Inequality, redistribution and growth - interrelations and directions

In this thesis I set out to examine the relationships between growth, inequality and redistribution. It has been difficult to disentangle cause and effect definitively in these relationships. Still, the literature is mainly focusing on the effect of inequality and redistribution on growth, keeping the possible reverse relationship out of the picture. Typically, growth will be treated as the endogenous variable placed on the left hand side of a regression model with inequality and/or redistribution entering as explanatory variables. Even though most researchers are fully aware of the possible endogeneity problem arising from this approach, to my knowledge few attempts have been made to systematically sort out the directions of how growth, inequality and redistribution interrelate. The historic lack of reliable cross-country time series data on inequality may have been a constraint. Recent developments in data-access allowing for simultaneous analysis of inequality, redistribution and growth have inspired new studies on how inequality and redistribution affects growth. The data innovations also give new opportunities to reach conclusions regarding the directions of the relationships. In this thesis, I investigate these relationships in a dynamic cross-country perspective with Granger causality tests being conducted inside the framework of a system GMM-model. When treating growth as the dependent variable, testing wether or not the time-series of inequality and redistribution contribute to the explanation of the time-series of growth, the findings are weak. As I flip the coin, however, I find a substantial and highly significant effect of growth on both inequality and redistribution. This challenges the common view that growth is to be treated as the endogenous variable when exploring the relationship between inequality, redistribution and growth. The analysis suggests that the effect of growth on inequality prior to taxes and transfers is substantially higher than the effect on inequality post taxes and transfers. Thus, the increased inequality appears to be partly offset by a redistributive response. This result can be read as a confirmation of the redistribution hypothesis, predicting that more unequal societies will tend to redistribute more