The release last week of new data from the Institute on Taxation and Economic Policy (ITEP) documenting who pays what in taxes across all 50 states confirmed that North Carolina continues to fall short of ensuring that our state and local tax system doesn’t ask more from middle and low income people.

Despite claims by the architects of North Carolina’s failed tax-cut experiment, policy choices since 2013 have not ensured that middle and low-income taxpayers are paying lower shares of their income in state and local taxes. Instead the richest taxpayers—whose average income is more than $1 million—continue to pay 33 percent less in state and local taxes as a share of their income than taxpayers who have averages incomes annually of $11,000, a threshold that aligns with deep poverty.

A useful way to look at whether incomes in the states are more equal, or less equal, after taxes has been developed by ITEP with the release of their Who Pays data last week. The ITEP Inequality Index compares incomes at various points throughout the income distribution both before and after state and local taxes are collected. The actual calculation involves numerous steps, but the following example helps illustrate the basic idea underpinning the Index.

In North Carolina, before state and local taxes are collected the top 1 percent of taxpayers earn an average income that is 97 times larger than the average income of the state’s poorest 20 percent of residents. Our tax system, which is among the 45 regressive tax codes in the nation, only exacerbates this divide. After state and local taxes are collected, the average after-tax income of North Carolina’s top earners stands at 100 times the size of the average after-tax income of the state’s low income residents. This is the predictable result of charging low-income families a 9.5 percent effective tax rate, while asking high-income families to pay just 6.4 percent of their income in tax.

While state tax codes are not a cure-all for economic inequality, well-designed systems can help lessen the problem while steeply regressive systems only make it worse.