Depreciating Licensing Model.
Welcome, premium subscribers! Thank you for subscribing. Your names will be recognised on the next video š
TLDR below. This is not financial advice.
Catch the episode on YouTube
General Conclusion
Today we will discuss property rights from real life to NFT space with Anthony Lee Zhang. He co-authored a paper with Glen Weyl on depreciating licenses (DL). We will go from existing problems with property rights in the physical world to the digital world and how the world is changing to bring more rights and fairness to everyone.
In this discussion, we will refer to DL model including Privatisation, Holdout and Allocation. Next will be the trade-offs of allocation for investment in the short and long term. Finally, we will discuss governance in DL, game theory for DL model and applications in NFT.
Assistant Professor of Finance UChicago Booth, Anthony Zhang
If I could teleport back to two years ago I would tell myself to watch out for this paper. It is exactly what I was looking for back then. Property rights and tokenising property rights as well as finding more efficient ways to be embedded within property rights for efficient allocation is beautiful. We can do that now! Anthony discusses his research paper on property rights.
About Anthony Zhang
Anthony: I’m Anthony Zhang, an assistant professor of finance at the University of Chicago Booth. I mainly work in financial market intermediation and market design. I analyse questions about how we can analyse financial markets, figure out how efficient they are and how we can make them work better. Some of the more classical markets that I study are the Banking industry, Housing markets, and Derivative markets.
This is an earlier paper actually which I’ve recently revised. We mainly were aiming this paper at analysing questions of natural resource license design but there seems to be an application in the crypto space too. I think the paper predates the NFT boom. When we were seeing all this NFT action going on, we realised that some of the ideas here might also apply to the NFT space, so this is where Iām coming from.
Introduction to Depreciating Licenses Model
The way we often motivate this is that the government has a bunch of natural resources. This can be land but also things like radio spectrum, oil drilling rights, fishing rights, etc. All these things are sources of things that somebody can generate a value from like fishing, building radio towers and selling 5G spectrum. The government wants to decide who should get access to use these resources, and they want to raise revenue from the use of these resources. Privatisation has been the classic way to do this since the Chicago Free Market Revolution.
Property rights in: land but also things like radio spectrum, oil drilling rights, fishing rights Generate a value: like fishing, building radio towers and selling 5G spectrum.
Privatisation
Before this privatisation boom, the government would take these licenses and give them to the best user. This did not work so well because normally the government is constrained regarding how much information it can gain about who will be the most efficient user. Hence, the government decided to privatise and auction off these licenses based on the assumption that the highest bidder for a license would be the best user of the asset. This started happening from the 90s onwards.
Holdout
Privatisation works fairly well but it generates distortions. One of the biggest distortions is that it generates holdouts ā when the government sells a spectrum license or a land-use license, that allocation is often efficient at the time it first sells the license. But 20 years later the same company is stuck with the license and much more efficient companies have come in and they want the license but the old company basically holds onto the license and waits to sell it at the highest price possible.
Allocation
Pure privatisation does not work perfectly. What we noticed is that there is a way to do slightly better than privatisation in terms of allocating resources efficiently. Resource license design affects the efficiency of use of these resources. Pure privatisation selling really long-term use rights gives buyers security in their assets and gives them incentives to invest. But it generates holdout problems because people own inalienable rights and can hold on to these assets longer than they socially should. It is hard to reallocate them to new entrants.
Trade off: Efficient Allocation vs Long-Term Investment
We are talking about two different problems here:
-
One is whether you are holding it long enough for you to invest additional resources to build up the radio towers or infrastructure and maximise the utility of the spectrum that is awarded to you versus
-
Allocation efficiency or asset allocation where we can allocate it to the person with the highest value right now.
We are trying to figure out how to balance both of them because they are on opposite spectrums: efficient allocation right now that will end after a short term only where you trade-off in terms of investment, or long-term investment where you trade off in terms of asset allocation.
Long Term VS Short Term License
Old licensed designs face this trade-off that long-term licenses give you high investment incentives but are not very good for allocation and short-term licenses have the opposite problem. Where we come in is that we can propose a better way to navigate this trade-off. You can think of a long-term licensee as owning 100% of a resource forever and a short-term licensee as owning 100% of the resource for a few years.
We can design a license that lasts forever but it depreciates. It is like you own spectrum-use rights for 100% one year then 90% next year and 81% the year after and so on. We basically have property rights that decay over time.
Example:
Let’s say every year the government prints 10% more stock in the license and the idea is that every year to use the license or the resource you have to own 100% of the stock. If you owned the license last year then this year you have to participate in an auction where either you can buy 10% of the license from the government and then continue using the asset or you can sell your 90% to a buyer and leave. This is a partial property right because if you own something this year you own a fraction of it next year and a smaller fraction the next year and so on. You have to pay to keep using the resource and so you have a lower incentive to hold out for really high prices. This, then, is the idea of a Depreciating License.
Government Quantitative Easing VS Asset Dilution
Lisa: Sounds like this is similar to quantitative easing?
Anthony: There is a difference between a Depreciating License and Quantitative Easing (QE).
