I endorse Matt's point here, but there's more that's problematic with it and it captures something a bit problematic in some recent research on housing markets. (A thread) https://twitter.com/mattyglesias/status/1317403030732607489

The paper finds that new construction causes prices on high end real estate very near (within 300m) the construction fall in price, but prices on low-end real estate to rise, compared with property slightly further away (300-800m).

It then suggests this is policy relevant. But it is has such an impoverished normative and theoretical world view that it really can’t (or at least shouldn’t) be useful for policy makers...

The reason people have studied this particular question (and there are a number, going in different directions) is because neighborhood groups who oppose new development often say they care very much about affordability in their neighborhood specifically.

And thus the papers test the claims they make in these political discussions.

But there are 3 major problems with the way this study is posed that make it less than useful for anyone not committed to seeing the world exactly as those groups do. I’ll go through these critiques from narrow to broad

(1) The DiD compares the effects close to the new construction to properties in the same neighborhood but further away. But there’s no reason to assume that the amenity value of new construction (prices up) and the supply effect (prices down) operate on the same geographic scale

Prices across the full 800 meters area may be pushed down by supply effects, but amenities may effect only nearby buildings. But this doesn’t mean new construction increases prices, unless by this you mean relative prices near the construction and further away from it.

(2) More theoretically, why care what happens to prices within 300 meters of a new project? I understand why groups near the properties care (they usually want the prices to go up!) BUT ...

the authors don’t even make any argument about why the reader should care about nearby prices independent of broader effects on prices

If new development causes shifts in demand across a city, increasing prices in one place while decreasing them in another, what should I take from this.

The aggregate effects are important but it is not clear why it is relevant that there are increases in one place (within 300m) and decrease in another.

There are reasons to care about affordability and growth at the regional and citywide level. (Much of my career is dedicated to these questionsl) But you’d have to have a very specific normative worldview to think we should always to be to keep prices steady at the block level.

(3) The negative effect that is increasing prices in the paper is new amenities — nicer buildings, better coffee places or whatever. But new amenities are good! People like new amenities, which is why they make prices go up.

When prices go up because of restrictions on supply, it is bad for the same reasons cartels are bad — it is producing monopoly rents and reducing the availability of a product people demand.

But when prices go up because things get nicer — a new park, say — that makes lots of people better off. Owners are clearly better off, renters who stay or come are very likely better off People who are forced to leave if prices go up are worse off.

But it’s not like you can simply assume that new amenities are bad! Were this the case, all investment — better gov’t services for instance — in poor areas would be bad, because it capitalizes into property prices.

And new public investment doesn’t usually increase housing supply, so there’s no countervailing effect -- new public spending would be worse in the world view than new development

Also, unless people are magically appearing from thin air, they are coming from somewhere. The demand shift they talk about necessitates declines elsewhere — people might move from elsewhere in the city, or from outside of the city. What’s the story why this is bad?

Unless outcomes in other neighborhoods matter zero, then you can’t draw the normative conclusions (that the authors hint at but don't quite argue for)

It’s such a weird world view and one that the authors don’t try to justify even a little. You have to argue for why a question is important!

So here’s my suggestion, people. Try to think through what it is you are studying before you try to test something. Why and how is the question relevant? What would different answers mean? A little theory goes a long way here….

Might be of interest to people who I saw commenting on it @ushanti @marketurbanism @evansoltas @danimmergluck @kristoncapps @drschweitzer @yonahfreemark @idothethinking @mattyglesias

Didn't mean to subtweet @tony_damiano. Also might be of interest, @RickHills2