Joel, this piece is super inspiring. You mention that we should treat ground rents as continuous, not discrete functions. What if we think of the UBI the same way? Redistribute continuous rents to the people that are actively in the city on a second-by-second basis. This way you incentivize people to come to the city and make it valuable on any timescale, even if they're just vising briefly. You wouldn't have any distorting effects around "what is the threshold of residency for getting the UBI each month." Capital could be allocated other ways, of course, but this just seems like the strongest implementation of the UBI that would most directly encourage population growth. And since population size determines the net productivity of cities, this would be the fastest way to grow the city and the net wealth of the residents.
Thank you for the nice words! I really like this approach. I think the hardest thing for me to figure out with this model is how to avoid any type of rent-seeking behavior to sneak back in, and a more continuous approach like that would certainly be a step in that direction. I'm reading some of your posts now and really like the way you are thinking through these topics - we should chat sometime!
Hey one more thought, I was re-reading Vitalik's post on Crypto Cities, and you both have a lot of similarities in how you are thinking about having people share in wealth of the whole city. If you haven't already, check out "experiment 1" in the post: https://vitalik.ca/general/2021/10/31/cities.html
Thanks for the reply! I like the Harberger Tax, but exempting the place of residence. That's interesting.
For the appraisal and cap rate example, the new appraisal would be done annually? And then the current lease owner would have the option to renew?
One thing I can't quite put my finger on is how important the length of a lease is. In a harberger tax system that doesn't matter, but in theory a wealthy person could lock up all the high value property if they are allowed to lease at indefinite lengths. This is somewhat mitigated by them having to pay rent on that high value, but it still seems less than ideal. Any thoughts on lease length?
Haven't heard of Lars' work before, will check it out!
Lars does some fantastic work on LVT/Georgism. Check out his new book “Land is a Big Deal”!
That’s correct. There is probably a discussion around the right cadence of appraisals, but I think annually makes most sense.
I’m not sure the lease length matters if the lease is priced appropriately - that is, they pay close to 100% of the land’s ground rent annually. The problem today isn’t necessarily that individuals hold title perpetually but that they face hardly any opportunity cost for holding land out of use or underused, and are actually punished by the tax code for improving it. That said, if someone is building a multi million dollar building then they’re probably going to insist on a minimum lease term (as they should).
The most important thing is that the lease rate approximates the full ground rent of the land being leased. This is a problem in Amsterdam and I think some in the U.K. where technically the city/government own most of the central land and leases it to individuals on long term ground leases. These leases are like 75+ years in length. The problem is they fix the lease payments so over time all the increased value of the land still accrues to the land lessee/building owner. This is almost functionally equivalent to perpetual title like in the US because the land owner is still capturing most of the incremental value created by the community. The only difference in the lease system is the purchase price of the land is continuous instead of discrete like I point out in the article (so they’re paying over time for use of land rather than all at once at the beginning), but doesn’t really solve the distribution of land value problem.
OK one more detail question... thinking of a developer building a multi million dollar building, would they be able to enter into a lease that guarantees they will have right of first refusal for X years after initial lease? The price wouldn't be locked in (right?), but you'd need some way to give the developer assurance that they could build and recoup the cost. I'm thinking every year the annual rent may change (set by the municipality during this right of first refusal period) and the developer would get to decide if he wanted to pay that price, or let someone else buy the lease. The "developer option" would continue for a set period, and when it ended at it would revert back to the highest bidder. Are you thinking of something like that, or am I off on a tangent?
A lot of this can be thought of as trying to set the market value of the lease itself as close to zero as possible (similar how with a LVT you try to set the buying/selling price of land close to zero). Because if the lease has market value that's the same as the landowner capturing some of that ground rent.
In the case of what you're asking - the leases would be fully transferable. So if a developer wants out he can simply sell the building + transfer the lease to anyone else, just like they would today. In this case they're only selling the building itself and its rents/cash flows and again the lease representing use of the land should be worth close to zero.
Another strategy is to implement a cap on annual rent increases, so ground rent couldn't increase more than 5% per year for the initial 5-10 period, for example. As a developer myself I can anticipate some of what developers will demand/expect, but I'm sure there will be things I'm not thinking of that they will want in the lease. My instinct is the fact that they have no upfront land cost is a really big benefit that can sweeten the deal for them. Another approach is to negotiate extensions in term mid-lease, so let's say on a ten year lease you're five years in. You can go ahead and negotiate/request a five or ten year extension, but then you open yourself up to some of the other terms changing.
