Harberger Staking — Improving the ENS market

A proposal to improve the allocative efficiency of the ENS domain system without losing investment efficiency by using an adapted Harberger tax with economic staking.

This blog proposes a potentially better system to the current ENS vickery auction style distribution method for ENS domains. It uses an adapted Harberger tax with demonstrable economic staking to more effectively allocate ENS resources.

ENS

Harberger Taxes; an intro

Harberger taxes are an economic system that attempts to have both optimal allocative and investment efficency. Property owners only pay tax on what they think their property is worth, but the price that they establish for their property, is the price that the property is to be sold at in the wider market. This allows the property to be easily re-distributed at fair market price.

Allocative Efficiency refers to the efficiency of a system to continually and optimally distribute goods and services, taking into account consumer preferences. Optimal allocative efficiency means that the marginal cost of a product equals its marginal utility. An example of this is land acquisitions — if a new highway that is set to benefit a wider community needs to be built through a section of land that is currently occupied by private owners, then ideally the land moves from the private owners to the public good. However, without a mechanism for the land to be fairly re-allocated, the original private owners can simply hold and prevent the highway from going ahead. In this case, this system has poor allocative efficiency since the land would be more optimally used as a highway than private occupations, but there is no ability to fairly re-allocate.

Investment Efficiency refers to the characteristic of a system whereby owners can acquire property rights and know that they always retain the ability to make a sell decision at a market price they agree with (and that it won’t be seized off them). This typically comes at a cost of allocative efficiency, since a system with high allocative efficiency normally means that a government or body can seize or acquire assets from private users in order to re-distribute to users that could put it to better use. In the example of the land and highway above, if the government simply seized the land (or paid the market price to the owners and forceably moved them on), then this system would have better allocative efficiency, but comes at a cost of investment efficiency — since nobody would want to invest in property that could be siezed off them.

Harberger taxes demonstrate optimal allocative and investment efficiency since the market price that a property is valued at is what it is either sold at, or paid taxes on. This proposal adapts the Harberger tax idea in order to achieve high allocative and investment efficiency for ENS domains.

ENS Domains

ENS, or Ethereum Name Service, is a system to allocate .eth domains to the wider community. Users can acquire human-readable ENS domains and link them to ethereum account addresses.

ENS Vickery Auction

ENS has already distributed a large ENS domains to users. The current system of distribution is a complex and high-effort Vickery auction where new domains are bid on and acquired after a 5 day process of bidding and revealing. The minimum bid is 0.01 Ether and each year the owner needs to re-bid on their domain.

Apart from being quite complex, the current system has the following issues:

1. Poor price discovery. Since there no ability to engage with an existing owner on their domain (apart from third-party ENS marketplaces), it is very hard to price a domain.
2. Poor allocative efficiency. Since price discovery is hard, ENS resources are not easily bought and sold. Even though the domain is exposed to the auction process once a year, waiting a year to engage and bid on an ENS domain is not an ideal time cycle to re-allocate the domain.
3. Poor investment efficiency. Since it is hard to engage in price discovery with potential buyers, and a market is only established around a domain once per year, it does not allow domains to be effectively bought and sold in a dynamic and continuous market.

Proposal: “Harberger Staking”

This proposal is quite simple. Users acquire domains by staking ether against it at an amount they think it is worth. Anyone else can also stake ether against the same domain at a price they think it is worth. If they think it is worth more than the original owner, then a harberger tax is paid to the interested party based on the difference between the two prices.

Party A stakes X. Party B stakes Y. Party A pays a tax to Party B based on Y-X, if Y>X.

Implementation

Alice wishes to secure the alice.eth domain. As it is currently unclaimed, she places a stake of 1 Ether in the ENS staking contract and earns the rights to the name. Alice has a 2 week grace period on the domain.

After 2 weeks Bob wishes to acquire the alice.eth domain. He stakes 0.5 Ether against alice.eth. Since Alice’s stake > Bob’s stake, she demonstrates that she has a higher economic preference to the domain and the domain remains hers with no further action.

Bob increases his stake to now 3 Ether against the alice.eth domain. Since he has proved he has a higher economic preference to the domain, (Bob’s stake > Alice’s stake), Alice has three options:

1. Alice can increase her stake to 3 Ether to match Bob’s stake. This demonstrates she can match the economic preference of Bob to the domain in question. She will retain her domain since she acquired the rights to it first and nothing further is required.
2. Alice does nothing and will begin paying a harberger tax to Bob, based on the difference between his stake and hers (3–1 = 2 ether). This harberger tax may be set to 1% per month, for a total of 12% per year. This would result in 0.24 Ether per year (0.1 * 2* 12), drained from Alice’s stake to Bob’s stake, which he can withdraw at any time, or retain in the contract. As the difference between the two stakes increase, the payments would increase. If it reduces to the point that Alice is completely drained of Ether, then Bob immediately acquires Alice’s domain. In the example above, Bob would acquire the complete rights to Alice’s domain after 5 years, since that is the time the harberger tax of 12% pa completely drains Alice’s 1 ether. Bob now has 4 Ether, and the alice.eth domain, but Alice owned the domain for 5 years, at a realised cost of 1 Ether.
3. Lastly, Alice can immediately sell the domain to Bob, for the price of the difference of Bob’s ether to Alice (2 ether). Alice now has 3 ether and Bob has the domain with 1 ether remaining staked on it. Bob now has 2 weeks to increase his stake of ether against his domain to protect it from anyone who values it more than 1 ether.

