As fun as it may be to roll rhetorical balls down the lane and topple simple propaganda and otherwise erroneous talking points on the Rhode Island left, the state would profit from a more credulous and balanced discourse. Plainly put, Oswald Krell’s supposed proof that “there is no negative correlation” between “high taxes and percentage of high-end earners” is incorrect on every level of analysis.
On the first level, Krell is wrong about the argument that he’s professing to disprove. Inasmuch as he cites me specifically, I can attest his error without qualification. The argument is not that taxes are the end-all-be-all of people’s decisions about where to live. Krell and friends wish to attack a simplistic strawman because the actual observations of the other side (our side) are unarguable. The argument is that taxes are a factor, and changing them amidst the other factors that people are already considering will have a predictable outcome. If all else remains constant, raising taxes within a state will decrease the population that pays those taxes. Consider an automobile: increasing or decreasing the amount of gas that flows through the engine will affect its speed, even though the results will vary based on other factors, such as the vehicle’s prior speed, the size and type of the engine, the weight of the car, and the terrain that it’s currently traversing.
On the second level, Krell is wrong about the necessary disproof even of his strawman. He ranks the states according to their tax burden and then according to the percentage of tax returns showing income above $200,000. He takes as the hypothesis for disproof that the lowest taxes would instantly equate to the highest percentage of rich residents, but failure to obtain those results hardly disproves correlation. To even get close to so dramatic a demand, he’d have to be much more specific in his rankings. The question wouldn’t be tax burden, but tax burden on those making over $200,000 per year; if a state taxes those households less, but the majority of its citizens, whose income is much lower, more, then its tax burden could be high while its wealthy remain in place. (I’ve long been arguing that this has been happening in Rhode Island, although the tragic solution of the General Assembly and, even more, the state’s progressives is to increase taxes on the wealthy rather than lower taxes on the majority.)
Even more egregious, though, is that Krell takes the extremes of tax burden (the ten highest and the ten lowest), ranks them with respect to rich folks in isolation from the other thirty states, and then raises his arms with a giant “See!” that there’s no correlation. To the contrary, by this foolish methodology, if there’s any intermixing of the states at all, there is correlation. It appears that the only outcome that Krell would accept as evidence of his opposition’s position is if the tax burden ranking and the rich percentage ranking were perfectly inverted. But there are degrees; it’s only a question of the strength of the correlation.

Which brings us to the third level, on which Krell’s assessment of his own data is just wrong. At least in his first attempt at this analysis, he made some attempt to keep all categories limited to quintiles. Simply putting the top and bottom quintiles of states from one ranking in their relative order for another ranking is useless. Consider the table at left, for which I ranked all 50 states using Krell’s sources. (Although, I had different results with three states; given that all the rest are the same and my double-checking of the data for those states, I suspect he miskeyed a couple of numbers on his spreadsheet.) The left column is the ranking by tax burden, with highest at the top; the right column is the ranking by percentage of $200,000 tax returns, again with the highest at the top.

Clearly — and as nobody is denying — wealthy people find the states between Washington, DC, and Boston to be desirable places to live. It makes sense that the state governments whose territory is desirable for reasons that don’t relate directly to taxation (geography, resources, proximity to industrial and financial centers, etc.) can get away with higher taxes. The tax burden becomes a premium that residents are willing to pay. The question that the people of each state must answer for themselves is whether their level of taxation exceeds the theoretical premium that productive people are willing to bear given other attractions of the locality. In Rhode Island’s case, high taxes, combined with poor public services and crumbling infrastructure, constitute a severe restraint on the economy and, therefore, our well-being.
Returning to the table, though, a few interesting observations work very well with the broad theory that I’ve been proposing. The Northeast and California (which amounts, essentially, to our entire region on the West Coast) hold up pretty well in the wealthy column, but high-tax states outside of our region drop precipitously. It fits the general hypothesis to suggest that taxation matters more in the mass of states that have less to distinguish themselves from each other, which describes (e.g.) Ohio and Wisconsin.
The states with no shading, which make up the middle quintile on the tax burden side, distribute with remarkable evenness across the the rich ranking. The orange states appear to have tax burdens well out of proportion to their attractiveness, and they fall away into the bottom three quintiles.. The green states, by contrast, appear to benefit significantly from their low tax burdens, and the light-blue states more or less stand in place while the green states surpass them in percentage of wealthy residents, with three rocketing up the chart.
As I always disclaim, making grand assertions from this data would require much more extensive research and analysis. But for the purposes of online discussion of local policies, it makes for another interesting consideration among the mound of evidence concerning the direction in which Rhode Island should head. And for the purposes of disproving Oswald Krell’s disproof, I’d say that it’s decisive.