Both these variables (rent increase cap and rolling renewals) are just sort of negotiation points. So for the first handful of developers you may offer a really good deal (like lease #1 is 20 years with $1 per year in ground rent, or something). Then as demand grows to build there you start getting closer to market rates for rent, term, etc.
I hope this kind of answers the question but if I missed the mark let me know! Great questions.
I love a lot of things about this proposal. I'm 100% onboard with allowing people to buy a piece of the whole city, but I'm a little stuck on the operational mechanics of the leasing bit. Is the rate of the lease determined by the city or by the market? And is there a limit on how long I can lease the property? Just trying to think through the mechanics!
Nicholas, these are great questions! Early on it has to be set by the city, likely at a subsidized rate for a few years for the first developers.
Then I think you have two options. One is to appraise land value in the city accurately, likely not through individual assessments but through some kind of software. (This is what Lars Doucet & his team are working on to help cities get accurate land value assessments) Then there would be a cap rate applied to that value because we're charging the value over time and not all at once. As an example if a parcel was valued at $1M by the appraisal team/software and our cap rate was set at 7% then the annual ground rent would be $70k per year. This cap rate should be tied to an index but I need to think more about which one. In this example I tied it to prime.
The second option would be even more market based, but is more "experimental". It would be to price the leases using a Harberger or COST tax. So the leaseholder would actually set the value of the land (not building, bc it is not taxed at all) themselves. This would incentivize leaseholders to set the true value of their property - not too low that it could be taken away and not too high they pay a ton of ground rent. I think this could be really fun to experiment with.
As you know, the biggest problem with the Harberger approach is the displacement problem. I think the way around that is to allow folks' primary residence to be leased at a fixed rate, similar to how cities offer homestead property tax exemptions today. So my own home I pay normal ground rent (based on land appraisal) on a long term ground lease, not a self-assessed value. So this approach would only apply to non owner occupied/commercial property like apartment buildings. Because they're held as an investment and then leased to multiple tenants. So if the investment owner changed because they set their lease rate too low, it wouldn't affect the tenants at all.
Again, wonderful questions and please keep them coming if you think of any more!
Joel, this piece is super inspiring. You mention that we should treat ground rents as continuous, not discrete functions. What if we think of the UBI the same way? Redistribute continuous rents to the people that are actively in the city on a second-by-second basis. This way you incentivize people to come to the city and make it valuable on any timescale, even if they're just vising briefly. You wouldn't have any distorting effects around "what is the threshold of residency for getting the UBI each month." Capital could be allocated other ways, of course, but this just seems like the strongest implementation of the UBI that would most directly encourage population growth. And since population size determines the net productivity of cities, this would be the fastest way to grow the city and the net wealth of the residents.
Thank you for the nice words! I really like this approach. I think the hardest thing for me to figure out with this model is how to avoid any type of rent-seeking behavior to sneak back in, and a more continuous approach like that would certainly be a step in that direction. I'm reading some of your posts now and really like the way you are thinking through these topics - we should chat sometime!
Hey one more thought, I was re-reading Vitalik's post on Crypto Cities, and you both have a lot of similarities in how you are thinking about having people share in wealth of the whole city. If you haven't already, check out "experiment 1" in the post: https://vitalik.ca/general/2021/10/31/cities.html
Awesome, thanks for sharing! I had read this when it first came out but revisiting it now is very interesting, certainly a lot of the same thinking
Thanks for the reply! I like the Harberger Tax, but exempting the place of residence. That's interesting.
For the appraisal and cap rate example, the new appraisal would be done annually? And then the current lease owner would have the option to renew?
One thing I can't quite put my finger on is how important the length of a lease is. In a harberger tax system that doesn't matter, but in theory a wealthy person could lock up all the high value property if they are allowed to lease at indefinite lengths. This is somewhat mitigated by them having to pay rent on that high value, but it still seems less than ideal. Any thoughts on lease length?
Haven't heard of Lars' work before, will check it out!
Lars does some fantastic work on LVT/Georgism. Check out his new book “Land is a Big Deal”!
That’s correct. There is probably a discussion around the right cadence of appraisals, but I think annually makes most sense.
I’m not sure the lease length matters if the lease is priced appropriately - that is, they pay close to 100% of the land’s ground rent annually. The problem today isn’t necessarily that individuals hold title perpetually but that they face hardly any opportunity cost for holding land out of use or underused, and are actually punished by the tax code for improving it. That said, if someone is building a multi million dollar building then they’re probably going to insist on a minimum lease term (as they should).
