Six years ago I posted on the idea of using combinatorial auctions as a substitute for zoning. Since then, news on how badly zoning has been messing up our economy has only gotten worse. I included the zoning combo auction idea in my book The Age of Em, I’ve continued to think about the idea, and last week I talked about it to several LA-based experts in combinatorial auctions.

I’ve been pondering one key design problem, and the solution I’ve been playing with is similar to a solution that also seems to help with patents. I asked Alex Tabarrok, whose office is next door, if he knew of any general discussion of such things, and he pointed me to a long (110 page) 2016 paper called “Property is another name for monopoly” by Eric Posner and Glen Weyl. (See also this technical paper.) And that turned out to be a relatively general argument for using the specific mechanism that I was considering using in zoning combo auctions, get this, as a new standard kind of property right for most everything! Looking for web discussion, I find a few critical responses, and one excellent 2014 Interfuildity post on the basic idea. In this post I’ll go over the basic idea and some of its issues, including two that Posner and Weyl didn’t consider.

But let’s start with basics. Imagine that you are the sole power over some new large empty territory, able to do anything you want with it. You will soon invite many people to (pay to) enter your territory, and each of them will need some sort of local property rights to support their activities. So you will need to divide your large territory into many smaller property units. And you will want to divide things well, i.e., to “carve nature at its joints,” so that you can promote the productive use of this territory. After all, if people expect to be more productive in your territory, they’ll pay you more for your properties. Yes, if you bundle things together badly, people might be able to re-bundle them in better ways later. But that could be a slow and expensive process; better to get it right the first time.

For example, you’ll want to put things that need to be coordinated more closely together into the same bundles, and you’ll want boundaries between the units that are easy to monitor and enforce. You may also want each unit to contribute to and be subject to some sort of governance structure, to ensure a rule of law and sufficient production of public goods. Even in the best case of a single owner who can choose any property rules he or she likes, this general problem of designing efficient property rights is complex and hard. But the framework I’ve just outlined is in essence the usual account of ideal property rights within the law & economics field (a subject on which I teach often) .

One important dimension of property design is the strength of the property rights. For example, in land property you might let each owner do absolutely anything they want with their land, or your might limit some activities (like explosives, pollution, and blocking views or sunlight) if neighbors are likely to place a higher value on such limits than each owner would place on their absence. And one key area where property rights can be stronger or weaker is regarding one’s freedom to set a resale price.

When you announce an offer to sell your property at a certain price, you in effect give an option to buy that property to the rest of the world. The world values this option, and values it more the lower is its strike price. Options on properties given to the world at lower prices make it easier for others to buy the properties that they value, and to assemble and recombine property units into new bundles. So when you design a property rights system, you might want to add extra incentives for people to create and maintain lower price offers.

Consider the examples of people who bought internet domain names, like “walmart.com”, early on in the hope of being able to resell them to the big firms who go by those names. Or consider homeowners who demand huge prices to sell their land to someone trying to assemble a large area to build a shopping mall. In these and many other cases, economic harm is done because owners can pretend to value their property for much more than they actually do. This also seems a big problem in combo auctions; one-sided combo auctions, where everyone buys from one seller, seem to work much better than do two-sided auctions, where many sellers can demand very high prices even when their values are much lower.

In 1965 the economist Arnold Harberger (of Harberger triangle fame) published on self-assessment in property taxes. The idea is to let each owner assess their property at any value they want, but that assessed value becomes a price at which they offer to sell their property to anyone. (Harberger didn’t invent the idea; it goes back at least to ancient Rome.) Posner and Weyl consider applying this basic idea to most all property, and call it a “Harberger tax.” That is, a simple way to create an incentive for low offer prices on any kind of property is to require the owner of that property to continuously pay a fee proportional to their posted offer price, i.e., the price they announce at which they will sell their property. The lower their price, the less they pay.

Of course the owner of a property is well advised to not set their price much below the value that they actually have for this property, to avoid regret should someone accept their offer. When people become attached to a property, their value will be well above its market price, and given an honest valuation of it, the chance that it will be sold should be very low.