Multiple Stakers

In the case that there is multiple stakers against a domain, the harberger tax is only paid to the highest staker (sub-sorted by time). Ie, if Charlie also stakes 2 ether in the example above, then only Bob receives the tax, or is able to buy from Alice.

Characteristics

The Harberger Staking model achieves a hyper-efficient market with near-optimal allocative and investment efficiency.

Price Discovery

Harberger staking causes a very efficient market to result with a completely anonymous mechanism for price discovery and domain purchasing. Each domain is priced at exactly what the market believes it should be at any given time, and each owner must provide an economic stake to each asset at all times. Owners do not even need to communicate to each other to settle on a price, since they simply need to place ether into the ENS contract to communicate the value of the domain.

Domains can be bought and sold at any time, and the entire sale process is handled trustlessly with no requirement for third-party marketplaces and escrows.

Allocative Efficiency

In the case of this proposal, domains are continually re-distributed to those that are prepared to pay the most for it (and prove as such), but still offering the original owner an ability to retain the domain at the price they think it is worth. If Party A thinks that is is worth X, but another party (B) thinks it is worth Y, then Party A will pay a tax based on Y-X. This allows the domain to still reach a market price of Y, without adversely affecting Party A.

In the case that Party A never sells no matter the cost of Y imposed on them, then they would suffer a war of attrition in the ever increasing cost of the tax they have to pay. Eventually the domain will be re-allocated to Party B, assuming Party A has finite resources. In this case, Party A should simply sell once Party B has staked a price that Party A is happy with. Assuming Party A has finite resources, then the domain will ALWAYS be re-allocated at some price Y, either through attrition or a direct sell. If the 1% tax is paid every 30 days, then new domains would be re-allocated after the minimum 30 days if Party B is prepared to pay 100x for the domain — however for the first 30 days Party A has an option to exit at 100x their investment.

Investment Efficiency

For this proposal, the harberger staking model ensures that private owners can invest in domains and know that they can always sell at fair market price. They can still acquire and invest in large amounts of ENS domains, but must be prepared to pay an on-going tax based on the actual market price of the domains. If they do not wish to pay a tax, they can simply and immediately sell their investment. This method is also difficult to game, since both buyers and sellers have to stake economic value before a property is sold or re-allocated.

Preventing Domain Squatting

Domain squatting is currently an issue for TLDs since anyone can acquire a domain at some arbitrary price, and retain it indefinitely with very little resources. Since there is very little cost to squat a domain, domain squatters can horde large amount of digital property and prevent fair re-distribution.

In the case of harberger staking, a domain staker would quickly run out of resources if they attempted to horde large amounts of domains. Since they need to stake on every domain they own, they will begin fighting a war of attrition once the wider market begins staking into their domains.

Additionally, and perhaps more importantly, a domain can be perpetually re-allocated with no central coordinator, at always fair market prices. In the case that a domain staker loses access to their domains or goes offline for a very long period of time, any amount of stake that is higher than what they originally stake will mean that it is eventually fully depleted and re-allocated. Thus there will never be a lost domain.

Existing argument against Harberger taxes

On Vitalik’s blog he mentioned that Harberger taxes had already been proposed for ENS, but it was deduced that it was not going to be effective since a scammer may have more resources than a legitimate owner; case in point being the MyEtherWallet domain.

However, in the Harberger Staking model, it can be seen that a scammer would need at least 100x the resources of a legitimate owner. Take the case that MyEtherWallet were staking $10k of Ether against their domain. A malicious actor would need to stake $1.01m of Ether against the same domain to cause a tax of $10k/month to be incurred. In this case it may be more ideal that MyEtherWallet would simply take the $1.01m offered and spend it on a re-brand and marketing.

Business Case for ENS

Since it is possible that a large amount of ether will be staked and a much more forth-coming marketplace will be facilitated, the ENS registry can charge a small 1% fee on the transactions paid between parties. For example, if 0.24 Ether is paid as tax between Alice and Bob, 0.0024 Ether can be collected by the ENS registry to fund ongoing development around the system. This can accumulated in the staking contract itself.

This “transactional tax” is only paid if a market is created between two parties, thus does not impact anyone who is not engaged in a market.

Conclusion

The current Vickery style auction for ENS domains does not facilitate efficient markets, and has characteristics of poor allocative and investment efficiency.

We invite the ENS team to consider this upgrade to the ENS registry which will see a much more efficient marketplace and adoption of the ENS, alongside better engagement, as well as a sustainable and fair business model for the ENS team themselves. It will be trivial for new adopters of ENS to acquire and buy domains they want, with a “set and forget” style approach to staking.

Follow me on twitter, I write, share and talk about the decentalised future.

Twitter: @jpthor_