The most important thing is that the lease rate approximates the full ground rent of the land being leased. This is a problem in Amsterdam and I think some in the U.K. where technically the city/government own most of the central land and leases it to individuals on long term ground leases. These leases are like 75+ years in length. The problem is they fix the lease payments so over time all the increased value of the land still accrues to the land lessee/building owner. This is almost functionally equivalent to perpetual title like in the US because the land owner is still capturing most of the incremental value created by the community. The only difference in the lease system is the purchase price of the land is continuous instead of discrete like I point out in the article (so they’re paying over time for use of land rather than all at once at the beginning), but doesn’t really solve the distribution of land value problem.
OK one more detail question... thinking of a developer building a multi million dollar building, would they be able to enter into a lease that guarantees they will have right of first refusal for X years after initial lease? The price wouldn't be locked in (right?), but you'd need some way to give the developer assurance that they could build and recoup the cost. I'm thinking every year the annual rent may change (set by the municipality during this right of first refusal period) and the developer would get to decide if he wanted to pay that price, or let someone else buy the lease. The "developer option" would continue for a set period, and when it ended at it would revert back to the highest bidder. Are you thinking of something like that, or am I off on a tangent?
I love these, keep em' coming!!!
A lot of this can be thought of as trying to set the market value of the lease itself as close to zero as possible (similar how with a LVT you try to set the buying/selling price of land close to zero). Because if the lease has market value that's the same as the landowner capturing some of that ground rent.
In the case of what you're asking - the leases would be fully transferable. So if a developer wants out he can simply sell the building + transfer the lease to anyone else, just like they would today. In this case they're only selling the building itself and its rents/cash flows and again the lease representing use of the land should be worth close to zero.
Another strategy is to implement a cap on annual rent increases, so ground rent couldn't increase more than 5% per year for the initial 5-10 period, for example. As a developer myself I can anticipate some of what developers will demand/expect, but I'm sure there will be things I'm not thinking of that they will want in the lease. My instinct is the fact that they have no upfront land cost is a really big benefit that can sweeten the deal for them. Another approach is to negotiate extensions in term mid-lease, so let's say on a ten year lease you're five years in. You can go ahead and negotiate/request a five or ten year extension, but then you open yourself up to some of the other terms changing.
Both these variables (rent increase cap and rolling renewals) are just sort of negotiation points. So for the first handful of developers you may offer a really good deal (like lease #1 is 20 years with $1 per year in ground rent, or something). Then as demand grows to build there you start getting closer to market rates for rent, term, etc.
I hope this kind of answers the question but if I missed the mark let me know! Great questions.
I love a lot of things about this proposal. I'm 100% onboard with allowing people to buy a piece of the whole city, but I'm a little stuck on the operational mechanics of the leasing bit. Is the rate of the lease determined by the city or by the market? And is there a limit on how long I can lease the property? Just trying to think through the mechanics!
Nicholas, these are great questions! Early on it has to be set by the city, likely at a subsidized rate for a few years for the first developers.
Then I think you have two options. One is to appraise land value in the city accurately, likely not through individual assessments but through some kind of software. (This is what Lars Doucet & his team are working on to help cities get accurate land value assessments) Then there would be a cap rate applied to that value because we're charging the value over time and not all at once. As an example if a parcel was valued at $1M by the appraisal team/software and our cap rate was set at 7% then the annual ground rent would be $70k per year. This cap rate should be tied to an index but I need to think more about which one. In this example I tied it to prime.
The second option would be even more market based, but is more "experimental". It would be to price the leases using a Harberger or COST tax. So the leaseholder would actually set the value of the land (not building, bc it is not taxed at all) themselves. This would incentivize leaseholders to set the true value of their property - not too low that it could be taken away and not too high they pay a ton of ground rent. I think this could be really fun to experiment with.
As you know, the biggest problem with the Harberger approach is the displacement problem. I think the way around that is to allow folks' primary residence to be leased at a fixed rate, similar to how cities offer homestead property tax exemptions today. So my own home I pay normal ground rent (based on land appraisal) on a long term ground lease, not a self-assessed value. So this approach would only apply to non owner occupied/commercial property like apartment buildings. Because they're held as an investment and then leased to multiple tenants. So if the investment owner changed because they set their lease rate too low, it wouldn't affect the tenants at all.
Again, wonderful questions and please keep them coming if you think of any more!