I think you hit the nail on the head by pointing out that “gentrification” has no clear meaning. In NYC I find it’s often used as a vague catch-all by NIMBYs opposed to market-rate construction.
In my mind it’s always been a good thing: bringing people from the suburbs into the city where they can live a more sustainable lifestyle.
I hope you expand on this topic--particularly because it ties into the broken "filtering system" that you've discussed in prior posts.
I think your figure of 20 million homes figure you refer to in your comment is a bit aggressive, although it probably factors in the ongoing impact of population growth. I think that 2-4 million homes added to the metro areas that have the worst housing situations would make a big difference. And by "worst" I mean Boston, S.F., L.A. type areas.
I have a few reasons for going with the aggressive number, besides the initial reason that it matches some basic arithmetic in long term trends.
First, the other estimates you see will not attribute any of the shortage to pre-2008 conditions. But, pre-2008, there had been years of countercyclical migration out of economic centers. Not only was there clearly a supply problem, but it was serious enough to show up in historically anomalous migration flows. There wasn't just a shortage, there was an historically important shortage. Among other things, household size was certainly being affected by it. And it is largely unaccounted for in other estimates.
The other reason is that (per this post: https://kevinerdmann.substack.com/p/who-causes-problematic-housing-demand ) when there is a supply problem, people tend to get morally uptight about different sources of demand. But if we do start building the appropriate amount of homes, there will be smaller household sizes, young adults forming households sooner, people moving out of crowded or unhappy situations, rehousing the homeless, pieds-à-terre, vacation homes, vacancies, more homes for sale, short term rentals, vaguely used semi-occupied units, etc.
Some of those sound great to people. Some of them sound downright offensive to people who are "swimming in the water" of a supply drought. So, if you're trying to get support for your model that says we are x million homes short, it's not very popular to say, "And this model is accurate because we accounted for all the short term rentals and pieds-à-terre that this will lead to." People will respond with "But THAT'S not why we need more homes!"
The thing is, in any functional process of healing, all of those changes will happen, regardless of where they fit in our populist moral models. I think there are many ways this causes traditional models to undercount. So the numbers you see thrown around a lot, for at least a couple reasons, are likely well below the realistic lower bound of the true need.
I think your observation that the problem is becoming national has important implications. One frequent NIMBY argument is that more construction simply draws in more people (more congestion, parking problems, crowded parks etc) without lowering prices. When people are highly mobile, it's probably true that building a few extra apartment buildings will have little effect on rents because the supply-demand balance is national, not local. That is, it seems unlikely to me that we can significantly bring down rents in any one metro area by building only in that area. We have a national supply-demand problem that requires building at a scale to meet national demand. Maybe one determined state can in principle actually do this --- if California, for example, had the will, could we build enough to house the overflow from the entire rest of the country? But one metro region probably cannot, that is, no matter how strong the YIMBYs might be in Mountain View, there's just no way Mountain View can meet the excess demand even from Silicon Valley let alone the entire country.
Maybe you've already done this (sorry, I have not read everything you've written) but I'd be interested in the scale of incremental construction needed to bring rents down to "natural" levels, which I might define as the cost of capital of new construction plus operating expenses (insurance, taxes, maintenance) plus some margin for the owner, leaving out land value on purpose under the assumption that a new unit can use an arbitrarily small amount of land if you are willing to build high enough.
I think you're right at the metro area level. For instance, it seems like New Jersey might have better land use rules than the rest of NYC. You can see it in some price & rent trends across the NYC metro. But, they would have to build a lot to make up for the rest.
On the other hand, while it is true that a single metro can't solve the affordability problem for the whole country, the frictions that limit migration between metros keep growth in any one metro from being too high. Austin is the closest thing we have to a fast growing big metro, but a lot of cities grew at a faster pace decades ago, so I don't think we can say that any single metro is growing at a pace that isn't reasonably manageable.
Here's an old post with some links and my basic estimate of the need. I think, nationwide, normalization of rents would probably be associated with 20 million units or more.
I think you hit the nail on the head by pointing out that “gentrification” has no clear meaning. In NYC I find it’s often used as a vague catch-all by NIMBYs opposed to market-rate construction.
