Econintersect: The IMF has issued an “IMF Staff Discussion Note” entitled “Causes and Consequences of Income Inequality: A Global Perspective“, written by Era Dabla-Norris, Kalpana Kochhar, Nujin Suphaphiphat, Frantisek Ricka and Evridiki Tsounta. This is an exhaustive review of available data for a variety of empirical parameters examined across developed, developing and emerging economies. Factors reviewed include growth rates (GDP), education outcomes, personal income distributions and taxation.
Click for larger image at imgfave.com.
From the executive summary of the report:
First, we show why policymakers need to focus on the poor and the middle class. Earlier IMF work has shown that income inequality matters for growth and its sustainability. Our analysis suggests that the income distribution itself matters for growth as well. Specifically, if the income share of the top 20 percent (the rich) increases, then GDP growth actually declines over the medium term, suggesting that the benefits do not trickle down. In contrast, an increase in the income share of the bottom 20 percent (the poor) is associated with higher GDP growth. The poor and the middle class matter the most for growth via a number of interrelated economic, social, and political channels.
Second, we investigate what explains the divergent trends in inequality developments across advanced economies and EMDCs, with a particular focus on the poor and the middle class. While most existing studies have focused on advanced countries and looked at the drivers of the Gini coefficient and the income of the rich, this study explores a more diverse group of countries and pays particular attention to the income shares of the poor and the middle class – the main engines of growth.
One of the graphs presented for 18 developed economies shows a clear relationship between reduction in top tax brackets and growth of income inequality:
Econintersect has the data estimated from the IMF graph and created a scatter graph with the regression line and R-squared value calculated:
We consider R-squared of 0.60 to be high in the range of fair correlation (correlation coefficient r = 0.77). The boundary between fair and good correlation is 0.8 by our Correlation Concepts guidelines.
Let’s be careful what statements we can make with this result. Here are statements we think are proven by the results:
- It is incorrect to propose that cutting top income tax brackets will help incomes at all levels.
- There is a significant correlation between reduction of top income brackets and increased income inequality.
What cannot be proven based on these results:
- Cutting top income tax brackets increases income inequality. (Correlation does not prove causation – there could be a other variables which “caused” both tax bracket reduction and increased income inequality.)
Note the care that the authors of this paper used in declaring correlations and associations but avoiding the term “caused”. Correlations in and of themselves can disprove possible causations but additional analysis is required to accompany a correlation to prove causation.
Click on the cover page below to read the entire paper: