5.1.2. The High-Income States Display the Widest Income Gaps

The main idea is to oppose two series: 1) the level of state average income per household, and 2) the size of the state inequality gap. More precisely, the state average income per household corresponds, like earlier in this thesis, to the BEA state personal income adjusted to Piketty and Saez’s series. The inequality gap is measured by the inter-state ratio of the top percentile (derived from IRS tables), i.e. the departure of the state top incomes (top 1 percent) from the national average. The first series is ordered in descending sequence, and then partitioned into 3 groups of 17 states each: high-, medium-, and low-income states. Similarly, the second series partitions states of wide, moderate, and narrow income gaps. How many of the poorest states are featured by high inequality? Kuznets predicted the reverse correlation.

Figure 5.1 illustrate the results for selected years, with negative correlation (left panel) and positive correlation (right panel):

It is evident in Figure 5.1 that there are many more states where a narrow income gap is associated with a low level of income (and sharp inequality with high income levels) than states recording a low income average along with a wide income gap (and vice-versa for ‘high-narrow’). Moreover, the state count in either case barely varies over time. The standard deviation for the negative and positive correlations is 1.9 and 3.7, respectively. This result contradicts Kuznets hypothesis, but prior to drawing such a conclusion, we need to focus on how state growth rates evolved with inter-state inequality, measured in the next section by the ratio of the income earned by the top 1 percent relative to the national average income.