Epistemic status: this probably would be a bad idea, but has some really nice theoretical properties that I hope somebody can find a way to preserve whilst making this a good idea.
I find both Georgian land taxes and Harberger taxes to be fascinating proposals to improve the way we tax land. I’ll start off with a brief overview of both, then discuss benefits and advantages of both, suggest a model to combine some of the benefits of both, and analyze the combined approach. Finally I suggest areas for future research.
The key principle of Georgian land taxes is you tax natural resources at a rate such that their market price is 0.
In other words, if I own a piece of land, and build a house on it, then the government taxes the value of the land, but not the value of the house—so two identical pieces of land in the same area with different buildings on them get taxed at the same rate.
How high is that rate? At exactly the rate such that if I’d built nothing on the land I would be neutral between:
Giving it away to someone else for free so that he will pay the taxes on it instead of me.
Building a house on it and charging people rent to live in the house.
This feels right from an ethical perspective—you put effort into building stuff on the land, and you should reap the rewards of your labour. But the land belongs to everyone, and we should all share equally in natures bounty—so any benefit you get directly from the land should be given back to the people (or in practice, the government).
However it’s also got a lot of economic advantages too:
Economic Advantages
Unlike all other taxes, the supply of natural resources is fixed, and so taxes on them do not distort the market at all—supply of land remains exactly the same, and willingness to buy land does too. The tax to the government over time exactly equals the previous interest adjusted cost of buying the land, so the cost of using land remains exactly the same as it would be if there were no taxes. Thus you can replace a certain percentage of current market distorting taxes with this distortion free tax, increasing total production.
There’s no incentive to hold onto land for investing purposes, since the price of land is fixed at 0, no matter if the area becomes more valuable. This encourages land to be used, rather than left fallow for 20 years till it can be sold at a profit.
Relatedly everyone pays the same rate of tax, ideally at the market rate for renting the land, no matter how they use the land. So only those with ideas for how to use the land that are more profitable than the market average will be willing to hold onto the piece of land, encouraging land to be held by those who would make the best use of it.
Disadvantages
The main disadvantage is the difficulty of assessing the value of the land. There’s no market for land in this scenario, so it has to be done manually. This may be expensive, and will also not be perfectly accurate, preventing us from reaping the full theoretical economic benefits of these taxes.
Harberger Taxes
As seen on Overcoming Bias (lots of blog posts, follow the links from that one to read them all).
Harberger taxes require all citizens to declare a price at which they’re willing to sell the property they own. Anyone can offer to buy the property at that price, at fixed terms and conditions for all properties. In order to encourage people not to set the price of their property arbitrarily high, property tax is set as a percentage of this self evaluated price, meaning you have to trade off reducing the chance of your property being sold against minimizing your tax burden. Assuming taxes are reasonably high, and the market is reasonably liquid, the ideal level to set your house price is the approximately the one at which you actually value the house.
This has a number of economic benefits:
Economic Advantages
The property market becomes extremely liquid, meaning its easy to buy up a large cluster of buildings to build something that requires a lot of land (e.g. a road) - something that’s almost impossible in urbanized areas today.
Property prices accurately reflect property value of the house, allowing them to be used as a market for futarchy (see Overcoming Bias blog post linked above).
The person who values a property the most will end up owning it, as he will buy out anyone who undervalues the house, and set the self evaluated price higher than anyone else is willing to pay.
Disadvantages
High property taxes distort the market, discouraging improving the property as it will increase the property tax.
It’s not obvious what is the optimal level to set the taxes.
Imagine a decision is being made which will hugely impact the price of your house—e.g. a company is deciding whether to build a sewage treatment plant in the neighboring plot. The moment the decision is announced the value of your house will shoot up or down. If it shoots up and you don’t make the adjustments in self evaluated price before an automated bot snaps up your house, you will have to pay the full difference in price between its previous price and its current price to buy back the house. This could easily destroy peoples lives, and keeping abreast of all potential such changes so you can change price preemptively is extremely difficult. An insurance market might be able to solve this.
Can we do better?
