results of the nonlinear least squares regression show that the absolute convergence across 81 Turkish provinces, which is estimated as 1.2 per cent per year, is lower that than those estimated for the US states, Japan prefectures, and the members of the EU. However, once a proxy for human capital is included besides the initial level of per capita income as a second explanatory variable in the regression equation, the speed of convergence increases to 2.2 per cent per year. This implies that once the differences in educational attainment across Turkish provinces are held constant, there exists conditional income convergence of similar magnitude found in studies on convergence across regions of other economies. Taking into consideration the evidence that the attitude towards, especially girls' education differs to a large extent between eastern and western regions of Turkey, the same regressions are repeated for 41 western and 40 eastern provinces, separately. The results show that the estimated conditional convergence reaches as high as 4 per cent across eastern provinces, while it is around 2 per cent across western provinces once the human capital is controlled for. This result implies that an effort to change attitude towards education in eastern Turkey is critical in increasing the speed of convergence, thus reducing income inequality across Turkish provinces.