Determinants of Economic Growth article review
In the article Determinants of Economic Growth: Will Data Tell? By Antonio Ciccone and Marek Jarocinski, the authors strive to answer the economic question of why in some countries, the income growth is faster than in others. Due to the existence of a wide variety of multiple measures, the economic experts expect to find a non-robust relationship from the variable selections of the growth determinants, thus leading to most of them to discount on the cross-country evidence. To measure this, a group of variables of proxy are employed, and used for the same growth determinant, for example, trade. The authors used data from the Penn World Table (PWT), which has different models that produce differing results. Averaging using the Bayesian Model in the context of the regression representations estimated yn = α + xjnβj + εjn as the regression equation. While trying to find the robust growth determinants which are totally agnostic a priori, Ciccone and Jarocinski found that the data strongly consider only a number of growth determinants. Therefore, some determinants are robust, while some are not. Furthermore, the signs on their estimated coefficient were as they expected, and statistically significant. They proved that minor errors are very sensitive to the income estimates used because they cause substantial differences. In their study research, they concluded that such income growth disparities are due to significant international income estimates errors for such growth empirics. They proved this by comparing the data results from the PWT and World Bank’s World Development Indicators (WDI), which gave significantly different estimates. I strongly agree with the author’s conclusions because different modes of empiric estimates are employed in estimating income growths of different countries. Moreover, different proxy groups are used, meaning more disparity is created. However, I disagree with how this regression-model literature suggests using statistical approaches with prior uncertainty to deal with how variable selections lead to robust findings. This approach involves symmetrically treating all explanatory variables prior to estimation for stronger support from the data. All explanatory variables cannot be symmetrically viewed because each determinant differently affects income growth.