Scott Alexander posted an unusually short (for him) post on the nagging question of whether new supply that increases residential density can actually lower housing costs.
If I understand this correctly, then an increase in demand for housing leads to either:
1. An increase in housing supply, followed by an increase in rents for wealthier/denser zip codes, or
2. A constrained housing supply, followed by an increase in housing prices in cheaper/less dense zip codes due to speculative activity from real estate investors.
This is a detailed analysis that shows hard work and incredible dedication. However, it still doesn't explain the positive correlation between density and prices.
Why do denser cities face more supply constraints? Why do denser cities face more out-migration instead of in-migration?
Are there long-term, 2nd order effects, related to but separate from the short--term effects that you mentioned in this article? If we went back in time to Manhattan in the 1940s, and stimied any attempts at increasing density in the area, would prices be higher than they are now?
Could the higher costs reflect higher electricity and water prices, and higher taxes that stem from more difficulty in governing the area (i.e. more powerful construction unions, more complex infrastructure)?
Could the higher costs stem from higher costs of living? The landlords and the maintenance staff need to buy food in the local area, after all.
Could the higher costs reflect higher induced demand? You mentioned that demand-increasing agglomeration effects aren't perfectly in sync with density, but if there was even a weak link between the two, then we would still have a positive feedback loop where density increased demand and demand increased density. Even pro-YIMBY neighborhoods will eventually face supply constraints with the local geography; you can't build infinite housing in a finite space.
Your research provides strong evidence for the YIMBY side, but there are still questions that need to be answered before we can call this settled.
Dave, thanks for the thoughtful questions so that I can clarify these issues.
I think we can divide demand-side price appreciation into 2 categories:
1) Short term, which is mostly what a city like Austin is dealing with today, which broadly push up prices across a city and generally revert when cyclical demand pressures lessen.
2) Long term, which is basically the agglomeration/density affect Scott Alexander is talking about. On a per-square-foot basis, the added costs of dense development might add 50% or more to unit prices, but families that choose to live in dense areas tend to lower their square footage, so on a price/income basis, it's less. Mostly, households choose to live in smaller units rather than pay more, if the entire reason from the higher costs is density/agglomeration.
Then there are supply-side price pressures, which my research helps to identify. Price pressures because of supply constraints are very income-sensitive. They are loaded onto residents with lower incomes. It really doesn't have much to do with speculation. It mostly has to do with the fact that, really, the only way you can get persistent price trends where families are living in units with price/income ratios in the double digits , is through inertia. Costs have been rising on them while they attempted to simply remain in place. They are faced with a choice between displacement and higher costs, and the level at which price/income settles in a city like Los Angeles is determined by how much cost the pre-existing local families are willing to take on before they cry uncle and migrate away. I would turn conventional wisdom on its head here. All of the excess real estate cost is from compromise, not from speculation or irrational fervor. It comes from people resisting change - resisting displacement in a market where housing supply for families like them is effectively shrinking. I go into that a bit in this recent post:
I'm glad that you and Scott Sumner have done such a thorough analysis of this issue, because it's critical to point out that housing isn't some amazing exception to the law of supply and demand. That we've done such a magnificent job of creating distortions in our housing markets is a consequence of so many dead hand policies that many people believe to be good and wise. The Boston metro region started a war on housing affordability several decades prior to the adoption of any formal zoning codes. These early regulations included prohibitions on building heights, prohibitions on triple deckers, race exclusive neighborhoods, and a strangulation of the power of Boston to expand as a city.
The topic of density is particularly complicated because the gradient distribution for any urban area is so broad that using median levels for any type of analysis is practically worthless. A big part of the fat tail of housing density is the result of wealthy towns that enforce large lot zoning requirements and have set aside massive areas of green space for perpetuity. People value these restrictions, and perversely, "best practice" zoning regulations that date from the 1950's create large lot sizes in relatively dense suburbs that are closer to the urban core. For example, 8,000 s.f. lot size requirements will tend to be the minimum in neighborhoods where most of the lots that were established prior to those rules average 5,000 s.f. or less.
Although I tend to be fixated on the situations in older cities in the New England, the false paradox of "more supply raises prices" could be used in any community that is dealing with a project that alters the scale of their neighborhood in any way. I do wonder about those supertalls in Manhattan, however.
