Wednesday, November 20, 2013

one more thing about that

UPDATE: Feel free to read this post, but disregard all its numbers and therefore possibly its conclusions. Commenter Erik Hetzner points out fatal problems with the way me and Matt used these data. Two problems: first, "net cost of attendance," the measure Matt used to indicate out of pocket tuition cost, is not the right number. The right number is net tuition. Second, I didn't realize that when the College Board divides families into quartiles they are not evenly sized quartiles, but rather based on arbitrary cutoffs that the CB makes up. This means the calculation I tried to do here is basically impossible with publicly available data, at least as far as I know. Conclusion for now: de omnibus dubitandum.

Just a quick addendum. In that last post, I said you can't argue that making college free would be a windfall for rich kids, because it all depends on how it's paid for. That led to a back and forth in comments and on twitter about how progressive various kinds of state and federal taxes are.

But let's make this simpler and ask: how progressive would a new free-college tax have to be (in a strict bean-counting sense) to make it a better deal  for poor families than the status quo? And the answer is: it wouldn't have to be progressive at all.

That's because the status quo method of college financing, in which poor students have to pay for college but get a discount compared to rich kids, is itself highly regressive, notwithstanding the discount. Let me show you, using some calculations I've done based on Matt B. own numbers:

As you can see, every income group except the richest -- including the poorest -- currently contributes a bigger share of out-of-pocket tuition than their share of total income. What that means is that even if tuition were eliminated and replaced by an absolutely flat income tax levy -- say, a flat 0.5% surcharge for everyone, or a figure slightly higher but with the poorest exempt -- it would still represent a more progressive financing system, in dollars-and-cents terms, than the status quo.

And, I would argue, a more progressive system in lots of other ways too.


  1. Hi Seth,

    I'm not sure I can figure out this chart. But I think that you and Matt Bruenig may be mistaking a division of income into 4 parts (e.g. $0-$32499, $32500-$49990, ...) for quartiles. Have a look at the raw data: It's reasonably close to US household income quartiles but not quite. And probably nowhere near the household income quartiles of who attends college. -Erik

  2. You're right, I've updated the post to basically say ignore this post for now.

    1. Doesn't help that the college board's charts claimed it's quartiles.