Another price-setting mechanism: Token voting


Here’s another proposal for how to handle price setting in the permanent registrar: Using tokens and voting.

Suppose we allocate ENS tokens to existing stakeholders - domain owners, possibly others - by some algorithm. Stakeholders can vote on a target domain renewal price, in ether. We take the median of these votes, and use that as the renewal price. Token holders are rewarded with a share of renewal fees proportional to their token holdings.

If a user’s share of tokens owned is the same as their share of domains owned, then there’s no incentive to vote either way: They earn exactly as much for owning the tokens as they spend on paying for the domains.

If a user owns a smaller share of tokens than domains, their incentive is to vote the price towards zero, since their expenses exceed their income. However, such users by necessity also have a smaller effect on the price, due to their small token holdings.

If a user owns a larger share of tokens than domains, however, their incentive is to maximise profits - to vote for the price that results in the largest product of price and number of domains. While our overall goal is to minimise squatting rather than maximise income, it seems like hitting the economic equilibrium for price is a reasonable approximation, in the absence of hard numbers for the ideal price to discourage squatting.

This does seem to have the nice property that a large domain holder wanting to reduce their costs would likely have to obtain enough tokens to do so that it changes their incentives to instead maximising profit from those tokens.

Open questions:

  • Is the economically optimal price for domains close enough to the optimal price to discourage squatting to function?
  • Will people act rationally to find the optimum price?
  • Within what parameters does this system behave as desired? How large can the largest domain holder be before it’s economically viable for them to buy up tokens to vote the price down?


I haven’t thought about this too deeply, but my gut says that this will optimize for squatting.

Selling domains to squatters is an equilibrium in a perfectly efficient market, where the squatters estimate actual price for various names, sometimes guessing under and sometimes guessing over. Since the barrier to entry to squatting is near zero, it means it is a highly competitive market and thus we can expect the efficiency to be near perfect and the frequency of guessing over vs the frequency of guessing under averages out such that squatters in aggregate barely make any money.

Any system that tries to optimize for profit (e.g., token holder profit) I believe will naturally land on setting prices based on what squatters will pay, not what legitimate participants will pay. While there is also another equilibrium at “what legitimate participants will pay” I suspect it fluctuates more and is harder to find than targeting the squatter price.

If a goal is to minimize squatting, I suspect that this strategy won’t succeed. That being said, I’m not convinced that any strategy that doesn’t involve humans will succeed at both minimizing squatters and maximizing usage.


I don’t think I follow your argument. The optimal rent price for a squatter is $0, while the optimal profit point for token holders is higher than that.


Sorry, I meant the equilibrium is “priced such that squatters buy almost all names, rarely does a name go directly to its final owner.” There is some price point at which squatters buy almost every word/word combination ENS name and then resell to final owners. Many of the names will go unsold, while some will sell for very high margins. The profit for token holders is optimal in this scenario because the net profit margin of squatters is very low due to it being a highly competitive market with low barrier to entry (similar to how the margins for mining are low due to it being a very competitive market with low barrier to entry).

If the price goes up from there, then squatters buy fewer names (the ones less likely to sell) such that their net margins are still barely profitable (again, competition is what causes the “barely” aspect). However, moving the price up will edge out legitimate buyers from buying any names so the total revenue to token holders will go down.

I suspect that no matter what price you choose, squatters will ensure that they will dominate the buying of names. Since not all names are equal in value, there will always be a margin to be made by squatters squatting on any name that is worth more than the base price. Thus, unless you set the base price to the price of the highest value domain name or unless you have a different price for each domain name then you will end up with squatters.


I think this entire proposal assumes that token holders will act rationally when regulating a “common good”. However, I think it is more likely that prices will be pushed up to the point where ENS will fork because no one is willing to pay rent with this system anymore. It also doesn’t highlight the fact that currently 10% of domains are controlled by one person.


I initially liked the idea but if we think about it, aren’t most people end up voting to close to 0 as people were not thinking about making profit from the rent (including the squatters who were more thinking about profiting when selling) ?


If you have fewer tokens than domains (proportionally), you’d rationally vote to lower the price. But if it’s the other way around, you should vote to maximise profits - and if there are profits to be had, people should buy tokens with that in mind. The question is what distributions result in a majority voting to optimise profit.


This is a continuum - the optimal price for a squatter is 0, because that enables them to squat on the maximum number of domains for the minimum price. As you observe, we can’t eliminate squatting entirely, because it’ll always be worthwhile for valuable names, but we can reduce it by setting the price high enough it’s not cost-effective to squat on low value names in the hope of reselling them.

We now have @MicahZoltu concerned the price will go to 0, and @decanus concerned the price will go to infinity…


My concern isn’t that the price will go to zero. My concern is that this strategy will not minimize squatting (,ne of the stated goals).