It’s true that choosing a place to live not only determines your housing costs, but also your transportation costs, and so talking about “housing affordability” ought to include both. It’s also true, I guess, that amenities in a given city might lead people to be willing to spend more of their income on housing and transportation than they would in another city, and so having a given percentage of income – 30%, whatever – pegged as the national “affordability” line is not ideal.

But even so, I feel like this Atlantic Cities piece about an Urban Institute “adjustment” to standard calculations of affordability is missing an awful lot. The problem is basically that most housing is not sold at median prices, and most people do not earn median incomes, and so any index that involves just comparing the median home price with median income isn’t super helpful. The whole point, really, is how both of those numbers are distributed, both in a high-low sense and in a geographic sense.

This is where I think Ed Glaeser’s argument that there are actually two affordability crises in the U.S. becomes really important. Glaeser (in Rethinking Federal Housing Policy) says that the two crises are: A) the crisis of supply restriction in elite cities and suburbs that artificially inflates prices above the point where middle-income people can afford them, a crisis which, absent those restrictions, would not exist; and B) the crisis of income among the poor, who can’t afford market-rate housing no matter what. These median-v.-median comparisons might be helpful for illuminating crisis A, but they clearly have nothing to say about crisis B.

But even so, I don’t know that I would use median housing prices, even on crisis A. Why? Think about a city like Chicago. The median housing price sits below a bunch of supply-restricted, highly-desirable neighborhoods, and also above some neighborhoods with rock-bottom prices plus terrible local schools, no quality retail options, poor access to jobs, and high crime. I don’t find it very instructive, from the Having a Decent Place To Live Is a Human Right, Or At Least A High-Priority Issue For Any Decent Government-perspective, to have a bunch of really unacceptable housing – housing that is not just inconvenient somehow, but which, were you to live in it, would actually drastically worsen the life paths of you and your children – dragging down the median housing price. This applies also to analyses that try to compare the number of housing units affordable to low-income workers to the number of low-income workers. What does that housing look like?

I guess what I’m saying is that housing quality matters. Housing textbooks, as far as I’ve seen, make a big deal of the fact that “housing quality” supposedly isn’t an issue any more, because the number of housing units without plumbing or whatever is very, very low now. But it seems that in an urban landscape where ghettoization and economic segregation have skyrocketed at the same time as “housing quality” has vanished as an issue, neighborhood quality has basically taken its place. And is no less problematic; maybe more so. So when the Atlantic Cities declares that they have a new metric that shows that coastal cities are “less unaffordable” than we thought, my response would be that no, they’re almost certainly much more unaffordable than conventional statistics show.

Of course, there is also some value in having an easily-understood  number that allows laypeople – or legislators, or regulators – to quickly gauge housing affordability. I don’t want to demean this work; it’s obviously important. But people who care about this ought to be working towards some way of better conveying affordability metrics that take into account distribution and quality.