I’ve always thought the best argument for worrying about income inequality as such—which is to say, the best of a bad lot—was that wealth disparities tend to get translated into disparities in political power. I’ve also tended to think the best response was to have a government too limited in its economic power to merit buying, though there’s surely something of a chicken-and-egg problem there. But in any event, that’s the argument Brad Plumer is making over at TNR Online, but he seems to get a lot of different issues muddled together in his hurry to make his point.

First, his measure of how serious inequality is, and how fast it’s growing, centers on the incomes of the top 1 percent of households, while the rest of his analysis looks at how those earning more than $75,000 differ from the rest of us. But that doesn’t just capture the top one percent; if you look at the 2004 breakdown, it includes all of the top quintile and a sizable chunk of the fourth as well. Now, that makes a fairly big difference. Because it’s easy to get worked up about the disproportionate holdings of the ultra-ultra wealthy, but what if the differential political impact kicks in at a far lower level? Meanwhile, Plumer quotes analysts suggesting that legislators are wholly unresponsive to voters in the lowest third of the income distribution. So at least at a first pass, solving this problem via straight inequality reduction would entail serious redistribution affecting a lot more than the top 1 percent.

Brad also notes that the better-off (again, $75k and up) are far more likely to vote than those lower down the distribution. But while he considers a few reasons this might be the case, he neglects the most obvious: Education. We’ve known for a long time that educational attainment is the single strongest predictor of voter turnout, and education is also strongly correlated with income. Now, education, income, and voting don’t always trend together: In 2004, higher incomes were associated with more votes for Bush, while the opposite was true for education. But if we want to explain a pro-voting norm in higher income groups, it seems like a decent hypothesis would be that education is the main causal factor: People in those high income groups are going to have social circles disproportionately composed of folks with a BA or better, where voting and discussion of politics more generally (which affects things like participation in political advocacy groups) are normalized. (A possible point in favor of this is that the effect of education on voting propensity after controlling for income is significantly stronger than the effect of income after controlling for education.) It follows that reducing inequality as such is unlikely to do much to close the voting gap except indirectly, insofar as it boosts educational levels in the lower quintiles. Also, even if we’re focused just on income, this doesn’t really sound like a problem of inequality. That is, there’s no reason to think that people with low incomes and less education are having their turnout depressed magically by higher concentrations of wealth in the top quintile, or that capping all incomes at $200,000 would somehow spontaneously make the less affluent more likely to head to the polls.

None of this is to deny that money matters in politics; obviously it does. But the bulk of the participation-related figures Plumer’s citing seem to be—first—only indirectly correlated with wealth as a side-effect of the relationship between wealth, education, and status—and second—not obviously related to the extent of inequality in a way that should make us worry about faster relative income growth in the upper quintiles. To the extent that this is what Brad wants us to freak out over, it looks like he’s gathered a whole mess of data without bothering a great deal over whether it actually supports the specific point he wants to make.

ADDENDUM: I meant to link this Washington Post article on the effects of McCain-Feingold earlier, but only just found it. In the background of Plumer’s piece is the notion that inequality increases the political influence of those who support conservatives and Repulicans. But there’s a huge range in that top quintile, from what might (oxymoronically?) be called the “moderately wealthy” to the über-rich. The wealthy on the whole—a group we’d still have even in a far more equal America—do indeed lean more Republican. But the really outsized (pre-“reform”) contributions, the ones from people giving $1 million or more to political campaigns, went to Dems 92 percent of the time. If “reducing inequality” (or its effects) meant largely shrinking the variance within that top quintile, it’s not at all obvious the upshot would be good for people who share Brad’s politics.