Professor Erdmann,
If I understand this correctly, then an increase in demand for housing leads to either:
1. An increase in housing supply, followed by an increase in rents for wealthier/denser zip codes, or
2. A constrained housing supply, followed by an increase in housing prices in cheaper/less dense zip codes due to speculative activity from real estate investors.
This is a detailed analysis that shows hard work and incredible dedication. However, it still doesn't explain the positive correlation between density and prices.
Why do denser cities face more supply constraints? Why do denser cities face more out-migration instead of in-migration?
Are there long-term, 2nd order effects, related to but separate from the short--term effects that you mentioned in this article? If we went back in time to Manhattan in the 1940s, and stimied any attempts at increasing density in the area, would prices be higher than they are now?
Could the higher costs reflect higher electricity and water prices, and higher taxes that stem from more difficulty in governing the area (i.e. more powerful construction unions, more complex infrastructure)?
Could the higher costs stem from higher costs of living? The landlords and the maintenance staff need to buy food in the local area, after all.
Could the higher costs reflect higher induced demand? You mentioned that demand-increasing agglomeration effects aren't perfectly in sync with density, but if there was even a weak link between the two, then we would still have a positive feedback loop where density increased demand and demand increased density. Even pro-YIMBY neighborhoods will eventually face supply constraints with the local geography; you can't build infinite housing in a finite space.
Your research provides strong evidence for the YIMBY side, but there are still questions that need to be answered before we can call this settled.
Dave, thanks for the thoughtful questions so that I can clarify these issues.
I think we can divide demand-side price appreciation into 2 categories:
1) Short term, which is mostly what a city like Austin is dealing with today, which broadly push up prices across a city and generally revert when cyclical demand pressures lessen.
2) Long term, which is basically the agglomeration/density affect Scott Alexander is talking about. On a per-square-foot basis, the added costs of dense development might add 50% or more to unit prices, but families that choose to live in dense areas tend to lower their square footage, so on a price/income basis, it's less. Mostly, households choose to live in smaller units rather than pay more, if the entire reason from the higher costs is density/agglomeration.
Then there are supply-side price pressures, which my research helps to identify. Price pressures because of supply constraints are very income-sensitive. They are loaded onto residents with lower incomes. It really doesn't have much to do with speculation. It mostly has to do with the fact that, really, the only way you can get persistent price trends where families are living in units with price/income ratios in the double digits , is through inertia. Costs have been rising on them while they attempted to simply remain in place. They are faced with a choice between displacement and higher costs, and the level at which price/income settles in a city like Los Angeles is determined by how much cost the pre-existing local families are willing to take on before they cry uncle and migrate away. I would turn conventional wisdom on its head here. All of the excess real estate cost is from compromise, not from speculation or irrational fervor. It comes from people resisting change - resisting displacement in a market where housing supply for families like them is effectively shrinking. I go into that a bit in this recent post:
https://kevinerdmann.substack.com/p/we-need-affordable-housing-not-market
I'm glad that you and Scott Sumner have done such a thorough analysis of this issue, because it's critical to point out that housing isn't some amazing exception to the law of supply and demand. That we've done such a magnificent job of creating distortions in our housing markets is a consequence of so many dead hand policies that many people believe to be good and wise. The Boston metro region started a war on housing affordability several decades prior to the adoption of any formal zoning codes. These early regulations included prohibitions on building heights, prohibitions on triple deckers, race exclusive neighborhoods, and a strangulation of the power of Boston to expand as a city.
The topic of density is particularly complicated because the gradient distribution for any urban area is so broad that using median levels for any type of analysis is practically worthless. A big part of the fat tail of housing density is the result of wealthy towns that enforce large lot zoning requirements and have set aside massive areas of green space for perpetuity. People value these restrictions, and perversely, "best practice" zoning regulations that date from the 1950's create large lot sizes in relatively dense suburbs that are closer to the urban core. For example, 8,000 s.f. lot size requirements will tend to be the minimum in neighborhoods where most of the lots that were established prior to those rules average 5,000 s.f. or less.
Although I tend to be fixated on the situations in older cities in the New England, the false paradox of "more supply raises prices" could be used in any community that is dealing with a project that alters the scale of their neighborhood in any way. I do wonder about those supertalls in Manhattan, however.