With QE, the government has discretion regarding how much money it wants to print so there is this uncertainty aspect. The economy has to price in the possibility of QE but doesn’t know how much the government is going to commit to printing money. The possibility of hyperinflation is basically a loss of trust in the government as a controller of the currency.
The way we are pitching Depreciating Licenses is that the government should commit upfront to a path of selling equity. If it does that then concerns about dilution are less salient because everyone knows that when you buy a license it entitles you to 90% next year then 81% the year after and so on. In the paper, we have some models showing how you can calculate the effect of these depreciation rates on prices. People will put in a lower price because they will have to pay these fee payments in the future.
Lisa: It is not possible for partial ownership because you have to own 100% before you can actually use that license. You need to own 100% through paying the fixed fee and paying the rental fee and then you can use the entire spectrum.
Cost Theory VS DL Model
Lisa: You co-wrote this paper with Glen Weyl and he also has been talking a little bit about the Cost Model where you own 100%. But every period, you have to self-report what the value is and then you have to pay taxes which is kind of like a short-term rental. It is quite similar to the Depreciating License Model where you pretty much have a long-term incentive to invest but also pay short-term rental fees to make sure that you are constantly being allocated the asset where you have the highest utility of it. How would you compare the two models?
Anthony: This paper is now old ā Glen and I started working on it very close to when Glen and Eric Posner started working on their early paper on Harvard licenses so this paper used to be a quantitative model of Harvard licenses, but recently we have pivoted it slightly.
Similarity
You are entirely correct that many of the basic features of the two mechanisms are exactly the same and both are designed to trade off allocative and investment efficiency. They do so by creating these depreciation or self-assess license fees which have the effect of limiting the owner’s ability to hold out. The core difference is basically about how you determine this license fee payment.
Determining the Fees
In the Harvard Model, this payment is determined by self-assessment where individuals announce how much they value the asset. In our Depreciating License Model, these fees are determined in an auction ā a second price or an ascending auction.
Benefits and Trade Off
There are benefits and trade-offs to each model. The benefit of the auction model arises from a critique that we got when we were talking about the self-assessment model. People were not happy with this kind of property of self-assessment where you might announce a price and hope nobody buys it and if somebody buys it then you suddenly lose your house.
If, instead, you find a way to calculate these fees using an auction, at least it has a kind of no regret property. In an ascending auction you say you value your house this much and if somebody outbids you then you can match them and keep your house but pay higher taxes. Or you can give up and sell the house to them. Or it is possible that somebody bids high enough that you find it in your interest to sell. Whatever the case, you won’t be surprised by losing your house and so that is the core mechanistic difference between the DL model and the Cost Model and then, as a result, the game theory or price setting and standards are slightly different.
Application in NFT
Lisa: Let’s bring the conversation back to crypto for a little bit. Simon de la Rouviere has this model where he experimented with art. It uses the Cost Model where you buy the art and every time you have to self-assess the cost of it and pay some amount of taxes. He tried 5% taxes versus 100% taxes back to the artist. It’s kind of like licensing art because you fully own the art but it also allows you to resell the art to someone else who might value the art more than you. Part of the sale is taxed and given to the art creator. That’s asset allocation.
How do you think depreciating licenses can be added into the crypto space or NFT space?
Anthony: I think this is a really great application and you could almost literally take the model we are seeing and just apply it to NFTs.
Models in the Past
Let me draw one parallel which is things like this have been tried in the art world even before crypto as there was a desire at some point to make it so that artists would get paid when their art trades on the secondary market. There were systems where people tried to write some sort of legal constraints on artwork which says whenever this is transacted on the secondary market the original artist gets some percentage of the sale.
Issue and Solution
The issue with these systems is that it distorts sales because you have a tax that is levied on the sale but not when the artwork is not sold and that discourages people from selling the artwork. Relative to these systems Harvard Licenses or Depreciating Licenses can be seen as a system that basically lets the artist keep a stake in the artwork even after it is sold and then if an artist gets very famous they still get a flow of revenue from their earlier works.
Advice to NFT Developers using the DL Model
Anthony: Do it and contact me because I would be happy to talk with anyone who is trying this. I think it is an interesting area because a key principle of market design as an academic is that everything works well until you hit the real world and then a ton of stuff that you have never anticipated happens.
This is still a young mechanism cost, and depreciating licenses even more so. I think it would benefit the space if we have people actually doing practical things talking to theorists like me and then drawing these connections between the game-theoretic models of the world and the applications. It is a very interesting space because of both the theory and the applications and I would be very excited to talk to anyone who is even just at an early-stage of thinking about doing anything in this space.
Listen on podcast if you prefer audio version.
TLDR:
In the Depreciating License, Anthony and Glen come up with a simple game theory model of property ownership. Basically, you own a fixed percentage of the asset and the remaining percentage is auctioned or sold every time period. In this way, it combines both a full ownership model and a full rental model.
This is quite attractive because in allocating property rights, we often experience a trade-off between incentive investment (100% equity ownership) and asset allocation (allocating to the best at the time of allocating). The amortization license model combines the two and balances the trade-off.
Ps: Order the textbook “Economics and Math of Token Engineering and DeFi” today!