[_delirium]

> Opening space at the high end of a market can have a disproportionate effect on a total market by removing massive amounts of bidding power.

Are there studies on this bit playing out in practice? My impression so far, mostly looking at Manhattan redevelopment, is that building high-end condos typically does not reduce prices for other buildings in the area, but rather raises them: the presence of new high-end housing increases demand in the immediate vicinity by more than the new building's own supply provides, assuming the area (like Manhattan or SF) is rather desirable in general terms to begin with. That may be good if you want to raise the housing values in an area (each building provides a multiplier effect that raises the prices of neighboring ones, making the whole area more desirable). But probably not good if you want to lower the prices. Good statistics either way would be interesting, though: are there cases where opening new high-end housing actually lowered the other prices in an expensive urban area? E.g. when new condo towers open in Manhattan, does it produce an observable downward move in prices?

(The main point I'm getting at is that demand and supply aren't independent variables: demand for living in an area depends quite strongly on the housing stock in that area, so building more housing stock affects both supply and demand, in possibly complex ways, which need empirical data to tease out.)

[jacques_chester]

> But probably not good if you want to lower the prices.

The thing I'm thinking of is the city-wide market, rather than in a particular locale. One way or the other, though, the wealthy will get to live where they want. The best thing you can do is to try and save the furniture by putting them into golden ghettos, rather than having them fan out and drive off people in multiple districts.

> Good statistics either way would be interesting, though: are there cases where opening new high-end housing actually lowered the other prices in an expensive urban area?

Good on you for calling me out on numbers. I'd be interested to see that too.

NYC is a poor case study; you have the confounding factor of widespread rent control. It both creates an artificial limit on housing supply and distorts your official statistics because people have a strong incentive to behave deceptively.

[_delirium]

I agree the wealthy will generally get to live where they want. I think there can be some variation in where they want to live, though, depending on policies. Some wealthy definitely want to live in urban areas, and others definitely want to live in suburban areas, but for others it may depend: for example, if you build shiny new condo towers with modern amenities, there are people who currently wouldn't consider moving into a dilapidated Mission complex who would consider buying one of those new condos and moving up to the city. So you might bring in some new rich people to SF depending on what gets built. The worry some anti-gentrification people seem to have is that the end result will be that the new demand gets soaked up by new rich people, and the only place demand is lowered is not so much a city-wide basis as a metropolitan-area-wide basis: prices go down in outlying suburbs, which is where the poor people then have to move. That's what seems to be happening in NYC: every time a new neighborhood is gentrified, poor people have to move another 15 minutes further away from work, into yet another further-out area with worse transit.

Some people hope to make the city stay unattractive to rich people, so they prefer to live on the Peninsula or North Bay (or in NYC, maybe some nice condo towers on Jersey's "gold coast"). So you still have the golden ghetto, so to speak, but you try to put it elsewhere. Even leaving aside whether that's a good goal, I think it's certainly a good question whether current policy will actually do that; it's quite possible anti-growth policies won't have that effect anyway (e.g. Paris's anti-growth policy has certainly not resulted in a non-gentrified Paris). The interconnections are pretty interesting though, and seem hard to model, especially with the existing data points (e.g. NYC) having, as you point out, their own oddities.