In my mind it’s always been a good thing: bringing people from the suburbs into the city where they can live a more sustainable lifestyle.
I hope you expand on this topic--particularly because it ties into the broken "filtering system" that you've discussed in prior posts.
I think your figure of 20 million homes figure you refer to in your comment is a bit aggressive, although it probably factors in the ongoing impact of population growth. I think that 2-4 million homes added to the metro areas that have the worst housing situations would make a big difference. And by "worst" I mean Boston, S.F., L.A. type areas.
I have a few reasons for going with the aggressive number, besides the initial reason that it matches some basic arithmetic in long term trends.
First, the other estimates you see will not attribute any of the shortage to pre-2008 conditions. But, pre-2008, there had been years of countercyclical migration out of economic centers. Not only was there clearly a supply problem, but it was serious enough to show up in historically anomalous migration flows. There wasn't just a shortage, there was an historically important shortage. Among other things, household size was certainly being affected by it. And it is largely unaccounted for in other estimates.
The other reason is that (per this post: https://kevinerdmann.substack.com/p/who-causes-problematic-housing-demand ) when there is a supply problem, people tend to get morally uptight about different sources of demand. But if we do start building the appropriate amount of homes, there will be smaller household sizes, young adults forming households sooner, people moving out of crowded or unhappy situations, rehousing the homeless, pieds-à-terre, vacation homes, vacancies, more homes for sale, short term rentals, vaguely used semi-occupied units, etc.
Some of those sound great to people. Some of them sound downright offensive to people who are "swimming in the water" of a supply drought. So, if you're trying to get support for your model that says we are x million homes short, it's not very popular to say, "And this model is accurate because we accounted for all the short term rentals and pieds-à-terre that this will lead to." People will respond with "But THAT'S not why we need more homes!"
The thing is, in any functional process of healing, all of those changes will happen, regardless of where they fit in our populist moral models. I think there are many ways this causes traditional models to undercount. So the numbers you see thrown around a lot, for at least a couple reasons, are likely well below the realistic lower bound of the true need.
Hi Kevin,
Thanks for this article.
I think your observation that the problem is becoming national has important implications. One frequent NIMBY argument is that more construction simply draws in more people (more congestion, parking problems, crowded parks etc) without lowering prices. When people are highly mobile, it's probably true that building a few extra apartment buildings will have little effect on rents because the supply-demand balance is national, not local. That is, it seems unlikely to me that we can significantly bring down rents in any one metro area by building only in that area. We have a national supply-demand problem that requires building at a scale to meet national demand. Maybe one determined state can in principle actually do this --- if California, for example, had the will, could we build enough to house the overflow from the entire rest of the country? But one metro region probably cannot, that is, no matter how strong the YIMBYs might be in Mountain View, there's just no way Mountain View can meet the excess demand even from Silicon Valley let alone the entire country.
Maybe you've already done this (sorry, I have not read everything you've written) but I'd be interested in the scale of incremental construction needed to bring rents down to "natural" levels, which I might define as the cost of capital of new construction plus operating expenses (insurance, taxes, maintenance) plus some margin for the owner, leaving out land value on purpose under the assumption that a new unit can use an arbitrarily small amount of land if you are willing to build high enough.
$0.02
-Ken
Nice to hear from you Ken.
I think you're right at the metro area level. For instance, it seems like New Jersey might have better land use rules than the rest of NYC. You can see it in some price & rent trends across the NYC metro. But, they would have to build a lot to make up for the rest.
On the other hand, while it is true that a single metro can't solve the affordability problem for the whole country, the frictions that limit migration between metros keep growth in any one metro from being too high. Austin is the closest thing we have to a fast growing big metro, but a lot of cities grew at a faster pace decades ago, so I don't think we can say that any single metro is growing at a pace that isn't reasonably manageable.
Here's an old post with some links and my basic estimate of the need. I think, nationwide, normalization of rents would probably be associated with 20 million units or more.
https://kevinerdmann.substack.com/p/the-base-case-is-optimistic-homebuilders