Housing quiz

By Felix Salmon
April 30, 2010
Richard Florida:

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Chart of the day comes from Richard Florida:

HPR_HousingPrice.jpg

He’s re-running the numbers on the rent-vs-buy debate, and the first thing to notice about this chart, if you just look at the distribution along the y-axis, is that there really aren’t all that many cities below David Leonhardt’s cut-off of 20, and there are precious few indeed below Dean Baker’s cut-off of 15. Most of the country, it seems, is still pretty expensive on a rental-ratio basis.

But be careful with those numbers: Florida, here, is comparing house prices in these cities to rents in these cities. But that ratio is always going to be higher than the ratio between the price of any given house and the annual cost of renting it, since houses for sale tend to be bigger and more expensive and located in nicer neighborhoods than houses for rent.

That, in turn, helps to explain at least a little bit of the startling positive correlation we see in the chart: the cities at the top-right are more likely to be ghettoized into expensive neighborhoods where everybody owns and cheap neighborhoods where everybody rents. Meanwhile, I’ll happily hazard a guess that the cities at the bottom-left will see much less in the way of that kind of differentiation.

But I think there’s something real going on in this chart as well, and it’s rational, to boot. Let’s ignore the difference in housing stock between purchase and rental properties for the time being, and create an oversimplified model where prices are set, in a rational market, at the net present value of future rents. Does it make sense that the prospects for future rent increases are much more robust in LA, NY, and SF than they are in Tampa and Atlanta? Of course it does: vibrant urban centers are much better placed, economically speaking, than endless crumbling exurbs.

In turn, what that says to me is that the market does a pretty good job with respect to pricing in future appreciation or depreciation in rents. There’s a clear baseline cluster around the 20 mark, and then an adjustment is made to that number according to the rental outlook from city to city.

Still, the top five cities are clearly outliers and far too expensive: it would be foolish, I think, to buy in any of them. Which raises an important question for Richard: What are the other two cities in that cluster of five up in the top right hand corner?

I’ll drop Richard an email to find out, but in the meantime put your guesses in the comments, and I’ll try and get his publicist to send a copy of his new book to the first people who get the answer right.

Update: Quiz over. It’s Honolulu and San Jose.

23 comments

Comments are closed.