Georgian land taxes and Harberger taxes have almost opposite benefits. One has a really strong ethical basis, the other doesn’t have much of one, and risks significant harm to the average person. One reduces distortion in the market and discourages speculation, but doesn’t have a strong way of evaluating prices or increasing liquidity. The other increases distortion, encourages speculation, but gives every property a price, and makes the market extremely liquid.
Can we get the best of both worlds?
The obvious first step is to make a market for Georgian land taxes. Essentially this means that the government rents out the land to the highest bidder. At any point, whoever is willing to offer the highest rental rate to the government for use of the land gets it immediately, under standard terms and conditions.
This seems to offer all the benefits of Georgian and Harberger taxes, without any of the disadvantages (we’ll go through them in more detail later). However it obviously wont work as is—what do you do if somebodies already built on the land? The highest bidder rents the rights to the land itself from the government, but has no right to the property built on the land.
The next step is to say that whoever rents the land has to buy the property rights from the previous owner. But the previous owner will just charge the highest price he can, offer a land rent of 0, and we’re essentially back to where we are today. We could have somebody assess the property and calculate the fair value of the property, but then we essentially have the same problem as with Georgian taxes—how do we make sure that the assessment is accurate?
The proposal
Once again all land is owned by the government, and you have to rent land from them. Whoever offers the highest rent is given the land automatically. The exact details of the market can be hashed out separately (e.g. how soon do you get the land after offering a higher rent, how often can you drop/raise your bid, etc.) and will likely evolve over time.
Anyone who rents land has full rights to the land, but not to the property the previous owner built on that land. He can either buy/rent the rights to the property from the previous owner, or he can destroy the property, returning the land to a state no better than its base state, and rebuild whatever he wants on the land.
At what price then will the property be sold? The seller can’t charge higher than the cost of rebuilding the property + the cost of demolition, since otherwise the buyer will just destroy the property and he’ll get nothing. Any amount under this price is pure profit for the buyer, so assuming a competitive market the price should settle somewhere slightly below this mark, or about the cost of building the property (assuming demolition costs are much smaller than building costs). The rental price in a competitive market will then be value of using property—cost of building property = value of land, exactly as desired.
Analysis
Now the sharp eyed reader might have spotted a problem—this works well if I happen to want exactly the same property as you’ve built on your land. But what if I would want a slightly different property? Then I would be willing to pay a slightly lower amount before I decided to rebuild the whole thing from scratch because I gain slightly more from the rebuilt property than the current one.
In fact, what if I planned to build a skyscraper where you’ve built a house, and I need to knock the whole thing down anyway? Then I wouldn’t be willing to pay you a penny for your property—you would lose everything you invested in that property!
I would argue that this actually makes a lot of sense from an ethical perspective. The land, as argued earlier, belongs to everyone. If I want to build a skyscraper on it, and I’m willing to give the most back to the rest of the people in return (pay the highest rent), I should be able to, and the fact that you grabbed the land before me and built a house on it shouldn’t stop me. Bad enough that I have to pay to demolish your house, I should have to pay you the value of the house which I don’t even want too? It’s not your land that you should be able to force other people off it by building something there before they get to it!
But what about from an economic perspective? Doesn’t this hugely increase the risk of building property?
Well in most cases you wont lose very much at all. All you have to do is offer a slightly higher rent than the market rate, and the potential skyscraper builder will move on to the neighboring plot of land instead, which he can get at market equilibrium price. The only person willing to offer a higher rate than that will be somebody who actually wants your house, and he’ll negotiate a lower price from you for the house in return for the higher rent. Eventually though, as skyscrapers get built on every free plot of land around, the market rate might get so high that it’s greater than any benefit you get from the land. However this is at exactly the point where it would be economically beneficial for you yourself to knock down the house and rebuild a skyscraper anyway—your house is purely a liability at this point. In other words, you stand to lose exactly as much money as the amount of harm your bad investment decisions (building a house where a skyscraper would be more beneficial) cost the economy.
If people are still worried about the risk of losing everything, and so unwilling to invest, I presume that insurance may offer a reasonable solution to this problem.
