World Income Database (WID), based on administrative tax data, and Survey of Consumer Finances (SCF) data exhibit different dynamics of fat-tailed income distributions in the U.S.A. In particular, the former dataset suggests much faster tail divergence (extreme growth in the gap between the wealthy and the rest of the population) than the latter. An influential structural-model-based literature has emerged to explain these fat tailed income distributions, with a focus on the extreme scenario depicted in WID data. Doing so, I review the WID methodoligy of data imputation and extrapolation. I show that the same results obtained in this literature, using continuous time models, can be obtained more easily, and with less computational burden, using discrete time models. I use the latter to replicate the literature's ability to calibrate dynamics of the income distribution tail for WID. However, the model can only be calibrated simultaneously to dynamics of both the body and tail of WID by stipulating different driving parameters for the two parts of the distribution, and the calibrated parameters seem quite extreme. My more parsimonious model can, in fact, simultaneously calibrate income distribution body and tail dynamics for the SCF dataset, which exhibits less extreme divergence. The analysis suggests that WID tail divergence may be exacerbated by the economic theory behind the authors' methodology.
We develop a discrete time heterogeneous model of income distribution to verify wether this class of models can match empirical evidence about the dynamics of inequality in the United States. While most of the existing literature only calibrates models on empirically observed tail of income distribution, we develop a calibration exercise that takes advantage of recently released data about the entire distribution of income in the United States. We find that this model is able to match observed dynamics of top income inequality. Also, we find that fiscal policy seems to have limited impact on the performance of this model. At the same time, we find that this class of simple random growth models that do not account for any form of economic policy changes will work well only in the tail of income distribution. These findings suggest that economic policies, and fiscal policy in particular, probably play an important role explaining the divergence of the top 1% of the population, compared to the middle class of US citizens.
Inequality, Elections and Redistribution: Structural Differences Between North America and Western Europe
The baseline Political Economy of Voting and Inequality says that countries with high levels of inequality will experience high taxation, which, in turn, is detrimental for economic growth. However, this does not happen when we compare data on inequality, taxation and economic growth in the United States, Northern Europe, and Southern Europe. Southern Europe seems to violate the theory, while the United States seems to follow its predictions closely. I develop an overlapping generations model of voting and inequality where individuals decide the general level of redistribution in one economy. Introducing social preferences allows the model to take into account regional differences in voting results and levels of taxation. I find that societies with Utilitarian social preferences will follow the baseline theory of voting and inequality, while those with Rawlsian social preferences will generally violate it. Countries of Northern and Southern Europe exhibit levels of taxation that are compatible with Rawlsian social preferences. This fact is interpreted as the traditionally stronger relevance of social democratic ideals in European political culture.