13 May 2011

"The Not-Too-Rich"

National Review Online | Reihan Salam | Blame the Not-Too-Rich - Reihan Salam

It is widely understood that the American middle class has fuzzy boundaries. Many relatively poor and relatively rich people identify as middle-class. In an ambitious 2008 survey, the Pew Research Center found that an extraordinary 53 percent of Americans described themselves as middle class.
Is that really so extraordinary?

Let's say you want to divide things up into three groups. Certainly putting them in three equal bins is one way to do that. But is putting the middle two quarters in one bin, and the upper and lower quarters in the others so unreasonable?

Since we seem to love quintiles so much when it comes to income, I wouldn't even find it that weird to consider the three middle quintiles the "middle class" and lop off the upper and lower quintiles as separate. (At this point it becomes important what options Pew's survey takers were given. Did they have a chance to describe themselves with high or low granularity?)

I wouldn't even find it unusual if 68% of people considered themselves in the middle of some distribution. That would be everyone within a standard deviation of the mean of a Gaussian, after all.

I would find 33%, 50%, 60% and 68% to all be pretty reasonable numbers. (And that's before we consider the cognitive difficulty people always have in assessing their place in an ordering of damn near anything, to say nothing of something which is taboo to talk about politely like your salary.) Before you claim 53% is "extraordinary" you tell me: what would ordinary be?

PS This is actually a decent piece, but it sure gets off to a rocky start with that opening. The final two paragraphs are the best.

PPS There is still a lot of verbal confusion by Salam — and he is certainly not alone in this — between current annual gross income, the summation of income over time, and the summation of income minus consumption over time. Be careful of flipping back and forth between those, especially when considering the slippery notion of "class."

PPPS Another inaccuracy: as much as I dislike the mortgage interest deduction and agree that it benefits mostly the wealthy, it does not "cost the Treasury" anything. Failing to take taxes from people is not a "cost" to the fisc, unless you think that all money ultimately belongs to government and they just deign to let you keep it in certain circumstances. I know I sounds like a stickler for this sort of linguistic thing, but if even conservative people like Salam use language with such collectivist connotations what hope does the overall debate have of staying accurate? I want to get rid of many of the "tax expenditures" that Obama does, I just think the name "tax expenditures" is a bold-faced lie.


  1. Since it seems likely that class follows IQ, it seems reasonable that the middle class consists of those within one SD of the mean ...

    Of course, depending on the type of economy, more or less of the LHS might be pushed into the working class.

    Still, it does not seem unreasonable that 50+% identify as middle class.

  2. "Failing to take taxes from people is not a "cost" to the fisc, unless you think that all money ultimately belongs to government and they just deign to let you keep it in certain circumstances."
    I think all you need for common notions of the word "cost" is that in the counterfactual you would have something that you currently don't. So a thief could say that his bumbling partner cost him a big score they could have gotten if they hadn't tripped the alarm or something.