The next problem is that people will deliberately build their houses out of impossible to destroy materials so that destroying the house becomes prohibitively expensive. The simplest solution is to simply make this illegal, with steep fines if it can be proven that you deliberately built your house in such a way as to make destruction expensive without any other benefits. An alternative is to split the cost of demolition between the two parties in some fashion, so that the builder also has an incentive to keep costs down.
Comparison With Georgian Land Taxes
Let’s go through the 3 benefits of Georgian land taxes and see whether we still get them:
Supply of land is fixed, so imposing a rent on it doesn’t distort the market. Great.
You can’t own land, so holding onto land whilst not using it is just purely paying rent for no benefit.
The person who offers the highest rent for the land ends up owning it—i.e. the person who’s able to make the best use out of it.
It also solves the disadvantage of Georgian land taxes—the land rent is now set by the market so it doesn’t need to be assessed manually.
So we keep all the advantages of Georgian land taxes, without the major disadvantage.
Comparison with Harberger taxes
Let’s go through the 3 benefits of Harberger taxes and see whether we still get them:
Land becomes extremely liquid. Buying the property rights on the land still requires negotiations, but since you can always just destroy the property that will usually not be a showstopper.
Land rents accurately reflect the value of the land, enabling them to be used as a market for futarchy. For example you could have people bet on which zoning policy will most raise lend rents, and pick the one which does best.
The person who most values the piece of land will offer the highest rent for it and own it.
So we get most of the benefits of Harberger taxes but with slightly different details—the benefits all apply to the land rather than the property. It’s not obvious which is better—for example would a futarchy based on land rents suggest better policies than one based on property prices? I’d expect them to be pretty similar, unless they effect the cost of building—a policy which increases the cost of building would increase property prices, but decrease land rents.
What about the disadvantages of Harberger taxes?
There’s no property tax, so we don’t distort the market that way. On the other hand we distort the market differently—people will be worried that they’ll lose their investment if somebodies rents the land at a higher price and destroys the existing property. This encourages people to make improvements that are likely to be universal improvements, rather than ones which are specific to them. It’s not obvious if this is a bug or a feature.
The land rents are set at the market rate.
If somebody offers a higher rent than you for your land, you can just outbid them, so you don’t lose anything if you don’t proactively raise land rents. In fact this is a bit of a problem since current renters have no incentive to raise their bid until somebody else outbids them. This could be solved by suitable market strategies—e.g. fining people if they raise their bid in response to someone else’s bid rather than organically, or by just assuming that in a suitably competitive market, there will always be someone bidding on your land. On the other hand you stand to lose in a different way—if somebody offers a higher rent for your land than you can afford and destroys your property.
So here it’s a mixed bag. We solve the disadvantages of Harberger taxes, but replace them with some disadvantages of our own. It’s not immediately obvious which is better.
Areas for future investigation
How can we make this more stable? How can we encourage investment when your entire investment could be destroyed at any moment.
How exactly would the market for land be structured?
How soon do you take over land if you offer a higher bid? What if the current owner counteroffers? What if someone else counteroffers before you take it over.
How do you prevent someone offering a high bid to take over land, and then immediately dropping it to 0? Any time someone outbids him he does the same thing, thus keeping the land forever and paying no rent.
What are the terms and conditions under which you can use the land?
Do you have to return the materials from the property to the previous owner when you destroy it? If so are you required to take care to keep them in good condition?
I mentioned that insurance markets might protect your investment if your property was destroyed. What would those insurance markets look like? How would they distort the market by changing incentives?
How would this affect construction? I would expect modular housing which can be moved from location to location to become hugely popular, since that’s an obvious way to protect your investment if somebody takes over the land. This might in fact be one of the major benefits of such a policy!
How could we create such land rent markets today without being unfair to those who currently hold the land? Whilst those who’ve held onto land for years have already more than reaped the benefits of their investment, those who recently bought land could end up millions of dollars in debt. Banks which have sold mortgages backed by land would collapse, bringing down everyone who has money in those banks.
How do we prevent malicious actors destroying your property not for personal gain but to damage you. E.g one company could destroy another company’s factory. (Asked by Richard Ngo).
