UPDATE 9/26: Looks like the TPC has “retracted” the estimates below citing an error “which involved rollover distributions from 401(k)s and similar retirement plans, caused us to significantly overstate the income of some high-income taxpayers and thus understate the tax rates they paid.” I don’t know how much of a difference this will make, but I suspect that the overall distribution will only be moderately effected, and only at the very top. I’ll post the new results when the TPC makes their correction.

***

The table below, adapted from a Tax Policy Center table here, shows the effective federal tax rate people pay in different income categories. (The “effective” rate is simply the total taxes paid divided by income, and will be lower than the statutory marginal rate because there are a variety of deductions, credits, and tax preferences in the code.)

See below as I’ve added some color to the table to help show how rates vary with income levels. Darker reds represent lower effective rates, and darker greens represent higher rates.

Paul Krugman (and others) have used this data to demonstrate that even though, on average, millionaires pay a higher effective rate than the average of those in the middle, there are still many millionaires that pay less than most in the middle class. For example, at least 25 percent of millionaires pay a lower effective rate (12.6 percent or below) than most people making between $40,000 and $50,000 (13.1 percent or higher).

When interpreting the Buffet principle, we need to decide what rate to use to set as a minimum rate for millionaires. For example, do we want millionaires to pay on average more than the middle on average? If so, we’re already there.

Click figure to enlarge

Or do we want to ensure that all millionaires pay more than the middle-class pays on average? If so, then we need to change the tax code so that millionaires pay something like a minimum of 13 percent, if we set the “middle-class” at the $30,000-$75,000 range. This would mean an increase for about a quarter of millionaires.

Or do we want to ensure that all millionaires pay at least as much as just about anyone in the middle class? In this case, the target would be closer to  a 25 percent effective tax rate, increasing taxes on about half of millionaires.

The spirit of the Buffet rule is clearly not the first category – the outrage stems from the fact that some significant fraction of millionaires do indeed pay less than a large share of the middle. This points to policy changes that would indeed increase revenue from many of those at the top (at least a quarter, and perhaps as much as half or more) that have found ways to lower their tax share to levels that are below many in the middle class.