Epistemic status: I’m not an economist. Please look carefully for errors in my reasoning.
Weaknesses of unimproved land value tax
Unimproved land value is hard to assess.
If someone has a lot of wealth stored in land that has been significantly improved, they could bear a larger tax burden than they would be paying under an unimproved land value tax.
Version 1
All real estate is constantly up for auction on a centralized website. If anyone enters a bid on a property, they need to have money stored in a special account tied the website that is ready to be paid to the owner if the owner accepts the bid (and their bids on other properties are immediately recalled if someone’s balance then drops too much). The highest bid on each property is considered the value of that property.
The government offers an open challenge for people to submit computer programs that predict the value of a property given:
the coordinates of the vertices of the polygon forming the boundary of the property (the “property coordinates”)
the property coordinates and the history of property values, going back 10 years, of every other property within a circle that is N times the area of the property itself.
The objective for the program submissions is to minimize is the squared error of predictions. The program must be within a capped file size. Programs that people submit are tested as time goes on, and the ones that do the best are awarded prizes.
People pay tax on the predicted value of their property, using an aggregated version of the best submissions. Unlike the unimproved land value, this value is easy to assess. Improving the value of one’s property does not increase one’s tax burden (unless it increases the value of neighbors’ properties, in which case the increase in the tax burden should still be fairly small.) And the taxable value of the property is closer to the true value than it would be if unimproved land value were taxed.
Problems with Version 1
Two identical neighboring properties could stabilize to different values; the low value one remains low value because it has a higher property tax because the neighbor’s property is higher value, and vice versa.
A speculator could buy up properties in a neighborhood, then sell them to himself at an arbitrarily large price. This could drive up the tax on neighboring properties, and drive down the price of those properties. Then the speculator could buy those properties cheaply, and then sell the properties he bought originally, which should have gone down in price because neighboring properties went down in price, so the property tax is lower now. Then he could sell the cheaply-bought properties at something resembling the appropriate price, now that the property tax will have returned (closer) to normal.
Good predictors might avoid these problems by exploiting the data about the historical prices of the neighboring properties, but they might not succeed perfectly.
Version 2
On the same real estate website, the taxable value of the property is also listed. The tax liability can be bought and sold separately. Bidders place (negative) bids, and the owner of the tax liability may sell to the highest bidder. (Owners of tax liability would be required to take out insurance to limit their liability if they don’t own the property). Certainly, the property and the property tax liability can be bundled together, but the objective for the program submissions is to predict the property value, not the property value plus the (negative) price of the tax liability. The government could provide a standing offer to buy a property alongside its tax liability at $0.
If we don’t like the idea of people having sold off their tax liability, and paying no more taxes, there could be alternative minimum property tax that the owner of the property has to pay if they don’t own the tax liability on that property.
People would not be allowed to short-sell tax liability. Thus, one cannot make a profit from driving up a tax liability with phony inflated-price sales of neighboring properties. The worst one can do is damage competitors who own this liability. But this significantly decreases the size of the set of people for whom this would be profitable, so a ban on this behavior could be feasibly enforced, especially since the programs predicting property value could themselves be used to flag suspicious transactions.
More thoughts
If people are concerned that their tax burden from year to year is uncertain, they can take out insurance against this. An additional tax is levied on properties according to how long it has been since they last had their home inspected. The result of the home inspection is posted on the real estate auction website. At least one picture must be posted of the exterior. If the property is sold, and doesn’t look the same as the pictures, the buyer can sue (so people can’t sell a property immediately after a fire, before bidders have had time to adjust their bids). If the property doesn’t look like the pictures, but hasn’t been sold, a smaller fine is paid to whoever noticed this.
Predicted Land Value Tax: a better tax than an unimproved land value tax
Epistemic status: I’m not an economist. Please look carefully for errors in my reasoning.
Weaknesses of unimproved land value tax
Unimproved land value is hard to assess.
If someone has a lot of wealth stored in land that has been significantly improved, they could bear a larger tax burden than they would be paying under an unimproved land value tax.
Version 1
All real estate is constantly up for auction on a centralized website. If anyone enters a bid on a property, they need to have money stored in a special account tied the website that is ready to be paid to the owner if the owner accepts the bid (and their bids on other properties are immediately recalled if someone’s balance then drops too much). The highest bid on each property is considered the value of that property.
The government offers an open challenge for people to submit computer programs that predict the value of a property given:
the coordinates of the vertices of the polygon forming the boundary of the property (the “property coordinates”)
the property coordinates and the history of property values, going back 10 years, of every other property within a circle that is N times the area of the property itself.
The objective for the program submissions is to minimize is the squared error of predictions. The program must be within a capped file size. Programs that people submit are tested as time goes on, and the ones that do the best are awarded prizes.
People pay tax on the predicted value of their property, using an aggregated version of the best submissions. Unlike the unimproved land value, this value is easy to assess. Improving the value of one’s property does not increase one’s tax burden (unless it increases the value of neighbors’ properties, in which case the increase in the tax burden should still be fairly small.) And the taxable value of the property is closer to the true value than it would be if unimproved land value were taxed.
Problems with Version 1
Two identical neighboring properties could stabilize to different values; the low value one remains low value because it has a higher property tax because the neighbor’s property is higher value, and vice versa.
A speculator could buy up properties in a neighborhood, then sell them to himself at an arbitrarily large price. This could drive up the tax on neighboring properties, and drive down the price of those properties. Then the speculator could buy those properties cheaply, and then sell the properties he bought originally, which should have gone down in price because neighboring properties went down in price, so the property tax is lower now. Then he could sell the cheaply-bought properties at something resembling the appropriate price, now that the property tax will have returned (closer) to normal.
Good predictors might avoid these problems by exploiting the data about the historical prices of the neighboring properties, but they might not succeed perfectly.
Version 2
On the same real estate website, the taxable value of the property is also listed. The tax liability can be bought and sold separately. Bidders place (negative) bids, and the owner of the tax liability may sell to the highest bidder. (Owners of tax liability would be required to take out insurance to limit their liability if they don’t own the property). Certainly, the property and the property tax liability can be bundled together, but the objective for the program submissions is to predict the property value, not the property value plus the (negative) price of the tax liability. The government could provide a standing offer to buy a property alongside its tax liability at $0.
If we don’t like the idea of people having sold off their tax liability, and paying no more taxes, there could be alternative minimum property tax that the owner of the property has to pay if they don’t own the tax liability on that property.
People would not be allowed to short-sell tax liability. Thus, one cannot make a profit from driving up a tax liability with phony inflated-price sales of neighboring properties. The worst one can do is damage competitors who own this liability. But this significantly decreases the size of the set of people for whom this would be profitable, so a ban on this behavior could be feasibly enforced, especially since the programs predicting property value could themselves be used to flag suspicious transactions.
More thoughts
If people are concerned that their tax burden from year to year is uncertain, they can take out insurance against this. An additional tax is levied on properties according to how long it has been since they last had their home inspected. The result of the home inspection is posted on the real estate auction website. At least one picture must be posted of the exterior. If the property is sold, and doesn’t look the same as the pictures, the buyer can sue (so people can’t sell a property immediately after a fire, before bidders have had time to adjust their bids). If the property doesn’t look like the pictures, but hasn’t been sold, a smaller fine is paid to whoever noticed this.