Two Bad Metrics for a Housing Shortage
wikimedia
At this point it's a cliche to say that we have an affordability crisis, although it's more accurately described as a housing shortage. There are many powerful anecdotes that support this thesis, but unfortunately policymakers lack the metrics to indicate if their housing strategy is working.
In the news media, we see two data series that don't do the job. One is the median single-family home pricein Seattle. This metric accurately measures the misery of someone shopping exclusively, for whatever reason, in the detached home market. It also measures the pleasure of anyone selling out of Seattle. However, it's also an expression of the inevitable: due to geometry, there will be essentially no new detached homes built in Seattle. Barring destruction of jobs, more crime, worse schools, or a long speculative bubble, the long-term price trend will be upwards.
At the other end of the market, a simple thought experiment shows the problem with median rent. Imagine a city with two rental buildings: 100 units at $750 per month and 75 units at $1500. Then, a dreaded luxury developer comes in and turns a parking lot into 75 more units at $2000 each. Even if no one's rent increases, the median rent doubles to $1500. The mean goes up as well. Not only does a rent crisis emerge where none actually exists, but the luxury developer superficially appears to make things worse!
[Zillow Research, which makes a scientifically serious attempt at meaningful statistics, says that "the impact of new high-end units will be minimal when taking the median over all Zillow rent estimates."]
What cities like Seattle really need is a professionalized rental index. Much like formal inflation indices, it would measure the change in the rent of the same units over time, with adjustment for the demolition of existing housing. To understand how this would work, let's return to my fictional city:
The rental index starts at 100 with the two buildings. In the first month, the rent of both buildings goes up 10%, which leaves our index at 110. In the second month, the luxury building opens and existing rents stay flat. Although the new building is now in the index, the score remains 110.
The follwing month, the luxury building raises rent 30%. As this is 30% of the units, it is a 9% increase overall, and the index moves to 110 * 1.09 = 120. Finally, someone tears down the cheap building and puts up 150 units, 75 at $1500 and 75 at $3000, while other rents remains stable. The cheapest 100 of these units are effectively replacing the teardowns, meaning that the rent for these households increases 100% for 75 and 300% for the other 25. That averages out to a 60% increase, bringing the index to 192. A bad series for our city!
There are other things one could do to add more richness - in particular, accounting for unit quality in some way. But we will probably not get good policy until we measure the right thing.