>The difference in rents between a single family house and a multi-story apartment building is "scraps"?
No, but if the land is densely populated and high value the potential stream of rental income from building an average house will be lower than the land taxes and you'll lose money building it.
"Scraps" is what speculating over the value of that housing will get you when the only rents to be derived are from the value you create as a property developer.
>And you couldn't even calculate it using the fact that someone paying money for land indicates that it's under-assessed, because if that was expected to work then no one would ever pay more than $0 as the value of unimproved land
Well, that's the point isn't it? If somebody does pay more than $0 for unimproved land (or land with buildings that are worth $0), that's a data point that you can use to adjust your model.
No, but if the land is densely populated and high value the potential stream of rental income from building an average house will be lower than the land taxes and you'll lose money building it.
"Scraps" is what speculating over the value of that housing will get you when the only rents to be derived are from the value you create as a property developer.
>And you couldn't even calculate it using the fact that someone paying money for land indicates that it's under-assessed, because if that was expected to work then no one would ever pay more than $0 as the value of unimproved land
Well, that's the point isn't it? If somebody does pay more than $0 for unimproved land (or land with buildings that are worth $0), that's a data point that you can use to adjust your model.