How Confirmation Bias Shapes the Debate Over Income Inequality

Derek Thompson Apr 12 2012, 4:17 PM ET

Confirmation bias is a weird and powerful thing. Two people can look at the exact same information and come to opposite conclusions. Take, for example, the economic fact that tax revenue has averaged less than 19 percent of GDP for the last few decades. Republicans and Democrats both acknowledge this number to be true. But what does it mean? Everybody thinks they know.

Conservatives know that if government revenue won't reliably exceed 19 percent, the only solution is to cut spending. Liberals know that revenue would exceed 19 percent if we didn't cut taxes every few years. Opposite sides are finding evidence for their opposite positions in same piece of data. Remarkable? Not really. It's just confirmation bias.

I like to call these sort of numbers "Rorschach Statistics" because they reveal so much about the prejudices of the reader. Under this paragraph, I've got a great Rorschach Statistic. Actually, it's a Rorschach Chart. It tracks income growth in America -- from the poorest 20% to the richest 5% -- between 1979 and 2007, according to a few different measures.

Blogging from the American Enterprise Institute, Jim Pethokoukis knows what he sees here: evidence that Obama is wrong about stagnating wages in the lower half of the country. "Obama's inequality argument just utterly collapsed," he proclaims, since the data shows incomes growing 26.4% among the bottom quintile. That would be pretty strong growth in wages, indeed!

But I see a three different points.

First, I see more evidence that market wages have clearly declined for the bottom half of earners. Just look at the far left column. Pre-tax, pre-transfer wages dropped 33% for the bottom fifth. That's horrible. Companies have found ways to replace or negotiate down low-income workers. For me, that statistic is the beating heart of income inequality concerns.

Second, I see evidence that without many of the same government transfers that Pethokoukis and other conservatives hate, income inequality would be much worse. This chart suggests that, in the country Pethakoukis wants to live in, which cuts government aid, income inequality would be much more severe than it is today.

Third, I think the last column isn't a good measure of income equality because it lumps the rising cost of health insurance in with straight-up income transfers. If the government gives me $1,000 in refundable tax credits, that makes me richer. If Medicare or my employer's health insurance pays $10,000 for a surgery that five years ago cost only $1,000, does that make me $9,000 richer? I'm not sure it does.

My purpose isn't to embarrass Jim, but to point out that it's interesting, if predictable, that we've both found confirmation for our old biases in this chart. Still, I think he should acknowledge that his, and Paul Ryan's, and Mitt Romney's vision of America would reduce the kind of income and medical transfer payments that so clearly mitigate income inequality.