Now this fee is not mainly intended to raise revenue for redistribution or central services. It is instead intended mostly to promote efficient use of property, so that the best people use each item in the best ways. Because of this, I’d rather not call this fee a “tax”, and so I’ll instead call it an “stability rent.” Just as you can buy something by paying all at once or by making “installment” payments spread out across time, part of this stability rent is payment for the property.  But another part is a price you pay over time to increase your chance of keeping the property. That is, to increase the stability of your relation to this property. You can never guarantee perfect stability, however, as that would require an infinite rent.

At this point you may feel that the more familiar kind of property, where you buy something at once for good and then never have to sell it or pay any property tax on it, is the more “natural” kind. Relative to that, this alternate form may seem to you an unnatural mutant abomination. But note that in fact today you usually have to make continuing payments to preserve most property. You pay for maintenance, cleaning, repair, and for a place to store things. You usually also suffer substantial chances of losing most property, via obsolescence, destruction, misplacement, theft, and legal and government policies, such as eminent domain. And you already assign explicit monetary values to property when you consider if to buy or sell, and how much to insure it for. So these new continuing fees, property loss chances, and assessment tasks are mostly just changes in degree from what you accept now.

You can also think of familiar property as a special case of this new property regime, with the stability rent set to zero. And I am persuaded by the analysis of Posner and Weyl to believe that the most economically-efficient rent level is usually not zero. At least if we can ignore some complications, that I will mention soon. Yes, increasing the fee decreases the incentive of a property owner to maintain and improve that property. But for small fees that loss is outweighed by gains in making property easier to transfer to new owners. (And zoning combo auctions might work better.)

To deal with various problems, Posner and Weyl propose that owners be allowed to update prices at anytime, to lump items into bundles that must be sold together, to specify “non-additive prices on subsets of goods”, to deduct observable investments from rents, and to get up to several months to actually transfer property once sold. They also suggest that potential buyers can pay a small refundable fee to inspect properties, and that some (as yet not worked out) combination of blockchains and IOT (internet of things) be used to allow direct transfer of property control while protecting owner privacy and preventing mischief by officials. They note that private insurance could help people to ensure their ability to make future rent payments, and that private apps could help people to estimate and frequently update their many property values. With everyone always posting prices on all of their property, such apps would have a lot of data to work with.

I’ll also note that previous owners should be liable for any damage to property during a transfer period, and that if stability rents were distributed back to pools of owners with roughly the same wealth and other social characteristics, then individuals would only pay net stability rents when they put higher than average values on their properties, relative to others in their pool. Gramma is then only at risk of losing her longtime home if she values that home much more than do other people much like her.

Yes, you might not like having to pay more in this system to get stability of your property. But that’s because in this system you are paying more of the social cost of stability. Its like how polluters prefer a world where they can pollute for free, and don’t have to pay for the social cost of their pollution. The rest of us who have to breathe their pollution shouldn’t sympathize much with their plight.

Yes, this system adds some complexity to property law, but it can also greatly simplify many areas of law, including property tax assessment, eminent domain, necessity exceptions to contract, accident liability levels, corporate buyouts, and patent and copyright licensing,

Posner and Weyl identify several factors that influence the ideal fee level: property turnover rate, sensitivity of value to owner investment, variance and multimodality of value distribution, and whether value tends to grow or decline over time. Even so,

We would advocate a relatively coarse system with a small number of easily distinguishable categories such as natural resources, equipment, real estate, corporate securities, general personal property, keepsakes, and heirlooms. … Rates within these categories being set per a coarse and easily auditable heuristic, such as the currently observed turnover rate.

They estimate that an ideal fee level to be roughly half the rate of property turnover:

A 2.5% annual rate is likely to be nearly optimal on this basis for a wide range of assets, like factories, natural resources, and houses, where investment plays a significant role but allocation can also be seriously distorted. … [This] would transfer about a third of use value to [rent recipients] … [It] achieves 70-90% of the maximum possible allocative welfare gains and the investment losses erode only 10-20% of these gains. .. [This is] likely to raise .. 10-20% of national income in most developed countries. .. The [purchase] prices of assets would be only a quarter to a half of their current level.