Combining the best of Georgian and Harberger taxes
Epistemic status: this probably would be a bad idea, but has some really nice theoretical properties that I hope somebody can find a way to preserve whilst making this a good idea.
I find both Georgian land taxes and Harberger taxes to be fascinating proposals to improve the way we tax land. I’ll start off with a brief overview of both, then discuss benefits and advantages of both, suggest a model to combine some of the benefits of both, and analyze the combined approach. Finally I suggest areas for future research.
A Brief overview of Georgian Land Taxes
As seen on AstralCodexTen.
The key principle of Georgian land taxes is you tax natural resources at a rate such that their market price is 0.
In other words, if I own a piece of land, and build a house on it, then the government taxes the value of the land, but not the value of the house—so two identical pieces of land in the same area with different buildings on them get taxed at the same rate.
How high is that rate? At exactly the rate such that if I’d built nothing on the land I would be neutral between:
Giving it away to someone else for free so that he will pay the taxes on it instead of me.
Building a house on it and charging people rent to live in the house.
This feels right from an ethical perspective—you put effort into building stuff on the land, and you should reap the rewards of your labour. But the land belongs to everyone, and we should all share equally in natures bounty—so any benefit you get directly from the land should be given back to the people (or in practice, the government).
However it’s also got a lot of economic advantages too:
Economic Advantages
Unlike all other taxes, the supply of natural resources is fixed, and so taxes on them do not distort the market at all—supply of land remains exactly the same, and willingness to buy land does too. The tax to the government over time exactly equals the previous interest adjusted cost of buying the land, so the cost of using land remains exactly the same as it would be if there were no taxes. Thus you can replace a certain percentage of current market distorting taxes with this distortion free tax, increasing total production.
There’s no incentive to hold onto land for investing purposes, since the price of land is fixed at 0, no matter if the area becomes more valuable. This encourages land to be used, rather than left fallow for 20 years till it can be sold at a profit.
Relatedly everyone pays the same rate of tax, ideally at the market rate for renting the land, no matter how they use the land. So only those with ideas for how to use the land that are more profitable than the market average will be willing to hold onto the piece of land, encouraging land to be held by those who would make the best use of it.
Disadvantages
The main disadvantage is the difficulty of assessing the value of the land. There’s no market for land in this scenario, so it has to be done manually. This may be expensive, and will also not be perfectly accurate, preventing us from reaping the full theoretical economic benefits of these taxes.
Harberger Taxes
As seen on Overcoming Bias (lots of blog posts, follow the links from that one to read them all).
Harberger taxes require all citizens to declare a price at which they’re willing to sell the property they own. Anyone can offer to buy the property at that price, at fixed terms and conditions for all properties. In order to encourage people not to set the price of their property arbitrarily high, property tax is set as a percentage of this self evaluated price, meaning you have to trade off reducing the chance of your property being sold against minimizing your tax burden. Assuming taxes are reasonably high, and the market is reasonably liquid, the ideal level to set your house price is the approximately the one at which you actually value the house.
This has a number of economic benefits:
Economic Advantages
The property market becomes extremely liquid, meaning its easy to buy up a large cluster of buildings to build something that requires a lot of land (e.g. a road) - something that’s almost impossible in urbanized areas today.
Property prices accurately reflect property value of the house, allowing them to be used as a market for futarchy (see Overcoming Bias blog post linked above).
The person who values a property the most will end up owning it, as he will buy out anyone who undervalues the house, and set the self evaluated price higher than anyone else is willing to pay.
Disadvantages
High property taxes distort the market, discouraging improving the property as it will increase the property tax.
It’s not obvious what is the optimal level to set the taxes.
Imagine a decision is being made which will hugely impact the price of your house—e.g. a company is deciding whether to build a sewage treatment plant in the neighboring plot. The moment the decision is announced the value of your house will shoot up or down. If it shoots up and you don’t make the adjustments in self evaluated price before an automated bot snaps up your house, you will have to pay the full difference in price between its previous price and its current price to buy back the house. This could easily destroy peoples lives, and keeping abreast of all potential such changes so you can change price preemptively is extremely difficult. An insurance market might be able to solve this.
Can we do better?
Georgian land taxes and Harberger taxes have almost opposite benefits. One has a really strong ethical basis, the other doesn’t have much of one, and risks significant harm to the average person. One reduces distortion in the market and discourages speculation, but doesn’t have a strong way of evaluating prices or increasing liquidity. The other increases distortion, encourages speculation, but gives every property a price, and makes the market extremely liquid.
Can we get the best of both worlds?
The obvious first step is to make a market for Georgian land taxes. Essentially this means that the government rents out the land to the highest bidder. At any point, whoever is willing to offer the highest rental rate to the government for use of the land gets it immediately, under standard terms and conditions.
This seems to offer all the benefits of Georgian and Harberger taxes, without any of the disadvantages (we’ll go through them in more detail later). However it obviously wont work as is—what do you do if somebodies already built on the land? The highest bidder rents the rights to the land itself from the government, but has no right to the property built on the land.
The next step is to say that whoever rents the land has to buy the property rights from the previous owner. But the previous owner will just charge the highest price he can, offer a land rent of 0, and we’re essentially back to where we are today. We could have somebody assess the property and calculate the fair value of the property, but then we essentially have the same problem as with Georgian taxes—how do we make sure that the assessment is accurate?
The proposal
Once again all land is owned by the government, and you have to rent land from them. Whoever offers the highest rent is given the land automatically. The exact details of the market can be hashed out separately (e.g. how soon do you get the land after offering a higher rent, how often can you drop/raise your bid, etc.) and will likely evolve over time.
Anyone who rents land has full rights to the land, but not to the property the previous owner built on that land. He can either buy/rent the rights to the property from the previous owner, or he can destroy the property, returning the land to a state no better than its base state, and rebuild whatever he wants on the land.
At what price then will the property be sold? The seller can’t charge higher than the cost of rebuilding the property + the cost of demolition, since otherwise the buyer will just destroy the property and he’ll get nothing. Any amount under this price is pure profit for the buyer, so assuming a competitive market the price should settle somewhere slightly below this mark, or about the cost of building the property (assuming demolition costs are much smaller than building costs). The rental price in a competitive market will then be value of using property—cost of building property = value of land, exactly as desired.
Analysis
Now the sharp eyed reader might have spotted a problem—this works well if I happen to want exactly the same property as you’ve built on your land. But what if I would want a slightly different property? Then I would be willing to pay a slightly lower amount before I decided to rebuild the whole thing from scratch because I gain slightly more from the rebuilt property than the current one.
In fact, what if I planned to build a skyscraper where you’ve built a house, and I need to knock the whole thing down anyway? Then I wouldn’t be willing to pay you a penny for your property—you would lose everything you invested in that property!
I would argue that this actually makes a lot of sense from an ethical perspective. The land, as argued earlier, belongs to everyone. If I want to build a skyscraper on it, and I’m willing to give the most back to the rest of the people in return (pay the highest rent), I should be able to, and the fact that you grabbed the land before me and built a house on it shouldn’t stop me. Bad enough that I have to pay to demolish your house, I should have to pay you the value of the house which I don’t even want too? It’s not your land that you should be able to force other people off it by building something there before they get to it!
But what about from an economic perspective? Doesn’t this hugely increase the risk of building property?
Well in most cases you wont lose very much at all. All you have to do is offer a slightly higher rent than the market rate, and the potential skyscraper builder will move on to the neighboring plot of land instead, which he can get at market equilibrium price. The only person willing to offer a higher rate than that will be somebody who actually wants your house, and he’ll negotiate a lower price from you for the house in return for the higher rent. Eventually though, as skyscrapers get built on every free plot of land around, the market rate might get so high that it’s greater than any benefit you get from the land. However this is at exactly the point where it would be economically beneficial for you yourself to knock down the house and rebuild a skyscraper anyway—your house is purely a liability at this point. In other words, you stand to lose exactly as much money as the amount of harm your bad investment decisions (building a house where a skyscraper would be more beneficial) cost the economy.
If people are still worried about the risk of losing everything, and so unwilling to invest, I presume that insurance may offer a reasonable solution to this problem.