Let me finish this post by highlighting two issues that Posner and Weyl don’t seem to discuss. First, social norms may often arise to treat those who buy property through this system as illicit “scabs” who betray a community. For example, stamp collectors might create a shared social norm to post low official prices on their stamps, and to shun anyone known to have bought stamps via such prices. This coordination could greatly reduce the average stability rents paid by stamp collectors, a gain that comes at the expense of non-stamp-collectors.

Similarly, in some community members might typically put all of their property into one single bundle for one entire price of everything they own. This community might then try to identify and shun anyone known to have purchased such entire bundles. Members of this community might then get into the habit of greatly reducing the prices they put on total bundles, and thus greatly reducing their stability rent payments. If a scab were to actually try to buy such a bundle, the target would correctly feel that this scab threatened to grab a large fraction of the value of everything they own. This target might well respond emotionally and fiercely, and his or her allies might respond similarly. The threat of violence could be very real here.

I thus suggest that the size of property bundles be limited to much smaller scales. After all, it is hard to see many people actually buying such large bundles, and thus hard to see how huge bundles can add much to allocative efficiency. I also suggest that stability rents be refunded via small enough pools of socially-connected people. The idea is for pools to be small enough to discourage groups from coordinating to reduce their payments relative to outsiders, and yet large enough for fees to give individuals strong incentives to reveal their property values.

The second issue I want to highlight is that a second enforcement mechanism exists. The mechanism that Posner and Weyl consider for encouraging people to declare accurate values for their property is the prospect that someone might come along and honestly value it at more than the declared price. In commodity futures markets, however, most traders are speculators who don’t actually intend to take delivery of the commodity traded; they instead hope to get out of the market before the official delivery date, and to profit from having bought low and sold high. Similarly, in this system there could also be speculators who seek to profit from mispricings, but who don’t want to actually take delivery of property.

When the owner A of some property puts a price offer P1 on it, and then B accepts that offer on date D, then A might have a month M until they have to actually make the property available for B to control on date D+M. Before date D+M, it is B who officially owns the property, and by the rules of this system B must offer a new price P2 for anyone else to take. But if C accepts this P2 offer during this period, it makes more sense for C to gain the right to take control of the property from A on date D+M, instead of forcing B to actually take delivery and then have another month to transfer it to C on date D+2M.

But if this is the rule, then an original owner A who had underbid and set price P1 below their actual value V might be tempted to pay a higher price P2 to get their property back again, with P1 < P2 < V. And a speculator might take the role of B hoping for just this outcome, from which they would gain a pure cash profit of P2-P1. Given a chance Q of A having underbid and thus being willing to pay P2 to get it back, taking on such a speculator role to challenge A’s claim of value P1 should on average be profitable if

Q*(P2-P1) > (1-Q)*(T+P1-P0),

where T is the speculator’s transaction cost and P0 is the price they expect to get selling the property to a third party.

People who are in the market to buy a property like this, but who are flexible on which particular property they buy, are in an especially good position to take on this speculation role, as they have an especially low added transaction cost. After all, they were already going to buy and pay a transaction cost. Thus professional speculators and flexible buyers will watch for clues that indicate underbid properties that can be profitably challenged. And if society applies a norm of honestly to value declarations, we may not feel much sympathy for people who are “exploited” due to lying about their values for property.

As a result, owners are induced to limit how far they set prices below their actual values both by the prospect that sincere buyers will arrive who sincerely value the property at more than their declared price, and also by speculators looking to profit from underbids. This means that for any given fee rate, declared values will be more honest. We can thus set fees to lower higher levels than we otherwise might and still get good allocative efficiency.

Overall I find this proposal quite promising, though still underspecified. Someone really needs to work out what sort of blockchain and IOT arrangements might actually be workable here. I’d also like to see something better for dealing with asymmetric info on property features than the inspection proposal given. And of course we should do lab experiments and field trials before we actually adopted such property rules on a large scale.