The next problem is that people will deliberately build their houses out of impossible to destroy materials so that destroying the house becomes prohibitively expensive. The simplest solution is to simply make this illegal, with steep fines if it can be proven that you deliberately built your house in such a way as to make destruction expensive without any other benefits. An alternative is to split the cost of demolition between the two parties in some fashion, so that the builder also has an incentive to keep costs down.
Comparison With Georgian Land Taxes
Let’s go through the 3 benefits of Georgian land taxes and see whether we still get them:
Supply of land is fixed, so imposing a rent on it doesn’t distort the market. Great.
You can’t own land, so holding onto land whilst not using it is just purely paying rent for no benefit.
The person who offers the highest rent for the land ends up owning it—i.e. the person who’s able to make the best use out of it.
It also solves the disadvantage of Georgian land taxes—the land rent is now set by the market so it doesn’t need to be assessed manually.
So we keep all the advantages of Georgian land taxes, without the major disadvantage.
Comparison with Harberger taxes
Let’s go through the 3 benefits of Harberger taxes and see whether we still get them:
Land becomes extremely liquid. Buying the property rights on the land still requires negotiations, but since you can always just destroy the property that will usually not be a showstopper.
Land rents accurately reflect the value of the land, enabling them to be used as a market for futarchy. For example you could have people bet on which zoning policy will most raise lend rents, and pick the one which does best.
The person who most values the piece of land will offer the highest rent for it and own it.
So we get most of the benefits of Harberger taxes but with slightly different details—the benefits all apply to the land rather than the property. It’s not obvious which is better—for example would a futarchy based on land rents suggest better policies than one based on property prices? I’d expect them to be pretty similar, unless they effect the cost of building—a policy which increases the cost of building would increase property prices, but decrease land rents.
What about the disadvantages of Harberger taxes?
There’s no property tax, so we don’t distort the market that way. On the other hand we distort the market differently—people will be worried that they’ll lose their investment if somebodies rents the land at a higher price and destroys the existing property. This encourages people to make improvements that are likely to be universal improvements, rather than ones which are specific to them. It’s not obvious if this is a bug or a feature.
The land rents are set at the market rate.
If somebody offers a higher rent than you for your land, you can just outbid them, so you don’t lose anything if you don’t proactively raise land rents. In fact this is a bit of a problem since current renters have no incentive to raise their bid until somebody else outbids them. This could be solved by suitable market strategies—e.g. fining people if they raise their bid in response to someone else’s bid rather than organically, or by just assuming that in a suitably competitive market, there will always be someone bidding on your land. On the other hand you stand to lose in a different way—if somebody offers a higher rent for your land than you can afford and destroys your property.
So here it’s a mixed bag. We solve the disadvantages of Harberger taxes, but replace them with some disadvantages of our own. It’s not immediately obvious which is better.
Areas for future investigation
How can we make this more stable? How can we encourage investment when your entire investment could be destroyed at any moment.
How exactly would the market for land be structured?
How soon do you take over land if you offer a higher bid? What if the current owner counteroffers? What if someone else counteroffers before you take it over.
How do you prevent someone offering a high bid to take over land, and then immediately dropping it to 0? Any time someone outbids him he does the same thing, thus keeping the land forever and paying no rent.
What are the terms and conditions under which you can use the land?
Do you have to return the materials from the property to the previous owner when you destroy it? If so are you required to take care to keep them in good condition?
I mentioned that insurance markets might protect your investment if your property was destroyed. What would those insurance markets look like? How would they distort the market by changing incentives?
How would this affect construction? I would expect modular housing which can be moved from location to location to become hugely popular, since that’s an obvious way to protect your investment if somebody takes over the land. This might in fact be one of the major benefits of such a policy!
How could we create such land rent markets today without being unfair to those who currently hold the land? Whilst those who’ve held onto land for years have already more than reaped the benefits of their investment, those who recently bought land could end up millions of dollars in debt. Banks which have sold mortgages backed by land would collapse, bringing down everyone who has money in those banks.
How do we prevent malicious actors destroying your property not for personal gain but to damage you. E.g one company could destroy another company’s factory. (Asked by Richard Ngo).